Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
DRAFT FOR CONSULTATION 1 2 3 4 Medicines concordance and adherence: involving adults and carers in decisions about prescribed medicines 5 6 7 Full guideline 8 9 National Collaborating Centre for Primary Care 10 Draft for consultation, July 2008 11 12 When commenting please use page number and line number 13 Medicines concordance: full guideline DRAFT (July 2008) page 1 of 373 DRAFT FOR CONSULTATION 1 2 3 Table of contents 4 Recommendations ......................................................................................... 11 5 Interventions to increase shared decision-making about medicines (chapter 4) 6 ....................................................................................................................... 11 7 Information regarding medicines for patients and practitioners when patients 8 are discharged from inpatient care (Chapter 6).............................................. 15 9 Patients’ experience of medicine -taking (chapter 5) ..................................... 16 10 Assessment of adherence (chapter 7) ........................................................ 17 11 Interventions to increase adherence to prescribed medication (chapter 8) 18 12 Reviewing medicines (chapter 9) ............................................................... 19 13 Key Priorities for Implementation ..................................................................... 9 1 Introduction ............................................................................................. 21 14 1.1 Aim of the guideline ......................................................................... 22 15 1.2 How the guideline is set out ............................................................. 22 16 1.3 Scope .............................................................................................. 23 17 1.4 Responsibility and support for guideline development ..................... 25 18 19 20 21 22 23 The National Collaborating Centre for Primary Care (NCC-PC) 25 1.4.2 The development team ............................................................. 26 1.4.3 The Guideline Development Group (GDG) ............................... 27 1.4.4 Guideline Development Group meetings .................................. 30 1.5 Care Pathway .................................................................................. 30 24 1.6 Research recommendations ............................................................ 30 25 1.7 Acknowledgements.......................................................................... 34 26 1.8 Glossary .......................................................................................... 34 27 1.4.1 2 Methods .................................................................................................. 41 28 2.1 Introduction ...................................................................................... 41 29 2.2 Developing key clinical questions (KCQs) ....................................... 41 30 2.3 Literature search strategy ................................................................ 42 31 2.4 Identifying the evidence ................................................................... 45 32 2.5 Critical appraisal of the evidence ..................................................... 45 33 34 2.5.1 Choice of outcomes .................................................................. 45 2.6 Health Economics methods ............................................................. 46 35 36 2.6.1 2.6.2 Health Economic evidence review methodology ...................... 47 Cost-effectiveness modelling methods ..................................... 48 Medicines concordance: full guideline DRAFT (July 2008) page 2 of 373 DRAFT FOR CONSULTATION 1 2.7 Forming recommendations .............................................................. 48 2 2.8 Areas without evidence and consensus methodology ..................... 48 3 2.9 Consultation..................................................................................... 48 4 2.10 Relationships between the guideline and other national guidance .. 49 5 6 7 8 3 2.10.1 National Service Frameworks ................................................... 49 2.10.2 Related NICE Guidance ........................................................... 49 2.10.3 Other national guidance ........................................................... 49 Principles and concepts used in the development of the guideline ......... 51 9 3.1 Patients’ rights to be involved in decisions about medicines ........... 51 10 3.2 What is meant by involving patients in decisions about medicines? 52 11 3.3 What are we trying to achieve in involving patients in decisions about 12 medicines? ................................................................................................. 55 13 3.4 Roles and responsibilities of patients and health care professionals 56 14 15 3.5 Understanding the influences on medicine taking by patients ......... 57 16 3.6 Terminology and structure of guideline ............................................ 60 17 3.7 Shared decision-making about medicines ....................................... 62 18 4 Interventions to increase shared decision-making about medicines ....... 65 19 4.1 Recommendations *listed at the start of the guideline according to 20 chapter ....................................................................................................... 65 21 4.2 Introduction ...................................................................................... 65 22 4.3 Process of shared decision-making ................................................. 66 23 4.4 Methods ........................................................................................... 69 24 4.5 Key Clinical Question: Is it possible to increase patient involvement 25 in decisions about medicines?.................................................................... 71 26 27 28 4.5.1 Methods of the evidence review ............................................... 77 4.5.2 Evidence review ....................................................................... 78 4.6 Key Clinical Question: How can practitioners elicit patients’ 29 preferences for involvement in decisions about medicines?....................... 89 30 31 32 4.6.1 Methods of the evidence review ............................................... 92 4.6.2 Evidence review ....................................................................... 92 4.7 Key Clinical Question: What tools are available to help elicit patients’ 33 beliefs about medicines? ............................................................................ 97 34 35 36 4.7.1 Methods of the evidence review ............................................... 99 4.7.2 Evidence review ....................................................................... 99 4.8 Key Clinical Question: What tools are available to help elicit patients’ 37 information needs? ................................................................................... 103 Medicines concordance: full guideline DRAFT (July 2008) page 3 of 373 DRAFT FOR CONSULTATION 1 2 3 4.8.1 Method of the evidence review ............................................... 105 4.8.2 Evidence review ..................................................................... 105 4.9 Key Clinical Question: How can information about medicines be 4 provided for patients in order to enhance SDM in regard to medicines? .. 113 5 6 4.9.1 Evidence review ..................................................................... 114 4.10 Key Clinical Question: What information about medicines should be 7 provided for patients in order to enhance SDM in regards to medicine? .. 117 8 9 4.10.1 Evidence review ..................................................................... 118 4.11 Key clinical Question: Mental capacity narrative ............................ 128 10 11 12 4.11.1 **Roles and responsibilities of patients and health care professionals......................................................................................... 128 4.12 Key Clinical Question: What information about shared decision 13 making and adherence should be recorded in patients’ notes? ................ 130 14 4.13 15 in reaching an informed decision? How effective are these tools? ........... 131 16 17 18 4.13.1 Methods of evidence review ................................................... 134 4.13.2 Evidence review ..................................................................... 134 4.14 Key Clinical Question: How can a practitioner elicit whether a patient 19 agrees with the prescription recommended by the practitioner? .............. 142 20 21 22 4.14.1 Methods of the evidence review ............................................. 143 4.14.2 Evidence review ..................................................................... 143 4.15 Key Clinical Question: What aspects of consultation style increase 23 patient involvement in decision-making? .................................................. 145 24 25 26 4.15.1 Methods of the evidence review ............................................. 147 4.15.2 Evidence review ..................................................................... 148 4.16 Key Clinical Question: Do interventions to increase patient 27 involvement increase length of the consultation? ..................................... 154 28 29 30 4.16.1 Methods of the evidence review ............................................. 158 4.16.2 Evidence review ..................................................................... 159 Patients’ experience of medicine-taking ............................................... 169 5 Key Clinical Question: What tools are available to support the patient 31 5.1 Recommendations *listed at the start of the guideline according to 32 chapter ..................................................................................................... 169 33 5.2 Introduction .................................................................................... 169 34 5.3 Key Clinical Question: what are the barriers and facilitators for 35 individuals in medicine-taking ................................................................... 169 36 37 38 39 40 5.3.1 Methods of the evidence review ............................................. 173 5.3.2 Evidence review ..................................................................... 174 5.3.3 The evaluation of medicines and the difficulties encountered by people in evaluating medicines ............................................................. 175 5.3.4 Medicines and identity ............................................................ 177 Medicines concordance: full guideline DRAFT (July 2008) page 4 of 373 DRAFT FOR CONSULTATION 1 2 5.3.5 Ways people take their medication ......................................... 178 5.4 Update of qualitative evidence synthesis - Pound 2005 synthesis 180 3 4 5 5.4.1 Methodology of update ........................................................... 180 5.4.2 Evidence review ..................................................................... 181 5.5 Systematic reviews of barriers and facilitators for individuals in 6 medicine taking ........................................................................................ 190 7 8 9 5.5.1 Methods of the update ............................................................ 190 5.5.2 Evidence review ..................................................................... 190 Information regarding medicines for patients and practitioners when 10 6 patients are discharged from inpatient care ................................................. 193 11 6.1 12 chapter ..................................................................................................... 193 13 6.2 Introduction .................................................................................... 193 14 6.3 Key clinical question: What information regarding medicines should 15 be provided for patients and practitioners when patients are discharged 16 from secondary care ................................................................................. 194 17 18 19 6.3.1 Methods of the evidence review ............................................. 195 6.3.2 Evidence review ..................................................................... 195 Assessment of adherence .................................................................... 197 7 Recommendations *listed at the start of the guideline according to 20 7.1 21 chapter ..................................................................................................... 197 22 7.2 Introduction .................................................................................... 197 23 7.3 Key Clinical Question: What are the advantages and disadvantages 24 of self-report in assessing patient’s adherence? ...................................... 198 25 26 27 7.3.1 Methods of the evidence review ............................................. 203 7.3.2 Evidence review ..................................................................... 203 Interventions to increase adherence to prescribed medication ............. 218 8 Recommendations *listed at the start of the guideline according to 28 8.1 29 chapter ..................................................................................................... 218 30 8.2 Introduction .................................................................................... 218 31 8.3 Methods ......................................................................................... 219 32 8.4 Evidence to recommendations : difficulties in interpreting studies on 33 interventions to improve adherence ......................................................... 221 34 8.5 35 adherence? .............................................................................................. 224 36 37 Recommendations *listed at the start of the guideline according to Key Clinical Question: Does change in dosing regime affect 8.5.1 8.5.2 Methods of the evidence review ............................................. 225 Evidence review ..................................................................... 226 Medicines concordance: full guideline DRAFT (July 2008) page 5 of 373 DRAFT FOR CONSULTATION 1 8.6 2 adherence to prescribed medication ........................................................ 233 3 4 5 8.6.1 Methods of the evidence review ............................................. 235 8.6.2 Evidence review ..................................................................... 235 8.7 Key Clinical Question: Does drug formulation and/or packaging affect 6 adherence? .............................................................................................. 239 7 8 9 8.7.1 Methods of the evidence review ............................................. 241 8.7.2 Evidence review ..................................................................... 241 8.8 Key Clinical Question: Is there any evidence on interventions that 10 aim to minimize side-effects in order to increase adherence? .................. 247 11 12 13 8.8.1 Methods of the evidence review ............................................. 249 8.8.2 Evidence review ..................................................................... 249 8.9 Key Clinical Question: How does the way the information is 14 presented (e.g. pictorial vs. written) affects adherence? .......................... 257 15 16 17 8.9.1 Methods of the evidence review ............................................. 259 8.9.2 Evidence review ..................................................................... 259 8.10 Key Clinical Question: Do specific forms of therapy (e.g. CBT) affect 18 adherence? .............................................................................................. 263 19 20 21 8.10.1 Methods of the evidence review ............................................. 267 8.10.2 Evidence review ..................................................................... 268 8.11 Key Clinical Question: Would a contractual agreement between HCP 22 and patient affect adherence? .................................................................. 278 23 24 25 26 8.11.1 Methods of the evidence review ............................................. 279 8.11.2 Definitions of contracts ........................................................... 279 8.11.3 Evidence review ..................................................................... 280 8.12 Key Clinical Question: Do reminders (and what types of reminders, 27 text messaging etc) increase adherence to prescribed medication? ........ 282 28 29 30 8.12.1 Methods of the evidence review ............................................. 285 8.12.2 Evidence review ..................................................................... 285 8.13 Key Clinical Question : Does being involved in self-monitoring (e.g. of 31 own blood pressure) increase adherence to prescribed medication? ....... 293 32 33 34 8.13.1 Methods of the evidence review ............................................. 297 8.13.2 Evidence review ..................................................................... 298 8.14 Key Clinical Question : Does the use of dosette boxes increase 35 adherence to prescribed medication? ...................................................... 307 36 37 38 8.14.1 Methods of the evidence review ............................................. 307 8.14.2 Evidence review ..................................................................... 307 Reviewing medicines ............................................................................ 308 9 Key Clinical Question : Effect of prescription charges/costs on 39 9.1 Recommendations *listed at the start of the guideline according to 40 chapter ..................................................................................................... 308 Medicines concordance: full guideline DRAFT (July 2008) page 6 of 373 DRAFT FOR CONSULTATION 1 9.2 Introduction .................................................................................... 308 2 9.3 Key Clinical Question: does medication review increase shared 3 decision–making or adherence?............................................................... 310 4 5 6 9.3.1 Methods of the evidence review ............................................. 314 9.3.2 Evidence review ..................................................................... 315 10 Health Economics and Medicines Concordance ............................... 324 7 10.1 Introduction .................................................................................... 324 8 9 10 10.1.1 How terminology of compliance, concordance and adherence is used in HE literature ............................................................................. 324 10.2 Health Economics methods ........................................................... 325 11 12 13 10.2.1 Health Economic evidence review methodology .................... 326 10.2.2 Cost-effectiveness modelling methods ................................... 327 10.3 Shared decision making and medicine taking from a health economic 14 perspective ............................................................................................... 327 15 16 17 10.3.1 Patient involvement and shared decision making ................... 327 10.3.2 Medicine-taking ...................................................................... 330 10.4 Cost effectiveness literature review of Interventions to increase 18 adherence ................................................................................................ 332 19 20 21 22 23 24 25 10.4.1 Methods of the review ............................................................ 332 10.4.2 Interventions to increase adherence: cost effectiveness systematic reviews narrative ................................................................. 333 10.4.3 Summary of findings and methodological issues highlighted by systematic reviews ................................................................................ 333 10.4.4 Conclusion .............................................................................. 338 10.5 Update of the systematic review by Elliott ..................................... 338 26 27 28 10.5.1 Methods of the review ............................................................ 338 10.5.2 Update review narrative .......................................................... 339 10.6 Discussion ..................................................................................... 345 29 30 31 32 33 34 35 10.6.1 Theoretical discussion of review findings, caveats and opportunities for future research ........................................................... 345 10.6.2 Recommendations for minimum reporting requirements ........ 347 10.6.3 Recommendations for future research ................................... 349 Please note the appendices are available as separate files. Medicines concordance: full guideline DRAFT (July 2008) page 7 of 373 DRAFT FOR CONSULTATION 1 2 3 Preface *to be added to final version 4 Medicines concordance: full guideline DRAFT (July 2008) page 8 of 373 DRAFT FOR CONSULTATION 1 2 Key Priorities for Implementation 3 Offer all patients the opportunity to be involved in making decisions about 4 prescribed medicines. Establish what level of involvement the patient would 5 like. (1.1.1) 6 Ask if the patient has any specific concerns about their medicines, 7 whenever you prescribe, dispense or review medication. Address these 8 concerns. (1.1.8) 9 10 11 Offer patients information about medicines before the medicines are prescribed. (1.1.10) Actively discuss information on medicines with the patient rather than just 12 presenting it. The discussion should take into account the patient’s 13 understanding and beliefs about the diagnosis and treatment. (1.1.12) 14 The information offered to patients about medicines should include: 15 what the medicine is 16 how it works 17 how the medicine affects their condition (that is, its benefits) 18 potential side effects 19 any important instructions on how medicine should be taken 20 what to do if they miss a dose 21 if the medicine should be continued following the initial 22 prescription.(1.1.13) 23 24 Be aware that a shared decision may mean an agreement not to prescribe 25 a medication or for the patient to stop taking a medication. If in the 26 healthcare professional’s view this may have an adverse effect, then this 27 must be recorded. (1.1.20) Medicines concordance: full guideline DRAFT (July 2008) page 9 of 373 DRAFT FOR CONSULTATION 1 Accept the patient’s right to decide not to take a medicine, even when you 2 do not agree with the decision, as long as the patient has capacity to 3 consent. (1.1.22) 4 Record the discussion of decisions about medicine and medicine-taking in 5 patients’ records where there are concerns about medicines. This may 6 include: 7 patients’ beliefs and concerns about medicines 8 information given to the patient 9 potential problems with adherence plans for review of medication. (1.1.25) 10 11 12 Non-adherence can be assessed by asking the patient if they have missed 13 any medications recently. Make it easier for the patient to report non- 14 adherence by: 15 asking the question in a way that does not apportion blame 16 explaining why you are asking the question 17 using a specific time period such as ‘in the last week’ 18 asking about specific medicine-taking behaviours. (1.2.3) 19 20 Tailor any intervention to increase adherence to the specific difficulties with 21 adherence experienced by a patient. Be aware that adherence can be 22 improved but no specific intervention can be recommended for all patients. 23 (1.3.2) 24 Medicines concordance: full guideline DRAFT (July 2008) page 10 of 373 DRAFT FOR CONSULTATION 1 2 Recommendations 3 Interventions to increase shared decision-making 4 about medicines (chapter 4) 5 Establishing patients’ wish for involvement 6 1.1.1 Offer all patients the opportunity to be involved in making decisions 7 about prescribed medicines. Establish what level of involvement the patient 8 would like. 9 1.1.2 Establish the most effective way of communicating with each patient 10 and where necessary consider ways of making information accessible and 11 understandable (for example, via pictures, symbols, large print and different 12 languages). 13 1.1.3 Avoid making any unwarranted assumptions about patient preferences 14 about treatment. Talk to the patient to find out their preferences about 15 treatment and note any non-verbal clues that may indicate you need to 16 explore the patient’s perspective further. 17 1.1.4 Tailor your consultation style to the needs of individual patients to 18 ensure that all patients can be involved in decision-making in the way they 19 wish. 20 21 How to explore patients’ beliefs 22 1.1.5 Encourage patients to ask questions about their condition and 23 treatment. 24 1.1.6 Ask patients open-ended questions because these are more likely to 25 uncover patients concerns. 26 1.1.7 Ask about the patient’s beliefs about medicines both before starting 27 new treatments and periodically during medication review. Medicines concordance: full guideline DRAFT (July 2008) page 11 of 373 DRAFT FOR CONSULTATION 1 1.1.8 Ask if the patient has any specific concerns about their medicines, 2 whenever you prescribe, dispense or review medication. Address these 3 concerns. 4 5 How to provide information for patients 6 1.1.9 Explain the goals of medical treatment to patients, openly discussing 7 the pros and cons of proposed medication. The discussion should be at the 8 level expected by the patient. 9 1.1.10 Offer patients information about medicines before the medicines are 10 prescribed. 11 1.1.11 Check patients have any information they wish about medicines before 12 medicines are dispensed. 13 1.1.12 Actively discuss information on medicines with the patient rather than 14 just presenting it. The discussion should take into account the patient’s 15 understanding and beliefs about the diagnosis and treatment. 16 1.1.13 The information offered to patients about medicines should include: 17 • what the medicine is 18 • how it works 19 • how the medicine affects their condition (that is, its benefits) 20 • potential side effects 21 • any important instructions on how medicine should be taken 22 • what to do if they miss a dose 23 • if the medicine should be continued following the initial 24 prescription. Medicines concordance: full guideline DRAFT (July 2008) page 12 of 373 DRAFT FOR CONSULTATION 1 1.1.14 To help patients make decisions about medicines, offer them relevant 2 information which is easy to understand and free from jargon. 3 1.1.15 Suggest where patients might find more information and support after 4 the consultation: for example, by providing written information or directing 5 them to other resources or expert patient groups (for example, 6 www.patient.co.uk, www.easyhealth.org.uk). 7 1.1.16 Be careful not to make assumptions about a patient’s ability to 8 understand the information provided. Check with the patient that they have 9 understood the information. 10 1.1.17 Information for patients should be structured and when possible 11 tailored to the needs of the individual patient. 12 13 How to support the patient 14 1.1.18 Clarify the patient’s goals from treatment. 15 1.1.19 Healthcare professionals have a duty to help patients to make 16 decisions about their treatment which are informed by an understanding of the 17 likely benefits and risks rather than by patients’ beliefs. 18 1.1.20 Be aware that a shared decision may mean an agreement not to 19 prescribe a medication or for the patient to stop taking a medication. If in the 20 healthcare professional’s view this may have an adverse effect, then this must 21 be recorded. 22 1.1.21 Be aware that patients may have different views from healthcare 23 professionals regarding balance of risks, benefits and side effects of 24 medicines. 25 1.1.22 Accept the patient’s right to decide not to take a medicine, even when 26 you do not agree with the decision, as long as the patient has capacity to 27 consent. Medicines concordance: full guideline DRAFT (July 2008) page 13 of 373 DRAFT FOR CONSULTATION 1 1.1.23 Assess capacity to make an informed decision using the principles in 2 the Mental Capacity Act (2005). To lack capacity patients must (a) have 3 disturbance or malfunction of mind and (b) demonstrate lack of capacity to: 4 • understand the information relevant to the decision 5 • retain information for long enough to use it in the decision 6 • use or weigh information 7 • communicate the decision. 8 1.1.24 Encourage and support patients, families and caregivers to keep and 9 maintain an accurate list of all medications, including prescription and non- 10 prescription medications and herbal and nutritional supplements, and any 11 allergic or adverse reactions to medication. 12 1.1.25 Record the discussion of decisions about medicine and medicine- 13 taking in patients’ records where there are concerns about medicines. This 14 may include: 15 • patients’ beliefs and concerns about medicines 16 • information given to the patient 17 • potential problems with adherence 18 • plans for review of medication. 19 1.1.26 Consider that good communication is necessary for increasing patient 20 involvement and that there are methods for improving your skills for increasing 21 patient involvement in decisions about their care. 22 1.1.27 Be aware that simple interventions to increase patient involvement do 23 not necessarily increase the overall length of consultation. Any extra time 24 spent in a single consultation may be justified by benefits, particularly over the 25 course of a long-term condition. Medicines concordance: full guideline DRAFT (July 2008) page 14 of 373 DRAFT FOR CONSULTATION 1 Information regarding medicines for patients and 2 practitioners when patients are discharged from 3 inpatient care (Chapter 6) 4 1.1.28 Before discharge from hospital, offer patients information about their 5 medicines. This information should include: 6 • what the medicine is 7 • why it is necessary 8 • how the medicine affects their condition 9 • its benefits and harms 10 • potential side effects 11 • any important instructions on how the medicine should be taken 12 • how to get a further supply 13 • likely duration of treatment 14 • what to do in case of adverse effects 15 • any special considerations (for example, drug interactions, 16 storage). 17 1.1.29 On discharge from hospital, give all patients a report containing: 18 • diagnosis 19 • a list of all medications the patient should be taking 20 • clear identification of new medications initiated during the 21 hospital stay 22 • clear identification of medications stopped during the hospital 23 stay with reasons Medicines concordance: full guideline DRAFT (July 2008) page 15 of 373 DRAFT FOR CONSULTATION 1 • clear information on which medications should be continued 2 after discharge and for how long 3 • known adverse reactions and allergies 4 • any potential problems with adherence and any actions taken 5 (for example, a compliance aid or reminder system). 6 1.1.30 When a patient is transferred between services, the next provider of 7 care should be given a report with the following information: 8 • a written report of diagnosis 9 • a list of all medications that the patient should be taking 10 • clear identification of new medications initiated during the 11 hospital stay 12 • clear identification of medications stopped during the hospital 13 stay with reasons 14 • clear information on which medications should be continued 15 after discharge 16 • known adverse reactions and allergies 17 • any potential problems with adherence and actions taken (for 18 example, use of dosette box). 19 20 Patients’ experience of medicine-taking (chapter 5) 21 1.1.31 Be aware that patients’ beliefs about medicines, in particular patients’ 22 concerns about medicines and patients’ belief in their personal need for the 23 medicines, affect how and whether they take their medicine. Medicines concordance: full guideline DRAFT (July 2008) page 16 of 373 DRAFT FOR CONSULTATION 1 1.1.32 Elicit and address patients’ specific concerns about taking their 2 medicine. These may include concern about becoming dependent on 3 medicines and the side effects and adverse effects of medicines. 4 1.1.33 Discuss with the patient why they need the treatment. Adopt a common 5 sense approach that addresses the need for the treatment (the disease or 6 condition) and the solution (the treatment). 7 1.1.34 Be aware that patients may wish to minimise how much medication 8 they take. 9 1.1.35 Be aware that patients may wish to discuss: 10 • what will happen if they do not take medicine suggested by their 11 doctor 12 • non-pharmacological alternatives to medicines 13 • how to reduce and stop medication they may have been taking 14 for a long time 15 • how to fit the medicine into their daily routine 16 • how to make a choice between medicines if they believe they 17 are taking too many medicines. 18 1.1.36 Be aware that patients evaluate prescribed medicines using their own 19 subjective or objective indicators. These include stopping and starting the 20 medicine, altering the dose of the medicine and checking how the medicine 21 affects their symptoms. 22 23 Assessment of adherence (chapter 7) 24 1.2.1 Recognise that non-adherence is common, and most people are non- 25 adherent sometimes. Medicines concordance: full guideline DRAFT (July 2008) page 17 of 373 DRAFT FOR CONSULTATION 1 1.2.2 Routinely assess adherence in a non-judgemental way as an integral 2 part of medicine prescribing, dispensing and review. 3 1.2.3 Non-adherence can be assessed by asking the patient if they have 4 missed any medications recently. Make it easier for the patient to report non- 5 adherence by: 6 • asking the question in a way that does not apportion blame 7 • explaining why you are asking the question 8 • using a specific time period such as ‘in the last week’ 9 • asking about specific medicine-taking behaviours. 10 1.2.4 Consider using records of prescription re-ordering and pharmacy refill 11 records to alert prescribers and dispensers to non-adherence. 12 1.2.5 When a healthcare professional identifies issues around adherence this 13 should be recorded and communicated within the healthcare team to optimise 14 patient involvement and care. 15 16 Interventions to increase adherence to prescribed medication 17 (chapter 8) 18 1.3.1 Discuss with the patient who is not taking their medicines whether this 19 is because of beliefs and concerns about medicines or because of practical 20 problems. 21 1.3.2 Tailor any intervention to increase adherence to the specific difficulties 22 with adherence experienced by a patient. Be aware that adherence can be 23 improved but no specific intervention can be recommended for all patients. 24 1.3.3 Find out what form of support the patient would prefer to increase their 25 adherence to medicines. Together, you and your patient should consider 26 options for support. Medicines concordance: full guideline DRAFT (July 2008) page 18 of 373 DRAFT FOR CONSULTATION 1 1.3.4 Consider whether to suggest that patients record their medicine-taking 2 in a diary to aid adherence. 3 1.3.5 Be aware that encouraging patients to monitor their condition may 4 increase patients’ adherence to prescribed medication. 5 1.3.6 Simplify the dosing regimen if this is a problem for the patient. 6 1.3.7 Use special packaging for the medicine when the standard packaging 7 is a problem for the patient. 8 1.3.8 Side effects can be a problem for some patients and in this situation: 9 • discuss how the patient would like to deal with side effects 10 • discuss benefits, side effects and long-term effects with patients 11 to allow patients to make an informed choice 12 • consider adjusting medication dosage 13 • consider switching to an alternative that has a different risk of 14 side effects 15 • consider what other strategies other than adjusting dosage or 16 type of medication might be used (for example, timing of 17 medicines). 18 1.3.9 Reminder systems are likely to be helpful to some patients. The 19 method of delivery should meet individual needs. 20 1.3.10 Ask patients if the costs of prescriptions are a problem for them. If cost 21 of prescription is a problem, patients may wish to know which medicines are 22 most important. 23 Reviewing medicines (chapter 9) 24 1.4.1 Offer repeat information and review to patients when necessary, 25 especially when treating long-term conditions with multiple medications. Medicines concordance: full guideline DRAFT (July 2008) page 19 of 373 DRAFT FOR CONSULTATION 1 1.4.2 Review practitioner and patients’ beliefs about medicines at intervals 2 agreed with the patient because these may change over time. 3 1.4.3 Patients should have the decision to prescribe medicines reviewed at 4 regular intervals according to patient choice and patient need. 5 1.4.4 A review of medicines should include an enquiry into medicine 6 adherence and where non-adherence is identified, possible causes should be 7 clarified and an agreement made with the patient about any appropriate 8 action. Any plan should include a date for a follow-up review. 9 1.4.5 Healthcare professionals involved in medication review should inform 10 the prescribing doctor of the review and its outcome, particularly when the 11 review involves adherence issues and further review is necessary. Medicines concordance: full guideline DRAFT (July 2008) page 20 of 373 DRAFT FOR CONSULTATION 1 2 1 Introduction 3 The prescription of drugs is now a central part of the delivery of medical care. 4 The total drugs budget for the NHS in 2006/07 was approximately £10.6billion. 5 Of this around £7.6 billion was spent in primary care. Expenditure on primary 6 care drugs has increased by over 60% in the last decade or an average of 7 4.8% in real terms each year1. 8 Reviews across different disease areas report that between 30-50% of 9 patients do not take or use the medicines prescribed for them. The patient 10 was in the past thought to be the ‘source of the problem of compliance’ but it 11 is now acknowledged that medicine - taking is a complex behaviour and 12 influencing this involves not only patients but providers of health care and 13 health care systems (World Health Organization, 2003). The term adherence 14 is now used to describe the extent to which patients’ medicine taking 15 behaviour matches agreed recommendations from the prescriber. This 16 understanding presumes an agreement between patient and prescriber. 17 Attention is therefore required to the decision-making process between 18 prescribers and patients. 19 The estimated costs of unused or unwanted medicines in the NHS have been 20 estimated to exceed £100 million annually2 The WHO report notes that while 21 interventions to promote adherence are not generally seen to provide savings 22 in economic studies from an institutional perspective, cost savings are 23 demonstrated from a societal view point due to improvements in patient 24 quality of life, indirect costs avoided and effect on productivity. As with all 25 NICE guidelines a cost impact report will be developed for this guideline once 26 recommendations have been finalised and this will be published with the final 27 version of the guideline. 28 While objective health and cost impacts of patients’ behaviour in relation to 29 medicines is important, the right of patients to make decisions in regard to 1 2 http://www.official-documents.gov.uk/document/cm73/7341/7341.pdf http://www.official-documents.gov.uk/document/cm73/7341/7341.pdf Medicines concordance: full guideline DRAFT (July 2008) page 21 of 373 DRAFT FOR CONSULTATION 1 their own health is accepted in modern practice. The approach taken by this 2 guideline is that patients have a right to be involved in decisions about 3 medicines to the extent that they wish and it is the role of health professionals 4 to facilitate and support patients in their involvement in decision-making and to 5 support patients in taking medicine if the decision has been to prescribe. 6 Adherence to medicines taking should be based on shared decision making 7 between the patient and the practitioner. 8 9 We have not made separate recommendations for carers and families. The 10 principle relationship is between patient and health care professional and the 11 patient has a right to decide who should be involved in their care. With patient 12 consent carers should have access to same levels of information and support. 13 There are an increasing number of healthcare professionals now involved in 14 prescribing of medicines, dispensing and reviewing of medicines. It is not 15 within the remit of a guideline to recommend which health care professional 16 carries out these roles. Our recommendations apply to all healthcare 17 professionals who do carry out these roles but professionals also need to 18 work within legal and professional codes 19 20 1.1 Aim of the guideline 21 Clinical guidelines are defined as ‘systematically developed statements to 22 assist practitioner and patient decisions about appropriate healthcare for 23 specific clinical circumstances. 24 This guideline gives recommendations to clinicians and others on how to 25 involve adults and carers in decisions about prescribed medication. 26 1.2 27 The recommendations for all the topics in each chapter are listed at the start 28 of the chapter. Both the evidence statements and narratives of the research 29 studies on which our recommendations are based are found within each topic How the guideline is set out Medicines concordance: full guideline DRAFT (July 2008) page 22 of 373 DRAFT FOR CONSULTATION 1 section. The evidence statements precede the narrative for each topic. Also 2 included in each chapter is a brief explanation of why the GDG made the 3 specific recommendations. The evidence tables with details of the research 4 studies that describe the studies reviewed are found in Appendix C. 5 1.3 6 The guideline was developed in accordance with a scope given by the 7 National Institute for Health and Clinical Excellence (NICE, ‘the Institute’). The 8 scope set the remit of the guideline and specified those aspects of the 9 identification and management of medicines concordance and adherence to Scope 10 be included and excluded. The scope was published in April 2007 and is 11 reproduced here in Appendix A. 12 Whom the guideline is intended for 13 This guideline is of relevance to those who work in or use the National Health 14 Service (NHS) in England and Wales: 15 Population 16 Groups that will be covered 17 a) Adults, including those with comorbidities, learning disabilities or language 18 and/or cultural differences. 19 Groups that will not be covered 20 Children and young people. However, the guideline recommendations may be 21 considered for a child or young person who is deemed competent to express 22 a view on their prescription. 23 Healthcare setting 24 All consultations with healthcare professionals in any NHS setting that relate 25 to the initiation or review of prescribed medication. 26 Areas that will be covered Medicines concordance: full guideline DRAFT (July 2008) page 23 of 373 DRAFT FOR CONSULTATION 1 a) Shared decision-making about medicines and medicine taking as reported 2 by the patient or carer. The guideline will focus on the barriers (such as 3 communication difficulties, cultural issues, low health literacy and physical 4 limitations), facilitators (including structural or procedural factors), beliefs and 5 health behaviours that influence decision-making and adherence. 6 b) Shared decision-making about medicines and medicine taking as reported 7 by the healthcare professional. The guideline will focus on the barriers (such 8 as communication difficulties, cultural issues and time), facilitators (including 9 structural or procedural factors), beliefs and health behaviours that influence 10 decision-making and adherence. 11 c) The effectiveness and cost effectiveness of interventions to facilitate the 12 process of shared decision-making about medicines (looking at time of 13 intervention – before, during, or after the consultation with the healthcare 14 professional; and mode of delivery). The target of the intervention may be the 15 patient, the carer, the prescriber, any healthcare professional providing 16 ongoing support or a combination of these. 17 d) The effectiveness and cost effectiveness of interventions to promote 18 adherence in medicine taking (looking at time of intervention – before, during, 19 or after the consultation with the healthcare professional; and mode of 20 delivery). The target of the intervention may be the patient, the carer, the 21 prescriber, the dispenser or any other healthcare professional providing 22 ongoing support or a combination of these. 23 e) The evidence on single or multiple medications as it relates to issues 24 around decision-making and adherence. 25 f) The skills and competencies required by prescribers to involve patient in 26 decisions regarding prescribed medicines. 27 Areas outside the remit of the guideline 28 The administration of medicines will not be covered. Administration is defined 29 as giving a medicine by introduction into the body (for example, orally or by Medicines concordance: full guideline DRAFT (July 2008) page 24 of 373 DRAFT FOR CONSULTATION 1 injection), or by external application (for example application of an 2 impregnated dressing). 3 1.4 4 The NCC-PC is a partnership of primary care professional associations and 5 was formed as a collaborating centre to develop guidelines under contract to 6 NICE. It is entirely funded by NICE. The NCC-PC is contracted to develop four 7 guidelines at any one time, although there is some overlap at start and finish. 8 Unlike many of the other centres which focus on a particular clinical area, the 9 NCC-PC has a broad range of topics relevant to primary care. However, it Responsibility and support for guideline development 10 does not develop guidelines exclusively for primary care. Each guideline may, 11 depending on the scope, provide guidance to other health sectors in addition 12 to primary care. 13 The Royal College of General Practitioners (RCGP) acts as the host 14 organisation. The Royal Pharmaceutical Society and the Community 15 Practitioners and Health Visitors’ Association are partner members with 16 representation from other professional and lay bodies on the Board. The 17 RCGP holds the contract with the Institute for the NCC-PC. 18 1.4.1 19 PC) 20 The NCC-PC is a partnership of primary care professional associations and 21 was formed as a collaborating centre to develop guidelines under contract to 22 NICE. It is entirely funded by NICE. The NCC-PC is contracted to develop five 23 guidelines at any one time, although there is some overlap at start and finish. 24 Unlike many of the other centres which focus on a particular clinical area, the 25 NCC-PC has a broad range of topics relevant to primary care. However, it 26 does not develop guidelines exclusively for primary care. Each guideline may, 27 depending on the scope, provide guidance to other health sectors in addition 28 to primary care. 29 The Royal College of General Practitioners (RCGP) acts as the host 30 organisation. The Royal Pharmaceutical Society and the Community 31 Practitioners and Health Visitors’ Association are partner members with The National Collaborating Centre for Primary Care (NCC- Medicines concordance: full guideline DRAFT (July 2008) page 25 of 373 DRAFT FOR CONSULTATION 1 representation from other professional and lay bodies on the Board. The 2 RCGP holds the contract with the Institute for the NCC-PC. 3 The development team 4 1.4.2 5 The development team had the responsibility for this guideline throughout its 6 development. They were responsible for preparing information for the 7 Guideline Development Group (GDG), for drafting the guideline and for 8 responding to consultation comments. The development team working on this 9 guideline consisted of the: 10 Guideline lead 11 who is a senior member of the NCC-PC team who has overall 12 responsibility for the guideline 13 Information scientist 14 who searched the bibliographic databases for evidence to 15 answer the questions posed by the GDG 16 Reviewer (Health Services Research Fellow) 17 with knowledge of the field, who appraised the literature and 18 abstracted and distilled the relevant evidence for the GDG 19 Health economist 20 who reviewed the economic evidence and assisted the GDG in 21 considering cost effectiveness 22 Project manager 23 who was responsible for organising and planning the 24 development, for meetings and minutes and for liaising with the 25 Institute and external bodies 26 Chair 27 who was responsible for chairing and facilitating the working of 28 the GDG meetings 29 The members of the development team attended the GDG meetings and 30 participated in them. The development team also met regularly with the Chair Medicines concordance: full guideline DRAFT (July 2008) page 26 of 373 DRAFT FOR CONSULTATION 1 of the GDG during the development of the guideline to review progress and 2 plan work. 3 Other guidelines normally have a clinical advisor who is someone with an 4 academic understanding of the research in the area and its practical 5 implications to the service, who advised the development team on searches 6 and the interpretation of the literature. Due to the conceptual nature of the 7 guideline topic and the different academic stances on explaining such 8 behaviour the development team chose not to have a formal clinical advisor. 9 1.4.3 The Guideline Development Group (GDG) 10 A Chair was chosen for the group and his primary role was to facilitate and 11 chair the GDG meetings. 12 The GDG consisted of diverse multidisciplinary group with an interest and/or 13 expertise in medicines concordance. The Chair, a general practitioner with 14 special interest in epilepsy identified by the NCC-PC, oversaw the work of the 15 group. 16 Nominations for group members were invited from various stakeholder 17 organisations, selected to ensure appropriate combination of members 18 including healthcare professionals and patient representatives. 19 Each nominee was expected to act as an individual expert in their own right 20 and not as a representative of their parent organisation, although they were 21 encouraged to keep their nominating organisation informed of the process. 22 Group membership can be found in the preface to the guidance. 23 In accordance with guidance from NICE, all GDG members’ interests were 24 recorded on a standard declaration form that covered consultancies, fee-paid 25 work, share-holdings, fellowships, and support from the healthcare industry. 26 Details of these can be seen in Appendix E. 27 The names of GDG members appear listed below. 28 Full GDG members 29 Dr Henry Smithson (Chair) Medicines concordance: full guideline DRAFT (July 2008) page 27 of 373 DRAFT FOR CONSULTATION 1 GP Escrick Research Practice York, Senior Clinical University 2 Teacher Academic Unit of Primary Medical Care Sheffield 3 Professor Rob Horne 4 Professor of Behavioural Medicine, Head of Department of Practice 5 and Policy, The School of Pharmacy, University of London, London 6 Dr John Benson 7 Senior Lecturer in General Practice, General Practice and Primary 8 Care Research unit, Department of Public Health and Primary 9 Care, University of Cambridge. 10 Mr Shaun Johnson 11 Patient representative, Melton Mowbray 12 Mrs Alison Bowser 13 Independent patient advocate, patient representative for Medicines 14 and Healthcare Products Regulatory Authority, Cornwall and Isles 15 of Scilly PCT, Royal College of GPs 16 Dr Mahendra Patel 17 Lecturer/Research Fellow, Institute of Pharmaceutical Innovation, 18 Bradford 19 Mr Stephen Hemingway 20 Senior Lecturer in Mental Health, University of Huddersfield 21 Mrs Bunis Packham 22 Nurse Consultant-Thrombosis and Anticoagulation, Barnet 23 Hospital, Barnet 24 Mr Jim Blair 25 President of the Forum on Intellectual Disability at the Royal 26 Society of Medicine and Committee Member of the Royal College 27 of Nursing's National Learning Disability Forum 28 Professor Peter Crome 29 Professor of Geriatric Medicine, Keele University, Keele, 30 Staffordshire and Hon. Consultant Geriatrician, North Staffordshire 31 Combined Healthcare NHS Trust, Stoke-on-Trent. 32 Dr Peter Haddad Medicines concordance: full guideline DRAFT (July 2008) page 28 of 373 DRAFT FOR CONSULTATION 1 Consultant Psychiatrist, Cromwell House, Greater Manchester 2 West Mental Health NHS Foundation Trust Honorary Senior 3 Lecturer in Psychiatry, University of Manchester 4 Dr Jonathan Steel 5 Chairman GP Network, Royal College of Physicians, London. 6 General Practitioner, Uley, Gloucestershire 7 Dr Sean Kelly 8 Consultant Physician and Gastrolenterologist, York Hospital, York. 9 Honorary senior lecturer Hull York Medical School. 10 Dr Wendy Clyne 11 Assistant Director, Medicines Partnership Programme, NPC Plus, 12 Keele University Science and Business Park, Keele 13 14 15 16 17 Members of the GDG from the NCC-PC were: Dr. Norma O’Flynn Guideline Lead and Clinical Director, NCC-PC Ms Elizabeth Shaw 18 Guideline Lead and Deputy Chief Executive, NCC-PC (until 19 February 2007) 20 Ms Vanessa Nunes 21 Senior Health Services Research Fellow/Project Manager, NCC- 22 PC 23 Ms Julie Neilson 24 Health Services Research Fellow, NCCPC 25 Ms Stefanie Kuntze 26 Health Economist, NCCPC 27 Mr. Gary Britton 28 Health Services Research Fellow, NCCPC (until October 2007) 29 Ms Marian Cotterell 30 Information Scientist, NCC-PC 31 Medicines concordance: full guideline DRAFT (July 2008) page 29 of 373 DRAFT FOR CONSULTATION 1 Observers Ms Sarah Willett 2 3 Commissioning Manager, National Institute for Health and 4 Clinical Excellence (from December 2007) Ms Colette Marshall 5 6 Commissioning Manager, National Institute for Health and 7 Clinical Excellence (until August 2007) Dr. Francoise Cluzeau 8 Technical advisor, National Institute for Health and Clinical 9 Excellence 10 Guideline Development Group meetings 11 1.4.4 12 The GDG met on 10 occasions (with one two day GDG meeting), at 13 approximately 2 monthly intervals over a period of 11 months and 6 weekly 14 intervals over a period of 6 months to review the evidence identified by the 15 project team, to comment on its quality and completeness and to develop 16 recommendations for clinical practice based on the available evidence. The 17 final recommendations were agreed by the full GDG. 18 1.5 19 **SEE NICE version 20 1.6 21 The Guideline Development Group has made the following recommendations 22 for research, based on its review of evidence, to improve NICE guidance and 23 patient care in the future. The Guideline Development Group’s full set of 24 research recommendations is detailed in the full guideline. 25 The GDG noted the generally poor quality of research in the area of 26 medicines adherence and the potential clinical and economic gains that would 27 accrue from achieving and implementing a shared decision about medicines. 28 The GDG believe that there is an urgent need to provide specific adherence 29 funding streams to support structured programmes of research particularly 30 where the health gains from medicines adherence are likely to be high. The Care Pathway Research recommendations Medicines concordance: full guideline DRAFT (July 2008) page 30 of 373 DRAFT FOR CONSULTATION 1 central theme underpinning this guideline is that adherence to medicines 2 taking is an individual behaviour that should be based on shared decision 3 making principally between the patient and the practitioner. The key research 4 agenda is therefore one of behaviour-change for practitioners and patients. 5 However, medicines carry the potential for harm as well as benefit and there 6 are also questions about what is good-prescribing and good medicine-taking. 7 There are therefore two agendas: an empirical agenda to address the 8 question of how adherence might be improved and a normative agenda to tell 9 us what the right thing to do is. 10 The research recommendations from this guideline are for research 11 programmes which are described below under the themes of A: Shared 12 decision making and the consultation; B: Barriers and interventions to 13 adherence and C: Groups for special consideration 14 15 A. Shared decision making and the consultation 16 Research questions: 17 a. 18 making, partnership and other models of consultation? 19 b. 20 practices that approximate to joint decision-making look like? What are the 21 strengths and weaknesses of such practices seen from the vantage point of 22 various stakeholders (e.g. prescribers, patients, funders)? 23 c. 24 that could re-distribute accountability more from prescribers to patients (e.g. 25 waivers)? What are the legal, policy, practical and psychological implications 26 of trying to share accountability differently? 27 d. 28 and emotional challenges of prescribing consultations designed to promote 29 informed choice and adherence to medication? What are prescribers’ and patients’ attitudes to shared decision What joint decision-making processes are possible? What do real world Are there practical mechanisms in place (or available to put in place) How can clinicians and patients be supported to deal with the cognitive Medicines concordance: full guideline DRAFT (July 2008) page 31 of 373 DRAFT FOR CONSULTATION 1 e. How can we facilitate the honest disclosure of medication-taking 2 behaviours within prescribing-related consultations and medication use 3 reviews? How can we equip health practitioners to respond appropriately and 4 effectively? 5 f. 6 risk of nonadherence or who are a priority for medication-review and 7 adherence support and how can we provide it. 8 Why this is important 9 The principles of shared decision-making have largely been developed from How can we enable new and existing prescribers to identify patients at 10 theoretical and conceptual models. The competencies listed for shared 11 decision- making consist of a number of different skills and patients have 12 shown that they may value different aspects of shared decision making. While 13 the right of patients to be involved in decision-making in regard to their own 14 health care is accepted, the practice of shared decision making may result in 15 practitioners and patients playing different roles than they have to date in 16 health care consultations. This may have implications for responsibility and 17 accountability. Information asymmetries also need to be addressed and this 18 may require structural changes to health services and their delivery. Patient 19 related outcomes need to be included. There is a new and growing agenda 20 relating to non-medical prescribers (pharmacists, nurses etc.) This is a key 21 context issue and there are a range of questions relating to patient 22 perspectives on new prescribers and to new prescribers’ perceptions and 23 skills. The effects of new prescribers on patient adherence to medication 24 should be included in any research agendas designed to evaluate new 25 prescribers. 26 27 B. Barriers and interventions to adherence 28 Research Questions 29 a) What are the most clinical and cost effective methods for addressing 30 cognitive influences (such as beliefs and concerns about medication) and Medicines concordance: full guideline DRAFT (July 2008) page 32 of 373 DRAFT FOR CONSULTATION 1 capacity (such as memory, manual dexterity, changes in routine) that result in 2 reduced adherence? 3 Why this is important 4 Few interventions have been systematically developed, using appropriate 5 theoretical models, nor have they have been modelled and piloted with 6 assessment of process variables as well as outcomes (as recommended in 7 the MRC framework for complex interventions to effect behaviour change). 8 Consequently it is difficult to tell why some interventions work and others do 9 not. Interventions should be developed using an appropriate theoretical 10 framework with a phased approach to testing that includes assessment of 11 process (i.e. the things that are targeted for change) as well as outcomes and 12 a need for an individual approach. Research should explore both clinical and 13 economic benefits and include patient outcomes. 14 C. Groups for special consideration 15 a. 16 prescriptions and adherence to prescribed medication? 17 b. 18 (children, young adults, elderly people) influence adherence and what are the 19 implications for interventions? 20 c. 21 multiple pathologies (and their informal carers) and what interventions are 22 required? 23 Why this is important 24 The value an individual places on sharing decisions with their practitioner has 25 been found to differ by groups such as the elderly and severity of condition. 26 Research into the factors and impact on adherence could inform clinicians 27 and shape clinical care. Multi-morbidity is common and occurs at all ages but 28 increases with age. Patients with multi-morbidity are often prescribed large 29 amounts of medication. Research is required to assess particular barriers 30 experienced by people with multi-morbidity and to investigate what type of What are the effects of social disadvantage and ethnicity on accessing How do the perceptions and life circumstances of different age groups What are the particular barriers to medicines use for people with Medicines concordance: full guideline DRAFT (July 2008) page 33 of 373 DRAFT FOR CONSULTATION 1 support is required to allow people to make decisions about medicines and to 2 adhere to those decisions. 3 4 5 6 1.7 Acknowledgements 7 **TO BE ADDED for final version 8 1.8 Glossary Adherence Adherence – ‘the extent to which the patient’s behaviour matches agreed recommendations from the prescriber’. Adherence emphasises the need for agreement and that the patient is free to decide whether or not to adhere to the doctor’s recommendation.. Compliance Compliance – ‘the extent to which the patient’s behaviour matches the prescribers recommendations’. Concordance Concordance – this is a recent term whose meaning has changed. It was initially applied to the consultation process in which doctor and patient agree therapeutic decisions that incorporate their respective views, but now includes patient support in medicine taking as well as prescribing communication. Concordance reflects social values but does not address medicine-taking and may not lead to improved adherence Medicines concordance: full guideline DRAFT (July 2008) page 34 of 373 DRAFT FOR CONSULTATION Cost-benefit A type of economic evaluation where both costs and analysis benefits of healthcare treatment are measured in the same monetary units. If benefits exceed costs, the evaluation would recommend providing the treatment. Cost- A type of economic evaluation where various health consequences outcomes are reported in addition to cost for each analysis intervention, but there is no overall measure of health gain. Cost- An economic study design in which consequences of effectiveness different interventions are measured using a single analysis outcome, usually in ‘natural’ units (for example, life-years gained, deaths avoided, heart attacks avoided, cases detected). Alternative interventions are then compared in terms of cost per unit of effectiveness. Cost- An explicit mathematical framework, which is used to effectiveness represent clinical decision problems and incorporate model evidence from a variety of sources in order to estimate the costs and health outcomes. See also Markov model. Cost- An economic evaluation that finds the least costly minimisation alternative therapy after the proposed interventions has analysis been demonstrated to be no worse than its main comparator(s) in terms of effectiveness and toxicity. Cost-utility A form of cost-effectiveness analysis in which the units of analysis effectiveness are quality-adjusted life-years (QALYs). Decision analysis A systematic way of reaching decisions, based on evidence from research. This evidence is translated into probabilities, and then into diagrams or decision trees which direct the clinician through a succession of possible scenarios, actions and outcomes. Medicines concordance: full guideline DRAFT (July 2008) page 35 of 373 DRAFT FOR CONSULTATION Decision problem A clear specification of the interventions, patient populations and outcome measures and perspective adopted in an evaluation, with an explicit justification, relating these to the decision which the analysis is to inform. Discounting Costs and benefits incurred today have a higher value than costs and benefits occurring in the future. Discounting health benefits reflects individual preference for benefits to be experienced in the present rather than the future. Discounting costs reflects individual preference for costs to be experienced in the future rather than the present. For NICE economic evaluations, health outcomes will be discounted at 3.5% and costs at 3.5% per annum, following the recommendations of the UK Treasury. Dispensing Professional trained in dispensing medicine, generally a professional pharmacist or a general practitioner in a dispensing practice Dominance An intervention is said to be dominant if there is an alternative intervention that is both less costly and more effective. See also extended dominance. Dosette box A type of compliance aid. Other terms used are NOMAD, MANRAX and monitored dose system. Economic Comparative analysis of alternative health strategies evaluation (interventions or programmes) in terms of both their costs and consequences. Medicines concordance: full guideline DRAFT (July 2008) page 36 of 373 DRAFT FOR CONSULTATION Extended An intervention is extendedly dominated when it can be dominance dominated by a combination of two alternative interventions (i.e. if x% of the population are treated with intervention A, and y% are treated with intervention C, the overall result will be an intervention strategy that is both cheaper and more effective than intervention B). See also dominance. Extrapolation In data analysis, predicting the value of a parameter outside the range of observed values. Forgiveness The ability of a drug to sustain its pharmacological action after a dose has been missed GDG Guideline development group who developed the guideline Health care Any health care professional- specialists, general professional practitioner, nurse prescribers who are involved in the (HCP) prescribing of medicines or in the discussion with patients about those medicines. Health The study of the allocation of scarce resources among economics alternative healthcare treatments. Health economists are concerned with both increasing the average level of health in the population and improving the distribution of healthcare resources. Health-related A combination of an individual’s physical, mental and quality of life social well-being; not merely the absence of disease. Informed Informed adherence refers to an outcome of informed adherence choice in decision to take medicines and supported adherence Medicines concordance: full guideline DRAFT (July 2008) page 37 of 373 DRAFT FOR CONSULTATION Incremental Cost Is the difference in costs between two interventions being effectiveness compared divided by the difference in effect of the two ratio interventions. For instance if A and B are being compared Cost of A – costs of B divided by effects of A- effects of B. ICER Incremental cost effectiveness ratio – this is the difference between the mean costs in the population of interest divided by the difference in the mean outcomes in the population of interest Life-year A measure of health outcome that shows the number of years of remaining life expectancy. Life-years gained Average years of life gained per person as a result o fan intervention. Markov model A modelling technique used when a greater number of health states needs to be considered. They are particularly useful for disease in which events can occur repeatedly over time. Medicines The term medicines is used in the guideline to apply to drug treatments that patients may take orally or selfadminister such as creams to the skin and drops. Medication A face to face meeting between a professional and a review patient to discuss the patients medicines and medicinetaking behaviour Opportunity cost .The opportunity cost of investing in a healthcare intervention is the other healthcare programmes that are displaced by its introduction. This may be best measured by the health benefits that could have been achieved had the money been spent on the next best alternative healthcare intervention. Medicines concordance: full guideline DRAFT (July 2008) page 38 of 373 DRAFT FOR CONSULTATION Persistence The length of time from initiation to discontinuation of therapy. Persistence is measured in units of time. Perspective (or viewpoint): This determines which costs to include. For NICE evaluations the perspective is from the NHS and includes costs to the NHS and Personal Social Services. Costs to other public bodies and to patients and carers may be considered as an additional factor. Probabilistic Probability distributions are assigned to the uncertain sensitivity parameters and are incorporated into evaluation models analysis based on decision analytical techniques (for example, Monte Carlo simulation). Quality adjusted An index of survival that is adjusted to account for the life-years person’s quality of life during this time. QALYs have the (QALYS) advantage of incorporating changes in both quantity (longevity/mortality) and quality (morbidity, psychological, functional, social and other factors) of life. Used to measure benefits in cost-utility analysis, QALYS are calculated by estimating the number of years of life gained from a treatment and weighting each year with a qualityof-life score between zero and one. Shared Decision .Shared-decision making (SDM) is described as a model Making (SDM) of decision making where information exchange is a two way process in the consultation and both deliberation and decision are made by both health care professional and patient. Specialist One who has expertise in a particular field of medicine by virtue of additional training and experience. Medicines concordance: full guideline DRAFT (July 2008) page 39 of 373 DRAFT FOR CONSULTATION Time horizon The time span used in the NICE appraisal that reflects the period over which the main differences between interventions in health effects and use of healthcare resources are expected to be experienced, and taking into account the limitations of supportive evidence. Unit-dose Unit-dose packaging is the packaging of a single dose in a packaging nonreusable container. Utility This concept is applied in health care to mean the individual's valuation of their state of well-being deriving from the use of health care interventions. In brief, utility is a measure of the preference for, or desirability of, a specific level of health status or specific health outcome. Willingness to WTP refers to the amount that a decision maker is willing pay (WTP) to pay for an additional unit of outcome (e.g. an additional QALY). If the WTP is higher than the ICER, the intervention is cost effective. If not, the intervention is not cost effective. 1 2 Medicines concordance: full guideline DRAFT (July 2008) page 40 of 373 DRAFT FOR CONSULTATION 1 2 Methods 2 2.1 Introduction 3 This chapter sets out in detail the methods used to generate the 4 recommendations for clinical practice that are presented in the subsequent 5 chapters of this guideline. The methods are in accordance with those set out 6 by the Institute in ‘The guidelines manual’. April 2006. London: National 7 Institute for Health and Clinical Excellence. Available from: 8 www.nice.org.uk/guidelinesmanual. The Guideline Development Process – an 9 overview for stakeholders, the public and the NHS describes how 10 organisations can become involved in the development of a guideline. 11 2.2 12 A series of key questions created from the scope was the first step in the 13 development of the guideline. The key questions formed the starting point for 14 the subsequent evidence reviews and facilitated the development of 15 recommendations by the GDG. 16 The key questions were developed by the project team with the guidance from 17 the GDG. Where possible, the questions were refined into specific research 18 questions by the project teams to aid literature searching, appraisal and 19 synthesis. However, due to the generic nature of the guideline, full PICO 20 parameters were not applicable to the developed research questions. The full 21 list of key questions is shown in appendix B. 22 A full literature search and critical appraisal could not be undertaken for all key 23 questions due to the time and resource limitations within the guideline 24 development process. The GDG and project teams therefore agreed 25 appropriate review parameters (inclusion and exclusion criteria) for each topic 26 area in accordance with those outlined in the scope. Developing key clinical questions (KCQs) 27 Medicines concordance: full guideline DRAFT (July 2008) page 41 of 373 DRAFT FOR CONSULTATION 1 2.3 Literature search strategy 2 An initial scoping search for published guidelines, systematic reviews, 3 economic evaluations and ongoing research was carried out on the following 4 databases or websites: National Library for Health (NLH) Guidelines Finder, 5 National Guidelines Clearinghouse, Scottish Intercollegiate Guidelines 6 Network (SIGN), Guidelines International Network (GIN), Canadian Medical 7 Association (CMA) Infobase (Canadian guidelines), National Health and 8 Medical Research Council (NHMRC) Clinical Practice Guidelines (Australian 9 Guidelines), New Zealand Guidelines Group, BMJ Clinical Evidence, 10 Cochrane Database of Systematic Reviews (CDSR), Database of Abstracts of 11 Reviews of Effects (DARE) and Heath Technology Assessment Database 12 (HTA), NHS Economic Evaluations Database (NHSEED) National Research 13 Register and Current Controlled Trials. 14 The aim of the evidence review was to identify the most relevant, published 15 evidence in relation to the key clinical questions generated by the GDG. Due 16 to time constraints, full systematic reviews were not undertaken. However, the 17 evidence reviews were undertaken using systematic, transparent approaches. 18 The following bibliographic databases were searched from their inception to 19 the latest date available: Cochrane Database of Systematic Reviews (CDSR), 20 Database of Abstracts of Reviews of Effects (DARE), Health Technology 21 Database (HTA), MEDLINE, EMBASE, CINAHL, AMED (Allied and 22 Complementary Medicine Database), CENTRAL (Cochrane Controlled Trials 23 Register). When appropriate to the question PsycINFO was also searched. 24 The search strategies were developed in MEDLINE and then adapted for 25 searching in other bibliographic databases. Systematic reviews and 26 randomized controlled trials were searched for using methodological search 27 filters designed to limit searches to these study designs. These were devised 28 by the Centre for Reviews and Dissemination and the Cochrane 29 Collaboration. The economic literature was identified by conducting searches 30 in NHS Economic Evaluations Database (NHSEED) and in MEDLINE and 31 EMBASE using an economics search strategy developed by ScHARR at the 32 University of Sheffield. Medicines concordance: full guideline DRAFT (July 2008) page 42 of 373 DRAFT FOR CONSULTATION 1 Databases of the results of the searches for each question or topic area were 2 created using the bibliographic management software Reference Manager. 3 The search strategies for all questions or topic areas developed for the 4 Medline database are detailed in appendix B. Details of all literature searches 5 for the evidence reviews are available from the NCC-PC. Further references 6 were also suggested by the GDG. 7 In line with the Scope, literature searches were undertaken to produce 8 evidence reviews on each of the following key topics: 9 10 Interventions to enhance adherence in the context of prescribed medication 11 Shared decision-making in the context of prescribed medication 12 Barriers to adherence and shared decision-making in the context 13 of prescribed medication. 14 Literature searches for the guideline were undertaken for the designated topic. 15 It was decided that individual literature searches for each clinical question 16 would result in a considerable amount of duplicated work, as the retrieved 17 evidence would potentially overlap from question to question. 18 Other more focused literature searches were undertaken as appropriate for 19 some of the key clinical questions. These were indicated by the GDG as key 20 clinical questions that due to their importance and possible impact on clinical 21 practice would require a focused literature search to ensure that no important 22 study would be missed out. These were: 23 24 25 26 27 28 29 30 What tools are available to help elicit patients information needs about medicines? . What tools are available to help elicit patients beliefs about medicines? How can a practitioner detect whether a patient agrees/disagrees with recommendation to take medicines? How can practitioners elicit patient’s preferences for involvement in decisions about medicines? Medicines concordance: full guideline DRAFT (July 2008) page 43 of 373 DRAFT FOR CONSULTATION Do interventions to increase patient involvement increase length 1 of the consultation 2 3 Does change in dosing regime affect adherence 4 Does drug formulation/packaging affect adherence? 5 What is the effect of prescription charges/costs on adherence to prescribed medication 6 7 How can practitioners assess adherence? 8 Do medication reviews increase adherence to prescribed medication? 9 10 11 12 The specific search strategy for each topic area varied and was agreed with 13 the project team (with input from the GDG as necessary). The review 14 parameters were agreed with the GDG and aimed to provide the best 15 available evidence. For further details on the methodology and 16 inclusion/exclusion criteria please see individual evidence reviews. 17 The literature on barriers to shared decision-making and medicine taking, 18 shared decision-making and adherence to medication is not well indexed, 19 therefore, despite the comprehensive and detailed searching, some trials that 20 met our criteria may have been missed. 21 Updated searches were conducted for references published during the course 22 of the guidance development and final update searches were carried out in 23 May and June 2008. Only those studies where recommendations needed 24 substantial revisions were added in detail. 25 In line with NICE Equality scheme additional searches of the literature were 26 undertaken to ensure that general searches had located all evidence relevant 27 to vulnerable groups in the United Kingdom3. 28 3 www.nice.org.uk/aboutnice/howwework/NICEEqualityScheme.jsp Medicines concordance: full guideline DRAFT (July 2008) page 44 of 373 DRAFT FOR CONSULTATION 1 2.4 Identifying the evidence 2 After the search of titles and abstracts was undertaken, full papers were 3 obtained if they appeared to address the key clinical question. The highest 4 level of evidence was sought. However, other types of quantitative evidence, 5 qualitative evidence and expert formal consensus results were used when 6 randomised control trials were not available. Only English language papers 7 were reviewed. Following a critical review of the full text paper, articles not 8 relevant to the subject in question were excluded. Studies that did not report 9 on relevant outcomes were also excluded. 10 The reasons for rejecting any paper ordered were recorded and details can be 11 seen in Appendix D. 12 2.5 13 From the papers retrieved, the Senior Health Services Research Fellow 14 (SHSRF) and the Health Service Research Fellow (HSRF) synthesized the 15 evidence for each question or questions into a narrative summary. These form 16 the basis of this guideline. Each study was critically appraised using the 17 Institute’s criteria for quality assessment and the information extracted for 18 included studies is given in Appendix C. The content and delivery of 19 interventions was poorly defined in many studies and it was difficult to decide 20 which studies should be included or excluded. The GDG advised on which 21 studies to include and exclude in these circumstances. Background papers, 22 for example those used to describe the concepts used in the guideline, were 23 referenced but not extracted. 24 2.5.1 25 26 When agreeing key clinical questions the GDG discussed the choice of 27 outcomes for each search. A variety of outcomes are currently found in 28 studies on shared decision-making but the outcomes primarily looked at were 29 patient preferences, identification of beliefs and patient agreement to the 30 decision.. Any additional information on factors which may have influenced the 31 study results and had an impact on the wider implementation of an Critical appraisal of the evidence Choice of outcomes Medicines concordance: full guideline DRAFT (July 2008) page 45 of 373 DRAFT FOR CONSULTATION 1 intervention, such as participants’ age, ethnicity or social status; dropout rates 2 and payments or rewards given to participants, were recorded in the evidence 3 tables considered by the GDG. The primary outcome measure for all the 4 evidence reviews on interventions to increase adherence was adherence. 5 Adherence levels was the outcome also for studies examining medication 6 review. 7 8 2.6 Health Economics methods 9 Economic evaluation provides a formal comparison of benefits and harms as 10 well as the costs of alternative health programmes. It helps to identify, 11 measure, value and compare costs and consequences of alternative 12 treatment options. These outcomes are usually synthesised in cost 13 effectiveness (CEA) or cost utility analysis (CUA), which reflect the principle of 14 opportunity costs. For example, if a particular treatment strategy were found to 15 yield little health gain relative to the resources used, then it could be 16 advantageous to re-deploy resources to other activities that yield greater 17 health gain. 18 19 To assess the cost-effectiveness of interventions to increase adherence 20 (Interventions to increase adherence), we conducted a comprehensive 21 systematic review of the economic literature relating to medicines and 22 nonadherence. 23 24 In accordance with the NICE social value judgement the primary criteria 25 applied for an intervention to be considered cost effective were either: 26 27 a) The intervention dominated other relevant strategies (that is it is both less 28 costly in terms of resource use and more clinically effective compared with the 29 other relevant alternative strategies); or 30 31 b) The intervention cost less than £20,000 per quality-adjusted life-year 32 (QALY) gained compared with the next best strategy (or usual care) Medicines concordance: full guideline DRAFT (July 2008) page 46 of 373 DRAFT FOR CONSULTATION 1 Health Economic evidence review methodology 2 2.6.1 3 The following information sources were searched: 4 • Medline (Ovid) (1966-June 2006) 5 • Embase (1980-June 2006) 6 • NHS Economic Evaluations Database (NHS EED) 7 • PsycINFO 8 • Cumulative Index to Nursing and Allied Health Literature (CINAHL) 9 10 The electronic search strategies were developed in Medline and adapted for 11 use with the other information databases. The clinical search strategy was 12 supplemented with economic search terms. Titles and abstracts retrieved 13 were subjected to an inclusion/exclusion criterion and relevant papers were 14 ordered. No criteria for study design were imposed a priori. In this way the 15 searches were not constrained to randomised controlled trials (RCTs) 16 containing formal economic evaluations. Papers included were: 17 • Full/partial economic evaluations 18 • Considered patients over 16 years of age 19 • Written in English, and reported health economic information that could 20 be generalised to UK. 21 22 The full papers were critically appraised by a health economist using a 23 standard validated checklist. A general descriptive overview of the studies, 24 their quality, and conclusions was presented and summarised in the form of a 25 narrative review. 26 Each study was categorized as one of the following: cost-effectiveness 27 analysis or cost utility analysis (i.e. cost-effectiveness analysis with 28 effectiveness measured in terms of QALYs or life year gained). Some studies 29 were categorised as ‘cost consequences analysis’ or ‘cost minimisation 30 analysis’. These studies did not provide an overall measure of health gain or 31 attempt to synthesise costs and benefits. Such studies were considered as 32 partial economic evaluations. Medicines concordance: full guideline DRAFT (July 2008) page 47 of 373 DRAFT FOR CONSULTATION Cost-effectiveness modelling methods 1 2.6.2 2 De novo modelling was considered for aspects of medicine taking. However, 3 due to heterogeneity in the population covered by this guideline this was not 4 possible. This is discussed in more detail in Chapter 10. 5 2.7 6 In preparation for each meeting, the narrative and extractions for the 7 questions being discussed were made available to the GDG one week before 8 the scheduled GDG meeting. These documents were available on a closed 9 intranet site and sent by post to those members who requested it. Forming recommendations 10 GDG members were expected to have read the narratives and extractions 11 before attending each meeting. The GDG discussed the evidence at the 12 meeting and agreed evidence statements and recommendations. Any 13 changes were made to the electronic version of the text on a laptop and 14 projected onto a screen until the GDG were satisfied with these. 15 All work from the meetings was posted on the closed intranet site following the 16 meeting as a matter of record and for referral by the GDG members. 17 2.8 18 The table of clinical questions in Appendix B indicates which questions were 19 searched. 20 In cases where evidence was sparse, the GDG derived the recommendations 21 via informal consensus methods, using extrapolated evidence where 22 appropriate. All details of how the recommendations were derived can be 23 seen in the ‘Evidence to recommendations’ section of each of the chapters. 24 2.9 25 The guideline has been developed in accordance with the Institute’s guideline 26 development process. This has included allowing registered stakeholders the 27 opportunity to comment on the scope of the guideline and the draft of the full 28 and short form guideline. In addition, the draft was reviewed by an 29 independent Guideline Review Panel (GRP) established by the Institute. Areas without evidence and consensus methodology Consultation Medicines concordance: full guideline DRAFT (July 2008) page 48 of 373 DRAFT FOR CONSULTATION 1 The comments made by the stakeholders, peer reviewers and the GRP were 2 collated and presented for consideration by the GDG. All comments were 3 considered systematically by the GDG and the development team recorded 4 the agreed responses. 5 6 7 2.10 Relationships between the guideline and other national guidance 8 2.10.1 National Service Frameworks 9 The National Service Framework for Older People (2001) makes specific 10 recommendations for medication review in older people. 11 2.10.2 Related NICE Guidance 12 This guideline differs from most NICE guidelines in that it is not condition 13 specific but makes recommendations on how to involve patients in decisions 14 about medicines. This guidance should be used in conjunction with condition 15 specific NICE guidance which makes recommendations on what treatments 16 are clinically and cost effective. NICE and the National Patient Safety Agency 17 (NPSA) have recently produced joint guidance on medicines reconciliation 18 when adult patients are admitted to hospital (www.NICE.org.uk/PSG001). 19 2.10.3 Other national guidance 20 In formulating recommendations consideration was given to 21 22 23 24 Report of the Committee on Safety of Medicines Working Group on Patient Information’ (2005) General Medical Council document, Consent: doctors and patients making decision together 25 http://www.gmc-uk.org/news/articles/Consent_guidance.pdf 26 Mental Capacity Act 2005 27 http://www.opsi.gov.uk/ACTS/acts2005/pdf/ukpga_20050009_en 28 .pdf 29 Disability Discrimination Act 1995 Medicines concordance: full guideline DRAFT (July 2008) page 49 of 373 DRAFT FOR CONSULTATION 1 http://www.opsi.gov.uk/acts/acts1995/plain/ukpga_19950050_en_1 2 3 Through review of published guidance, personal contact and commenting on 4 guideline scope, endeavours were made to ensure that boundaries between 5 guidance were clear and advice was consistent. 6 Medicines concordance: full guideline DRAFT (July 2008) page 50 of 373 DRAFT FOR CONSULTATION 1 3 Principles and concepts used in the development of the guideline 2 3 Clinical guidelines generally provide guidance on the management of clinical 4 conditions. To inform the recommendations evidence is sought regarding the 5 benefits and harms, as well as cost of treatments. This guideline seeks to 6 inform patient involvement in decisions about medicines across clinical areas 7 and as such is more interested in patient and health care professional 8 behaviour than evidence for individual treatments. The development of the 9 guideline required the Guideline Development Group (GDG) to engage with 10 topics more usually found in psychological and sociological literature as well 11 as philosophical and legal issues such as rights and duties of patients and 12 professionals. The GDG discussed these issues at length to develop a 13 working consensus which then allowed them to examine the literature and 14 develop recommendations. These discussions are included in the relevant 15 sections of the guideline where they informed recommendations. The GDG 16 also wished to see the principles they used to develop the guideline brought 17 together in one chapter. This chapter brings together those discussions from 18 different parts of the guideline. Some of this content is therefore repeated in 19 different chapters and this is indicated where appropriate by the use of double 20 asterisks at start and end of sections. **In the final version of the guideline we 21 will use hyperlinks to link these sections** 22 3.1 23 Patients’ rights to be involved in decisions about medicines 24 The prescribing of medicines to patients has become a central part of the 25 delivery of modern medical care. Studies and commentaries on medicine- 26 taking by patients have often emphasized the objective health and cost 27 impacts incurred when patients do not take medicine as prescribed (World 28 Health Organization, 2003). There is an often unstated and perhaps 29 unrecognised assumption that patients should take medicines as suggested 30 by medical professionals (Pound, P., Britten, N., Morgan, M. et al , 2005). 31 While objective health and cost impacts of patients’ behaviour in relation to 32 medicines is important, the right of patients to make decisions in regard to Medicines concordance: full guideline DRAFT (July 2008) page 51 of 373 DRAFT FOR CONSULTATION 1 their own health is accepted in modern practice. The approach taken by this 2 guideline is that patients have a right to be involved in decisions about 3 medicines to the extent that they wish and it is the role of health professionals 4 to facilitate and support patients in their involvement in decision-making and to 5 support patients in taking medicine if the decision has been to prescribe. 6 It is particularly important for people who are known to frequently experience 7 inequalities in health to have their right recognized to be effectively engaged 8 in decision-making e.g. those with learning disabilities, mental health 9 problems and people of black and ethnic minority ethnic origin. These 10 individuals must be afforded equal opportunities for healthier outcomes 11 through the effective provision of appropriate access and support. 12 Practitioners must be aware of their legal duties and responsibilities to make 13 ‘reasonable adjustments’ in line with the Disability Discrimination Act (2005). 14 15 16 3.2 What is meant by involving patients in decisions about medicines? 17 Early analysis of consultations between medical doctors and patients 18 indicated that consultations were primarily led by the health care professional. 19 Bain (1976) (Bain, DJG, 1976) a general practitioner, tape-recorded his own 20 consultations to examine what actually happened in consultations and found 21 that he talked at least as much as the patients did. Tuckett and colleagues 22 (1985) (Tuckett, D, Boutlon, M, Olson, C et al , 1985) found that doctors 23 frequently inhibited patients from asking questions and did little to encourage 24 patients to present their own view. Historically medical professionals have had 25 the dominant role in making treatment decisions. Charles and colleagues 26 (1999) (Charles, C., Gafni, A., and Whelan, T., 1999) outline a number of 27 assumptions on which this dominant role was based: a single best treatment 28 existed and physicians would be well versed in current clinical thinking; 29 physicians would apply this knowledge consistently to all patients and were in 30 best position to evaluate treatment decisions and tradeoffs; and because of 31 their professional concern for the welfare of their patients, physicians had a 32 legitimate interest in each treatment decision. Significant asymmetry between Medicines concordance: full guideline DRAFT (July 2008) page 52 of 373 DRAFT FOR CONSULTATION 1 professionals and patients also existed in terms of education, income and 2 status. The assumptions underlying the dominant role of the professional have 3 been increasingly challenged as both medicine and society changed (Charles, 4 C., Gafni, A., and Whelan, T., 1997). More treatments for conditions have 5 become available with complex risk – benefit analyses required. It was 6 recognised that it is the patient who has to live with the consequences of 7 these decisions and might be in a better position than the professional to 8 evaluate and weigh these. Medical practice has also shifted away from acute 9 care towards both chronic long term care and preventative care which 10 requires a long-term commitment to medicines taking and may require 11 frequent monitoring of medications and illness. The principles of informed 12 consent and informed choice in relation to treatment decisions are now 13 enshrined in law and there is an inherent implication in these that the patient 14 is making a decision in relation to their own healthcare. Previous asymmetries 15 between health professionals and patients in terms of education and access 16 to information have also lessened. 17 The concepts of shared-decision making and patient-centredness are part of 18 the response to the need to recognise the role of the patient in medical 19 encounters and decisions. In the literature shared-decision making (SDM) is 20 described as one model of decision making. In this model information 21 exchange is a two way process in the consultation and both deliberation and 22 decision are made by both health care professional and patient. This is in 23 contrast to a ‘paternalistic’ model where information is given to the patient and 24 deliberation and decision are made by the health care professional or an 25 ‘informed’ model where information is given to the patient and the patient 26 makes the deliberation and decision (Charles, C., Gafni, A., and Whelan, T., 27 1999). Intermediate models are also recognised where decision may be led by 28 the professional or handed to the professional following full sharing of 29 information between both parties. Patient-centredness is an approach to the 30 patient which encompasses the sharing of power and responsibility. Mead and 31 Bower (2000) (Mead, N and Power, P, 2000) have described patient- 32 centredness as having 5 dimensions (1) adopting the biopsychosocial 33 perspective (this means using biological, psychological and social Medicines concordance: full guideline DRAFT (July 2008) page 53 of 373 DRAFT FOR CONSULTATION 1 understandings of disease and illness experience); (2) understanding the 2 patient as person; (3) sharing power and responsibility between the doctor 3 and patient; (4) building a therapeutic alliance and (5) understanding the 4 doctor as a person. Despite the interest in these concepts and an agreement 5 that there is a moral value inherent in them, it is accepted by many working in 6 the area that the operationalisation of these concepts is still evolving (Elwyn, 7 G., Edwards, A., Mowle, S. et al , 2001), (Charles, C., Gafni, A., and Whelan, 8 T., 1997), (Duggan, P. S., Geller, G., Cooper, L. A. et al , 2006). The 9 difficulties relating to how to enact shared decision making include 10 overcoming asymmetry in knowledge and experience between patients and 11 health care professional, the difficulty in recognising a shared decision and 12 what this concept means in terms of responsibility of the clinician. 13 Current evidence exploring health care professionals views indicate that they 14 perceive difficulties in implementing SDM. A recent systematic review which 15 examined qualitative and quantitative research on health care professionals’ 16 views about implementing SDM identified several perceived barriers (Gravel 17 2006) (Gravel, Karine., Legare, F., and Graham, I. D., 2006). The studies 18 included in the systematic review were primarily qualitative and the vast 19 majority of the practitioners were medical practitioners. The three most 20 commonly perceived barriers were time constraints, lack of applicability to 21 patient characteristics and lack of applicability to the clinical situation. 22 Thompson (2007) (Thompson, A. G., 2007) describes the current literature as 23 being primarily derived from the perspective of professionals – policy makers, 24 academics or medical professionals and not from the perspective of patients. 25 Using interviews and focus groups he describes patients wish for involvement 26 to be dynamic with demand varying according to the need for health care, 27 personal characteristics of the patient and the patient- professional 28 relationship. Wish for involvement is higher in this model when illness is 29 chronic, trust in the professional is low and the patient is active. Demand in 30 lower when illness is acute and serious, the patient passive and trust in the 31 professional is high. Medicines concordance: full guideline DRAFT (July 2008) page 54 of 373 DRAFT FOR CONSULTATION 1 Surveys of patients indicate that when asked patients as a group do ask for 2 increased involvement in healthcare decisions. The Picker Institute (2007) 3 (Richards, C and Coulter, A, 2007) published a report examining patient 4 experience of the NHS between 2002 and 2006. Using the results of 26 5 national patient surveys they report patients’ information needs not being met 6 and patients wanting more involvement in health care decisions. In regard to 7 medicines, patients reported wanting more involvement in medication choices 8 than they currently receive. The surveys indicate that professionals were 9 giving less information about side effects over the time of the surveys and 10 patients felt that fewer decisions were shared decisions in 2006 than in 2004 11 and 2005. 12 The literature on decision-making first evolved in the context of life-threatening 13 diseases such as cancer (Charles, C., Gafni, A., and Whelan, T., 1997) and 14 included decisions for example as to whether or not to have surgery (Murray, 15 E., Charles, C., and Gafni, A., 2006).These were often one off decisions with 16 major consequences. In the case of medicines the initial involvement in the 17 decision to prescribe a medicine is necessarily followed by ongoing, often 18 daily decisions by patients to continue taking the medicine prescribed. 19 Involving patients in decisions to take medicines concerns not just the 20 decisions made within a consultation but also attention to the ongoing 21 decisions that patients about their medicines following the consultation with a 22 health care professional. 23 3.3 24 What are we trying to achieve in involving patients in decisions about medicines? 25 The outcome that we wish to see from patient involvement is an informed 26 choice by the patient in regard to their use of medicines. The term informed 27 adherence has been used to describe an outcome of informed choice and 28 supported adherence (Horne, R and Weinman, J., 2004). In this 29 understanding achieving a shared agreement is limited if the patient is then 30 not supported to implement their intentions to take the medication as 31 recommended. Similarly, stipulating unconditional and unquestioning 32 adherence to prescribers’ instructions as our goal is, in most cases, not Medicines concordance: full guideline DRAFT (July 2008) page 55 of 373 DRAFT FOR CONSULTATION 1 justified if the patient has not made an informed choice about taking the 2 medication. 3 In most cases the patient is free to decide whether to take the treatment or 4 not. However, the healthcare practitioner has a responsibility to help ensure 5 that the choice is an informed one. Informed patient choice, rather than 6 ‘compliance’ is the desired outcome of the discussion. If the patient decides to 7 accept the prescription, then the aim is to facilitate appropriate adherence to 8 the agreed recommendations for how it should be taken to maximise its 9 efficacy and safety for the individual and optimise benefits and reduce risk. 10 Facilitating informed choice involves more than the provision of information. 11 Informing should be an active process, which involves more than simply 12 presenting the evidence. It also entails eliciting the patient’s beliefs and 13 identifying whether pre-existing beliefs might act as a barrier to an accurate 14 interpretation of the evidence. If the interpretation of information is influenced 15 by misconceptions about the illness and treatment, then the patient’s choice 16 may not be ‘informed’. 17 18 3.4 Roles and responsibilities of patients and health care professionals 19 20 Concern has been expressed by practitioners that sharing decisions with 21 patients may conflict with their duty of care to patients or their legal or ethical 22 obligations (Stevenson, F. A., 2003). While the UK General Medical Council 23 (GMC) (2001)4 considered one of the key duties of a doctor to ‘respect the 24 rights of patients to be fully involved in decisions about their care’ for many 25 clinicians there is a legitimate area of concern or indeed of conflict between 26 respect for autonomy of the patient and the duty of beneficence when a 27 clinician feels uncomfortable about a patient’s wishes. The GMC (2008)5 has 28 recently updated guidance for doctors about patient consent which makes 29 explicit the right of competent patients to make decisions about their own 4 5 http://www.gmc-uk.org/publications/annual_reports/review_archive/report2002.pdf http://www.gmc-uk.org/publications/business_plans/Business_Plan_2008.pdf Medicines concordance: full guideline DRAFT (July 2008) page 56 of 373 DRAFT FOR CONSULTATION 1 healthcare. The guidance emphasises the need for doctors to maximize 2 opportunities and patients abilities to make decisions for themselves and that 3 doctors must respect competent patients’ decisions even if they do not agree 4 with them. Doctors do not have to provide a treatment to patients which they 5 believe is not in the patients’ interest but must under such circumstances 6 explain their reasons and other options, including a second opinion to the 7 patient. It remains important however for health care professionals to do their 8 utmost to ensure that patients’ choices are informed choices as outlined 9 above. 10 The Mental Capacity Act (2005) 6governs the making of decisions for people 11 who lack capacity in England and Wales. Under the Act doctors are advised 12 that they must work on the presumption that every adult patient has the 13 capacity to make decisions about their care, and to decide whether to agree 14 to, or refuse, an examination, investigation or treatment. A patient is regarded 15 as lacking capacity once it is clear that, having been given all appropriate help 16 and support, they cannot understand, retain, use or weigh up the information 17 needed to make that decision, or communicate their wishes7. 18 Doctors are advised that assumptions must not be made that a patient lacks 19 capacity to make a decision solely because of their age, disability, 20 appearance, behaviour, medical condition (including mental illness), their 21 beliefs, their apparent inability to communicate, or the fact that they make a 22 decision that health professionals disagree with8. 23 3.5 Understanding the influences on medicine taking by patients 24 25 If health care professionals are to facilitate patient involvement in decisions 26 about medicines it is helpful and necessary to understand how patients 27 approach the taking of medicines, in particular the ongoing appraisal of 28 medicines that continues after a consultation. Investigation into why patients 6 7 http://www.opsi.gov.uk/ACTS/acts2005/pdf/ukpga_20050009_en.pdf http://www.gmc-uk.org/publications/business_plans/Business_Plan_2008.pdf 8 http://www.gmc-uk.org/publications/business_plans/Business_Plan_2008.pdf Medicines concordance: full guideline DRAFT (July 2008) page 57 of 373 DRAFT FOR CONSULTATION 1 do not take medicines as prescribed indicates that the decision to take 2 medicines and the continuing taking of medicines should be considered as a 3 complex behaviour. Fig 1 indicates a diagrammatic representation of current 4 evidence regarding factors and influences on medicine taking by patients. As 5 the figure shows there are a number of influences on patients. Many of these 6 factors interact and the arrows indicate the dynamic nature of the process. 7 Internal factors represent the beliefs and experiences of the patient. These 8 include the patient’s beliefs about their symptoms or disease, their beliefs 9 about medicines in general and their own reaction to medicines. These will 10 influence their intention to take a medicine as suggested by a health care 11 professional. Even when patients intend to take a medicine they can 12 experience difficulty in doing so because of practical problems such as 13 packaging or complexity of regime or they may forget to take medicines. 14 Patients conduct their own appraisal of the medicine they are taking and 15 evaluate its effects against their own expectations of what the medicine will 16 achieve. This feeds into their beliefs about their medicines which in turn 17 influences their intention to take or not to take the prescribed medicines. 18 External factors are those that feed into the patient’s internal appraisal 19 process. These include communication with family and friends and the 20 communication they have with their health care professionals. Information 21 about medicines will be available from multiple sources including 22 documentation patients gets with their medicines, from the pharmacist or 23 dispenser or from any other health care professional the patients comes into 24 contact with. Patients may get information from other patients who have taken 25 the same medicine. Patients may be influenced by the attitude in society to a 26 particular medicine or medicines in general and information may be received 27 from media sources. 28 The research evidence indicates that patients’ decisions about medicines are 29 made within the patients’ own frames of reference and make sense within the 30 patients’ understanding. Patients however often do not disclose to the health 31 professional any reluctance to take medicines or disagreement with the 32 doctor’s recommendation of a prescription or their non-taking of medicines. Medicines concordance: full guideline DRAFT (July 2008) page 58 of 373 DRAFT FOR CONSULTATION 1 The onus is on the health professional to elicit and explore patients’ beliefs 2 and experiences and facilitate the patient making an informed choice about 3 whether or not to take a prescribed medicine. 4 When patients do not take medication as prescribed they may therefore be 5 doing this on an intentional basis i.e. they have made their own appraisal and 6 have decided to take the medicine in their own way or even not at all. This 7 may be done with full information about medicines, illness and consequences 8 of taking or not taking recommended medication. INTERNAL FACTORS APPRAISAL: Assessment and interpretation of outcomes relative to expectations: - Did it work? Is it worth continuing? Beliefs about prescribed medicine(s) Perceptions of illness Symptom interpretations Background beliefs about pharmaceuticals as a class of treatment INTENTION TO TAKE ADHERENCE/ NONADHERENCE MEDICATION Treatment preferences PRACTICAL ABILITIEScapacity and resources Perceptions of personal sensitivity to medicines Barriers include forgetting and regimen complexity EXTERNAL FACTORS eg -Information -Communication (Health care professionals, friends etc.) -Media -Cultural influences cultural factors healthcare policy social support 9 10 Figure 1……………………………….Horne, R. Concordance, Adherence and 11 Compliance in Medicine-Taking. Report for the National Co-ordinating Centre 12 for NHS Service and Delivery Organisation R& D (NCCSDO) (2005), pp 139. 13 In routine clinical terms the factors included in figure 1 that are barriers to 14 medicine-taking can be considered as either practical barriers or perceptual 15 barriers. Perceptual barriers are ways in which individual patients think about Medicines concordance: full guideline DRAFT (July 2008) page 59 of 373 DRAFT FOR CONSULTATION 1 their illness or condition and treatments both in general and specifically. 2 Practical barriers are those such as cost, memory or dexterity that affect 3 individuals ability to use the medication recommended to them. 4 3.6 5 The terminology used in the area of medicine-taking highlights the changing 6 understanding of medicine taking behaviour and the changing relationships 7 between health care professionals and patients. The terms compliance, 8 adherence and concordance are now often used interchangeably and 9 inappropriately to describe medicine taking by patients. The approach taken 10 by this guideline is to use terminology as recommended in the Report for the 11 National Co-ordinating Centre for NHS Service Delivery and Organisation R & 12 D on Concordance, Adherence and Compliance in Medicine Taking (2005) 13 (NCCSDO) (Horne, R, Weinman, J., Barber, N. et al , 2005). 14 Compliance has been the most commonly used term in relation to medicine- 15 taking and can be defined as ‘the extent to which the patients’ behaviour 16 matches the prescriber’s recommendations’ (Haynes, T. B., Taylor, D. W., and 17 Sackett, D. L., 1979). The term has been criticised as it is suggested it carries 18 an implicit assumption that it is the prescriber’s role to decide on the correct 19 medication and the patient has a passive role which is to take the medication 20 as he/she has been instructed. Non-compliance indicates a failure to follow 21 instructions and can be used as a pejorative term indicating a patient who is 22 unwilling to do as instructed by a prescriber who knows what the patient 23 needs. 24 Adherence is defined as ‘the extent to which the patient’s behaviour matches 25 agreed recommendations from the prescriber’. Adherence shifts the balance 26 between doctor and patient to suggest there should be agreement between 27 doctor and patient about the prescriber’s recommendation. 28 Concordance is less easily defined and its meaning has changed. It was 29 initially used to describe the consultation process by which agreement about 30 therapeutic decisions was reached by doctors and patients. It presumed an 31 exploration of patients’ views and their incorporation into the decision made. Terminology and structure of guideline Medicines concordance: full guideline DRAFT (July 2008) page 60 of 373 DRAFT FOR CONSULTATION 1 The term therefore addresses consultation processes and does not include 2 any aspects of medicine taking as do compliance and adherence. Some uses 3 of the term has included both communication and support for patients in 4 medicine taking. While the term concordance has been useful in drawing 5 attention to the need to engage patients in decisions there is currently no 6 agreed way in which one can judge whether a consultation has been 7 concordant. Concordance does reflect current social values where patients 8 are seen as active participants in their own healthcare but does not address 9 medicine-taking and may or may not lead to improved adherence. 10 We have chosen to discuss the consultation process regarding medicines 11 separately from actual medicine taking. We use the term ‘shared decision- 12 making about medicines’ to refer to the healthcare professional–patient/carer 13 consultation and the term adherence to describe patients’ medicine taking 14 behaviour. The guideline has looked separately at evidence about 15 interventions to increase patient involvement in the decision to take medicine 16 within the consultation and at evidence about interventions to support patients 17 in taking of medicines. While this division is useful when examining the 18 literature and making recommendations, the dynamic nature of medicine 19 taking must not be forgotten. Patients’ perceptions and beliefs will change 20 over time and the experience of taking a medicine will also influence patients’ 21 intentions to continue taking that medicine and others prescribed. Fig 1 22 provides a representation of the patient’s pathway. The patient comes to the 23 consultation with their own beliefs and experiences (see section 5). In the 24 meeting with the health professional the patient’s complaint is assessed and a 25 prescription may be recommended by the health professional. Within the 26 consultation the decision as to whether or not the patient leaves with a 27 prescription may be led completely by the doctor or negotiated between 28 doctor and patient.. If the patient leaves the consultation with a prescription 29 they may or may not take the prescription to be dispensed, and even if the 30 medicine is dispensed they may or may not take the medicine. 31 Medicines concordance: full guideline DRAFT (July 2008) page 61 of 373 DRAFT FOR CONSULTATION P-HP encounter Before consultation P HP P-D encounter Medicine-taking Takes medicine P P D P Medicine prescribed but not dispensed P-Patient Medicine dispensed but not taken HP Healthcare Professional D- Dispensing professional Arrows to indicate how patients attitudes are influenced by previous experiences of medicines and medicine taking 1 2 Fig.1 Simplified representation of patient pathway 3 3.7 4 Models of shared decision-making for use in clinical practice have been 5 developed. Towle (1997) (Towle, A., 1997) suggested steps for shared 6 decision making and these have been adapted by Elwyn and colleagues 7 (1999) (Elwyn, G., Edwards, A., Dwyn, R et al , 1999) following exploratory 8 research with general practitioner registrars. They suggest that population 9 surveys cannot predict preference of individual patients for involvement and 10 patients’ preferences for involvement may vary according to clinical situation 11 so the involvement of patients in decision-making has to be explicitly 12 addressed. Neither can patients consider their role in actual making of a Shared decision-making about medicines Medicines concordance: full guideline DRAFT (July 2008) page 62 of 373 DRAFT FOR CONSULTATION 1 decision until they have information about their options and the risks and 2 benefits of those options. The following stages are suggested by Elwyn and 3 colleagues to involve patients in healthcare decisions and by implication this 4 also describes the competencies required by practitioners to involve patients 5 (Elwyn 1999) (Elwyn, G., Edwards, A., Dwyn, R et al , 1999). 6 Implicit/explicit involvement of patients in decision-making process 7 Explore ideas/fears and expectations of problem and treatments 8 Portrayal of options 9 Identify preferred format and provide information 10 Checking process: understanding of information and reactions 11 Acceptance of process and role in decision-making 12 Make, discuss, or defer decisions 13 Arrange follow up 14 While these models have been developed when considering a variety of 15 decisions we have used this model to provide an outline structure for our 16 discussion and recommendations about sharing decisions about medicines in 17 consultations. While a treatment can never be understood without reference to 18 the underlying disease, illness or symptom it is beyond the scope of this 19 guideline to make recommendations about general communication and about 20 how diagnosis and prognosis should be explained to patients. These do 21 overlap with our recommendations on communication about medicines but we 22 have not examined these areas explicitly. 23 Although we have used the term SDM in this guideline and have used 24 literature relating to SDM the understanding of the Guideline Development 25 Group is that this is a process and that we are addressing the ‘sharing’ of 26 decisions rather than defining what a shared decision is. Edwards and Elwyn 27 (2006) (Edwards, A. and Elwyn, G., 2006) have suggested that it is the Medicines concordance: full guideline DRAFT (July 2008) page 63 of 373 DRAFT FOR CONSULTATION 1 process of involving patients in decisions that delivers benefits for patients 2 rather than attaching importance to defining who makes the decision. Given 3 the difficulties in defining precisely what a shared decision is we cannot 4 advocate that the outcome from this process has to be a ‘shared – decision’. 5 For many patients this may be the preferred outcome, other patients will 6 prefer to give the decision to the professionals and others to make their own 7 ‘informed’ decision. Cox (2007) (Cox, K., Britten, N., Hooper, R. et al , 2007) 8 in a study about prescribing of medicines in general practice, found that 39% 9 of patients wanted the G.P to share the decision, 28% wanting the G.P to be 10 main decision-maker, 17% wanted the GP to be the only decision-maker, 14% 11 preferred that the patient be the main decision-maker and 2% the only 12 decision-maker. Given the evidence that patient involvement in choices about 13 medicine is lower than desired by patients and that doctors are not good at 14 recognising which patients want involvement our recommendations aim to 15 make the process of involvement more explicit and to increase overall patient 16 involvement. 17 Medicines concordance: full guideline DRAFT (July 2008) page 64 of 373 DRAFT FOR CONSULTATION 1 2 4 making about medicines 3 4 Interventions to increase shared decision- 4.1 Recommendations *listed at the start of the guideline according to chapter 5 6 4.2 Introduction 7 Shared decision making can be broadly defined ‘as a decision-making 8 process jointly shared by patients and their health care provider’ (Legare 9 2008) (F Légaré, S Ratté D Stacey J Kryworuchko K Gravel L Turcot ID 10 Graham, 2008). As discussed in chapter 3 the concept of shared decision- 11 making evolved as a development from a predominantly paternalistic model of 12 professional and patient interactions. Makoul and Clayman (2006) (Makoul, 13 Gregory and Clayman, Marla L., 2006) found thait the most commonly 14 occurring themes in discussion of shared decision making were patient values 15 and preferences, options, partnership and patient participation, with 17 other 16 concepts also given considerable weight. In a review of communication about 17 medicines Cox (2004) (Cox, F, Stevenson, N, Britten, N et al , 2004) sets out 18 how patients and professionals need to have two way discussions in which 19 they exchange information and views about medicines. 20 ** This section is a repeat of section 3.4 from chapter 3 ** 21 **The outcome that we wish to see from patient involvement is an informed 22 choice by the patient in regard to their use of medicines. The term informed 23 adherence has been used to describe an outcome of informed choice and 24 supported adherence (Horne, R and Weinman, J., 2004). In this 25 understanding achieving a shared agreement is limited if the patient is then 26 not supported to implement their intentions to take the medication as 27 recommended. Similarly, stipulating unconditional and unquestioning 28 adherence to prescribers’ instructions as our goal is, in most cases, not 29 justified if the patient has not made an informed choice about taking the 30 medication. Medicines concordance: full guideline DRAFT (July 2008) page 65 of 373 DRAFT FOR CONSULTATION 1 In most cases the patient is free to decide whether to take the treatment or 2 not. However, the healthcare practitioner has a responsibility to help ensure 3 that the choice is an informed one. Informed patient choice, rather than 4 ‘compliance’ is the desired outcome of the discussion. If the patient decides to 5 accept the prescription, then the aim is to facilitate appropriate adherence to 6 the agreed recommendations for how it should be taken to maximise its 7 efficacy and safety for the individual and optimise benefits and reduce risk. 8 Facilitating informed choice involves more than the provision of information. 9 Informing should be an active process, which involves more than simply 10 presenting the evidence. It also entails eliciting the patient’s beliefs and 11 identifying whether pre-existing beliefs might act as a barrier to an accurate 12 interpretation of the evidence. If the interpretation of information is influenced 13 by misconceptions about the illness and treatment, then the patient’s choice 14 may not be ‘informed’.** 15 This chapter outlines research evidence, the deliberations of the Guideline 16 Development Group and their recommendations in relation to patient 17 involvement in decision-making about medicines. The chapter is organised 18 around the key clinical questions in this area. These explored the process of 19 patient involvement . Involving patients in decision about medicines involves 20 attention to the process of patient involvement but is also informed by what is 21 known about patient views and experiences of medicines and medicine- 22 taking. Research studies have explored patients’ views about medicines and 23 although this information is not a substitute for discussion with individual 24 patients, it can sensitise health care professionals to how patients’ approach 25 medicine taking. The evidence about patient experience is presented 26 separately in chapter 5 for clarity but there is inevitable overlap between 27 recommendations developed from these separate evidence reviews. 28 4.3 29 ** This section is a repeat of section 3.7 from chapter 3 ** 30 **Models of shared decision-making for use in clinical practice have been 31 developed. Towle (1997) (Towle, A., 1997) suggested steps for shared Process of shared decision-making Medicines concordance: full guideline DRAFT (July 2008) page 66 of 373 DRAFT FOR CONSULTATION 1 decision making and these have been adapted by Elwyn and colleagues 2 (1999) (Elwyn, G., Edwards, A., Dwyn, R et al , 1999) following exploratory 3 research with general practitioner registrars. They suggest that population 4 surveys cannot predict preference of individual patients for involvement and 5 patients’ preferences for involvement may vary according to clinical situation 6 so the involvement of patients in decision-making has to be explicitly 7 addressed. Neither can patients consider their role in actual making of a 8 decision until they have information about their options and the risks and 9 benefits of those options. The following stages are suggested by Elwyn and 10 colleagues to involve patients in healthcare decisions and by implication this 11 also describes the competencies required by practitioners to involve patients 12 (Elwyn 1999) (Elwyn, G., Edwards, A., Dwyn, R et al , 1999). 13 14 Implicit/explicit involvement of patients in decision-making process 15 Explore ideas/fears and expectations of problem and treatments 16 Portrayal of options 17 Identify preferred format and provide information 18 Checking process: understanding of information and reactions 19 Acceptance of process and role in decision-making 20 Make, discuss, or defer decisions 21 Arrange follow up 22 While these models have been developed when considering a variety of 23 decisions we have used this model to provide an outline structure for our 24 discussion and recommendations about sharing decisions about medicines in 25 consultations. While a treatment can never be understood without reference to 26 the underlying disease, illness or symptom it is beyond the scope of this 27 guideline to make recommendations about general communication and about 28 how diagnosis and prognosis should be explained to patients. These do Medicines concordance: full guideline DRAFT (July 2008) page 67 of 373 DRAFT FOR CONSULTATION 1 overlap with our recommendations on communication about medicines but we 2 have not examined these areas explicitly. 3 Although we have used the term SDM in this guideline and have used 4 literature relating to SDM the understanding of the Guideline Development 5 Group is that this is a process and that we are addressing the ‘sharing’ of 6 decisions rather than defining what a shared decision is. Edwards and Elwyn 7 (2006) (Edwards, A. and Elwyn, G., 2006) have suggested that it is the 8 process of involving patients in decisions that delivers benefits for patients 9 rather than attaching importance to defining who makes the decision. Given 10 the difficulties in defining precisely what a shared decision is we cannot 11 advocate that the outcome from this process has to be a ‘shared – decision’. 12 For many patients this may be the preferred outcome, other patients will 13 prefer to give the decision to the professionals and others to make their own 14 ‘informed’ decision. Cox (2007) (Cox, K., Britten, N., Hooper, R. et al , 2007) 15 in a study about prescribing of medicines in general practice, found that 39% 16 of patients wanted the G.P to share the decision, 28% wanting the G.P to be 17 main decision-maker, 17% wanted the GP to be the only decision-maker, 14% 18 preferred that the patient be the main decision-maker and 2% the only 19 decision-maker. Given the evidence that patient involvement in choices about 20 medicine is lower than desired by patients and that doctors are not good at 21 recognising which patients want involvement our recommendations aim to 22 make the process of involvement more explicit and to increase overall patient 23 involvement.** 24 The GDG were first interested in whether interventions can improve patient 25 involvement in decision-making. The evidence in this area comes from a 26 literature which examines overlapping areas of patient communication, 27 patient-centeredness and shared decision-making. The outcomes from these 28 studies include a variety of patient –orientated outcomes such as knowledge 29 and satisfaction. A scale for measuring patient involvement, the OPTION 30 scale, has been developed and validated in the UK (Elwyn, G., Edwards, A., 31 Wensing, M. et al , 2003) (Elwyn, G., Edwards, A., Hutchings, A. et al , 2005). 32 The scale has not been used widely and recent comparison of this scale with Medicines concordance: full guideline DRAFT (July 2008) page 68 of 373 DRAFT FOR CONSULTATION 1 an instrument developed in the US called the Informed Decision Making 2 instrument showed poor overall agreement although good agreement on 3 elements that appeared conceptually similar (Weiss 2008) (Horne, R and 4 Weinman, J., 2004). 5 When exploring the process of shared decision-making on more detail we 6 used the stages and competences outlined by Elwyn (1999) (Elwyn, G., 7 Edwards, A., Dwyn, R et al , 1999) to structure the evidence. 8 For the purposes of this review the steps of the process of most interest to the 9 GDG were : 10 1. the elicitation of patients preferences for involvement in decision – making, 11 12 2. .the exploration of patient beliefs, 13 3. the identification of preferred format and provision of information, 14 4. .support for the patient in decision-making 15 5. .the elicitation of patient agreement 16 4.4 Methods 17 Searches were conducted to gather the most relevant evidence relating to the 18 key clinical questions on shared decision-making. Due to time constraints, 19 exhaustive systematic reviews were not conducted (see Cochrane Review 20 Methods). Initially an overall search of relevant Randomised Controlled Trials 21 (RCTs) and Systematic Reviews was conducted using shared decision- 22 making terms. The articles retrieved were then separated under the relevant 23 key clinical questions. However, this did not answer all the Key Clinical 24 Questions succinctly so further searches were done using lower grades of 25 study design. 26 After the key terms were searched we generated a list of abstracts which were 27 sifted to find relevant articles. Full articles of those deemed to have relevance 28 to the question were obtained. These were then further assessed as to their Medicines concordance: full guideline DRAFT (July 2008) page 69 of 373 DRAFT FOR CONSULTATION 1 match with our inclusion and exclusion criteria for the question and either 2 included or excluded. We were specifically looking for those studies regarding 3 medicine-taking, which ruled out a lot of studies from the oncology and 4 surgical fields. 5 We extracted the evidence from each included article for study quality, then 6 brought together the results into an evidence review for each key clinical 7 question. The evidence reviews were structured into explanatory narratives for 8 each article. These were then combined to provide evidence statements of the 9 overall evidence. It proved difficult to separate out the content of some of the 10 interventions contained in systematic reviews e.g. there is overlap in 11 interventions which explore improving communication between patient and 12 practitioner and those exploring how information should be presented and 13 discussed between practitioner and patient. We have therefore included some 14 studies in more than one section and have indicated this where appropriate. 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Medicines concordance: full guideline DRAFT (July 2008) page 70 of 373 DRAFT FOR CONSULTATION 1 2 4.5 Key Clinical Question: Is it possible to increase patient involvement in decisions about medicines? 3 Related References Evidence Statements (summary Evidence into of evidence) recommendations Patient-centred communication in the consultation Rao (2007) (Rao, 1. One systematic review of Jaya K., 2007) interventions to improve The GDG were interested communication behaviours of in what evidence was physicians and patients found that available to indicate that the interventions studied improved patient involvement in the patient-centred communication decisions about behaviours of physicians and medicines can be residents significantly. There were increased. The evidence modest effects for the interventions in this area is complex to improve patients’ communication and the GDG requested behaviours. further searches to look at process of patient Rao (2007) (Rao, 2. In one systematic review there involvement in more Jaya K., 2007) was a mix of significant and non detail. The GDG were significant results, depending on convinced however that the communication behaviour the systematic reviews studied – e.g. 5/7 found significant indicated that practitioner changes in patients’ skills could be improved communication patterns. All of and that these these included skills practice as improvements as well as part of the intervention, the other 2 some patient studies were low intensity; interventions did result in Rao (2007) (Rao, 3. One systematic review increased patient Medicines concordance: full guideline DRAFT (July 2008) page 71 of 373 DRAFT FOR CONSULTATION Jaya K., 2007) investigated intensity of involvement. The GDG interventions and found that higher considered it important intervention intensity was clearly that practitioners were related to improvements in aware that skills could be physician communications, improved through further whereas this was less pronounced training. for patient communication. Lewin (2001) (Lewin, 4. One systematic review found S. A., Skea, Z. C., that some interventions to promote Entwistle, V. et al , patient-centred care in 2001) consultations may lead to significant increases in the patient centredness of consultation processes. They concluded that there is limited and mixed evidence on the effects on health care behaviours or health status Patient involvement in the consultation Rao (2007) (Rao, 5. Three systematic reviews found Jaya K., 2007); mixed results as to whether or not (Harrington (2004) interventions increased patient (Harrington-Jane, involvement in the consultation Noble Lorraine, (described below). One RCT study 2004); Kinnersley found an increase in patient (2007) (Kinnersley, participation. P., Edwards, A., Hood, K. et al , 2007); Loh (2007) (Loh, Andreas, Medicines concordance: full guideline DRAFT (July 2008) page 72 of 373 DRAFT FOR CONSULTATION Simon, Daniela, Wills, Celia E. et al , 2007) Harrington (2004) 6. In one systematic review, 10 (Harrington-Jane, RCTs found a significant increase Noble Lorraine, in participation, while 5 RCTs found 2004) a non-significant increase. Kinnersley (2007) 7. In one systematic review patient (Kinnersley, P., participation was increased in 8 out Edwards, A., Hood, of 14 RCT studies and no effect K. et al , 2007) found in 5 RCT studies. Wetzels (2007) 8. One systematic review found (Wetzels, R., limited evidence of interventions Harmsen, M., van, aimed specifically at improving Weel C. et al , 2007) older patients’ involvement in consultations, they found only three RCTs. Wetzels (2007) 9. Those RCT studies that did (Wetzels, R., investigate older patients’ Harmsen, M., van, involvement found that the Weel C. et al , 2007) interventions resulted in the patients asking more and different questions, and so they may become more active in consultations due to pre-visit preparation. Medicines concordance: full guideline DRAFT (July 2008) page 73 of 373 DRAFT FOR CONSULTATION Type of behaviour evoked Harrington (2004) 10. Question-asking was found (Harrington-Jane, equal in significant and non- Noble Lorraine, significant trends in one systematic 2004); review (Harrington 2004). In another systematic review Kinnersley (2007) (Kinnersley, P., Edwards, A., Hood, K. et al , 2007); Wetzels (2007) (Wetzels, R., Harmsen, M., van, Weel C. et al , 2007) (Kinnersley 2007) 6 out of 17 studies found significant increases and the other 11 studies found no significant effects. Another systematic review (Wetzels 2007) found one study where more asked questions and one study where few prepared questions. Harrington (2004) 11. In one systematic review there (Harrington-Jane, was a significant increase in Noble Lorraine, clarifying issues. 2004) Harrington (2004) 12. One systematic review found a (Harrington-Jane, significant increase in patients’ Noble Lorraine, perception of control over health 2004) and for an active role in health care, recall of information, adherence to recommendations, attendance and clinical outcomes. Harrington (2004) 13. One systematic review (Harrington-Jane, (Harrington 2004) found a Noble Lorraine, significant increase in satisfaction Medicines concordance: full guideline DRAFT (July 2008) page 74 of 373 DRAFT FOR CONSULTATION 2004); Kinnersley for only two RCTs however there (2007) (Kinnersley, was high satisfaction found overall. P., Edwards, A., In one systematic review Hood, K. et al , (Kinnersley 2007) 14 out of 23 2007); Loh (2007) studies showed no change and 5 (Loh, Andreas, had increased satisfaction. In Simon, Daniela, another RCT (Loh 2007) Patient Wills, Celia E. et al , satisfaction was significantly higher 2007). in the intervention 29.8 (2.7) than the control group 27.0 (3.6), p=0.14. Type of intervention Rao (2007) (Rao, 14. In four systematic reviews most Jaya K., 2007); of the interventions were written, Harrington (2004) e.g. booklet/checklists, some were (Harrington-Jane, videotapes or face to face Noble Lorraine, coaching. 2004); Kinnersley (2007) (Kinnersley, P., Edwards, A., Hood, K. et al , 2007); Wetzels (2007) (Wetzels, R., Harmsen, M., van, Weel C. et al , 2007) Harrington (2004) 15. One systematic review found (Harrington-Jane, that face-to-face and video Noble Lorraine, interventions showed more 2004); Kinnersley significant results than written (2007) (Kinnersley, interventions (Harrington 2007). Medicines concordance: full guideline DRAFT (July 2008) page 75 of 373 DRAFT FOR CONSULTATION P., Edwards, A., Another systematic review found a Hood, K. et al , small to moderate significant 2007) increase for writing alone and coaching. Wetzels (2005) 16. One RCT of a consultation (Wetzels, R., leaflet (including open and pre- Harmsen, M., van, structured questions) to improve Weel C. et al , 2007) involvement of older patients, showed an increase in satisfaction but no effect of the leaflet on involvement, enablement or satisfaction. However there was a significant result of reporting more psychological symptoms to their GPs. Little (2004) (Little, 17. One RCT found that a general P., Dorward, M., leaflet which asked patients to list Warner, G. et al , issues they wanted to raise and 2004) knowledge that the doctors expected questions led to a significant increase in satisfaction but no other aspects. The leaflet was significantly more effective when consultations were short. The leaflet increased the number of investigations by the doctor. Wilkinson (2002) 18. One RCT of an education (Wilkinson, C. R. and guidebook to encourage and Williams, M., 2002) support participation found no significant differences in the effectiveness of their primary care Medicines concordance: full guideline DRAFT (July 2008) page 76 of 373 DRAFT FOR CONSULTATION visit. There were significant differences for those receiving preventative health care interventions. Ross (2004) (Ross, 19. One RCT found the use of a S. E., Moore, L. A., patient accessible online medical Earnest, M. A. et al , record found no significant results 2004) for self-efficacy, adherence, health status or patient satisfaction. General adherence to medication advice was significantly improved (no data given). Loh (2007) (Loh, 20. One RCT found no significant Andreas, Simon, differences in treatment adherence. Daniela, Wills, Celia E. et al , 2007) 1 Methods of the evidence review 2 4.5.1 3 This paper includes a narrative summary of the included evidence, structured 4 according to the category of the intervention, following the agreed reviewing 5 protocol: 6 Types of studies: Systematic Reviews of Randomised Controlled Trials 7 (RCTs) or Randomised Controlled Trials of interventions involving shared 8 decision-making in the clinical context. 9 Types of participants: people prescribed medication for a medical condition. 10 Duration of studies: no time limit specified. Medicines concordance: full guideline DRAFT (July 2008) page 77 of 373 DRAFT FOR CONSULTATION 1 Types of interventions: any interventions involving shared decision making 2 in a consultation between a health care professional and patient. 3 Types of outcome measures: Patient-centred communication in the 4 consultation; Consultation process outcomes: patient involvement, question 5 asking, preparedness; Patient care outcomes: satisfaction, knowledge, self- 6 efficacy. Type of interventions involved and type of information. 7 8 It should be noted that the remit is for conditions with prescribed medication 9 and this excludes conditions which require chemotherapy or screening. All 10 RCTs are within this remit, however many of the systematic reviews included 11 populations outside the remit, this is noted where applicable. 12 Evidence review 13 4.5.2 14 The narratives for this question are not grouped because they cover various 15 aspects of the question. The evidence review has been constructed under the 16 above subheadings of patient-centred communication, patient involvement, 17 type of behaviour, type of intervention and type of information. It should also 18 be noted that RCTs which were included in the Systematic Review were not 19 narrated separately so as not to duplicate results. 20 Rao (2007) (Rao, Jaya K., 2007) conducted a systematic review of 21 interventions to enhance communication behaviours among physicians and 22 patients in an outpatient setting. They included RCTS of interventions which 23 are designed to improve the communication behaviours of physicians and 24 patients. The primary outcome was an assessment of the patient-centred 25 verbal communication behaviours. They rated the intensity of the interventions 26 based on how often it was delivered and the personnel involved in the 27 delivery. Thirty six RCTs met the inclusion criteria, 18 trained physicians or 28 residents and 15 trained patients, 3 intervened on both. The majority (26 29 RCTs) were conducted in primary care clinics while 10 were conducted in 30 medical specialty settings. Sixteen of the studies were outside of the USA. 31 Half the studies participants were assessed before and after the physician- 32 patient interaction. Most of the studies were conducted on audiotapes or Medicines concordance: full guideline DRAFT (July 2008) page 78 of 373 DRAFT FOR CONSULTATION 1 videotapes and coders were blinded to group assignment. The 2 physician/resident interventions mainly involved a variety of types (e.g. 3 information, feedback, modelling, practice). Nearly all included written 4 instructional information. Some (10 interventions) included videotapes 5 modelling desirable communication behaviours. Most RCTs showed 6 significant improvements in communication behaviours of the 7 physicians/residents. The higher intensity interventions resulted in physicians 8 asking more open-ended questions and less biomedical questions than the 9 comparison groups. They were more likely to elicit the patients’ concerns they 10 had prior to the visit, and show a more patient-centred communication style 11 overall. There were significant improvements in their information provision to 12 patients. Rao noted that few RCTs (6) checked for patients’ understanding of 13 the information received. 18 studies intervened with patients and showed 14 modest effects. Most of the RCTs were informational (17), often written, some 15 instances the written information included models of desirable communication 16 with examples of questions to ask (8), eight included practicing or coaching 17 sessions and four were feedback. Six were moderately-intense, 2 were rated 18 high-intensity. Seventeen of the RCTs used different measures of patient 19 involvement. 7 assessed the degree to which patients spoke during the visit, 5 20 (moderate intensity interventions) of which showed significant changes in 21 communication patterns. All of these were skills practice interventions. The 22 other two RCTs were low-intensity and showed no significant results. To 23 conclude, the interventions enhanced communication behaviours with 24 physicians compared to controls and there were modest effects with the 25 patient interventions. Intervention intensity had a clear relationship to 26 improving the physicians’ behaviours but this was not so much for patients. 27 Harrington (2004) (Harrington-Jane, Noble Lorraine, 2004) conducted a 28 systematic review which looked at interventions to increase patients’ 29 participation in medical consultations. Therefore there were a mix of 30 populations including primary care and outpatient oncology. The inclusion 31 criteria were interventions designed to improve patients’ communication with 32 doctors in any setting; reporting data on the impact of the intervention on the 33 patients’ communication. Most of the studies were RCTs, with the others Medicines concordance: full guideline DRAFT (July 2008) page 79 of 373 DRAFT FOR CONSULTATION 1 being Quasi-experimental. Twenty five papers were retrieved from the search. 2 Most of the studies were from the USA, 2 from Australia, 5 from the UK and 3 one from the Netherlands. Most of the interventions were written followed by 4 face-to-face coaching and videotape. Written interventions were in booklet or 5 checklist form. The specific behaviours most encouraged were question- 6 asking, raising concerns and requesting clarification or checking 7 understanding. The process of communication was measured mostly using 8 interaction analysis from audio recordings. Researchers coding the 9 interactions were blinded in only six of the studies. Overall the effect of 10 interventions was that they encouraged patients to be more active in their 11 consultations. Of the 16 studies examining variables of participation, 10 12 reported a significant increase and five reported a non-significant increase. All 13 but one of the six face-to-face interventions and all of the video interventions 14 reported significant results in increasing participation. Of the 10 written 15 interventions, only two reported a significant increase. Question-asking was 16 equal in significant increases and non-significant trends. There was a 17 significant increase in requesting clarification on matters. There was no 18 significant increase in satisfaction due to the interventions apart from in two 19 studies. There was overall a high level of satisfaction reported. There was a 20 significant increase in perception of control over health and preferences for an 21 active role in health care, recall of information, adherence to 22 recommendations, attendance and clinical outcomes. 23 It should be noted that Lewis, Butow, Ford, Street and Brown were studies 24 with cancer patients exclusively. 25 Kinnersley (2007) (Kinnersley, P., Edwards, A., Hood, K. et al , 2007) 26 conducted a Cochrane Collaboration systematic review of the effects of 27 interventions before consultations designed to help patients address their 28 information needs. The settings and populations varied, and were from six 29 countries, mainly the USA. They stated that it complemented other Cochrane 30 reviews by Wetzels (2007) (Wetzels, R., Harmsen, M., van, Weel C. et al , 31 2007) and Lewin (2003). The inclusion criteria were RCTs of interventions 32 intended to help the patients, representatives or carers address their Medicines concordance: full guideline DRAFT (July 2008) page 80 of 373 DRAFT FOR CONSULTATION 1 information needs in a consultation. This was done by encouraging question 2 asking, to express their information needs, to overcome barriers to 3 communication and to clarify their understanding of the information provided. 4 Outcome measures were the consultation process, consultation outcomes 5 and service outcomes. 33 trials described in 35 studies met the inclusion 6 criteria. From the studies assessing single interventions for patients 15 7 included written materials, four were coaching. The multiple component single 8 interventions studies 4 had coaching and written materials. Results: 17 9 studies measure question asking with 6 finding statistically significant 10 increases and 11 studies finding no effects of the interventions compared to 11 controls. Patient participation was measured in 14 studies, it was increased in 12 8, and no effect in five studies. Patient satisfaction was measured in 23 13 studies, in 14 studies there were no changes and in 5 there was increased 14 satisfaction. Patient knowledge was measured in 5 studies with a reduction in 15 two studies and no changes in 3 studies. According to type of intervention, 16 comparisons between written alone and coaching alone showed similar, small 17 to moderate and statistically significant increases for both types for question 18 asking. Patient satisfaction was borderline statistically significant for written 19 materials, coaching the effect was small and statistically significant. Written 20 materials led to a small and statistically significant increase in consultation 21 length with coaching the increase was smaller but not significant. 22 Interventions immediately before the consultation led to a small statistically 23 significant increase in consultation length and patient satisfaction. 24 It should be noted that many of the studies were from other settings: Brown 25 (1999, 2001), Bruera (2003), Butow (1994, 2004), Davison (1997, 2002), Ford 26 (1995), Oliver (2001) were cancer studies. Finney (1990), Kim (2003), Lewis 27 (1991) were from paediatric and family planning settings. 28 Wetzels (2007) (Wetzels, R., Harmsen, M., van, Weel C. et al , 2007) 29 conducted a Cochrane systematic review assessing the effects of 30 interventions in primary care to improve older patients’ involvement. The 31 inclusion criteria were RCTs and quasi-randomised trials; older participants 32 (65 or over) in primary care either before during or after the consultation and Medicines concordance: full guideline DRAFT (July 2008) page 81 of 373 DRAFT FOR CONSULTATION 1 the interventions had to aim to improve the patients’ involvement. Studies of 2 other health care professionals were excluded. The outcome measures 3 included use of health care, preparation for contact with a care provider, 4 contact with the care provider (communication), feedback of care, health 5 status and behaviour, treatment outcomes and outcomes related to health 6 professionals. Three studies met all inclusion criteria (Cegala 2001; Kimberlin 7 2001; Tennstedt 2001) were published in English and conducted in the USA. 8 The interventions were either face-to-face sessions to coach patients in 9 question asking and participating in consultations (given before the 10 consultation) or written interventions (booklet or checklist). Questioning 11 behaviour of patients was the primary outcome measure in two of the studies. 12 The other study had self-reported active behaviour as the primary outcome 13 measure. The studies used mostly qualitative analysis to assess the 14 outcomes and none were assessed at a later date. In Cegala (2001) the 15 trained patients asked more medically-related questions, gained more 16 information and provided more information than control patients. They did not 17 verify information more than control patients and appointment length was not 18 longer overall. In Kimberlin (2001) there was more question asking about their 19 medication than the control group. Tennstedt (2000) found older patients are 20 in general not involved in their GP visit. Few prepared questions or identified 21 issues to discuss with their GP. However, only 21% found that the GP 22 dominated the visit. The intervention group reported more behaviours such as 23 bringing a list of problems to the visit (non-significant in intention to treat 24 analysis but significant in those who did attend, and higher satisfaction with 25 the interpersonal aspects of the encounter). In conclusion, there is little 26 evidence of interventions specifically for improving older patients’ involvement 27 and thus they cannot recommend the use of the examined interventions in 28 clinical practice. Those interventions that were included resulted in patients 29 asking more and different questions, and they became more active in 30 consultations due to pre-visit preparation. 31 Wetzels (2005) (Wetzels, R., Wensing, M., van, Weel C. et al , 2005) ran a 32 cluster-randomised trial to assess a consultation leaflet implementation 33 programme in which GPs and older patients were encouraged to improve Medicines concordance: full guideline DRAFT (July 2008) page 82 of 373 DRAFT FOR CONSULTATION 1 older patients involvement when visiting their GP. This study was conducted 2 in the Netherlands. 3 All patients aged 70 years or above in the intervention practices received a 4 consultation leaflet by mail. This leaflet included a short motivating text on 5 patient involvement and a mixture of open and pre-structured questions to 6 help patients prepare for the next consultation and prioritize which problems 7 they wanted to discuss with their GP. The questions were chosen as they 8 would help to explore patient’s ideas, fears and expectations and encouraged 9 them to address important issues. Pre-intervention 315 patients and post- 10 intervention 263 patients were included in the study. 11 Based on results from a previous qualitative study, the authors concluded that 12 it would be important to include GPs in the implementation of the patient 13 involvement, so GPs in the intervention group received a 30 minute practice 14 visit to motivate them to involve patients and instructed them in the use of the 15 consultation leaflet. 16 The main results reported that subjects were satisfied with their involvements 17 and the GPs behaviour during the consultation, however there was no 18 difference in effect as a result of the leaflet on involvement, enablement or 19 satisfaction were found between the intervention and the control group. 20 Intervention group leaflet users reported more psychological symptoms to 21 their GP compared with non-users of the leaflet (p=0.034). 22 Overall the main findings do not support the use of the implementation 23 programme on improving involvement, enablement or satisfaction of older 24 patients in their care. 25 26 Lewin (2001) (Lewin, S. A., Skea, Z. C., Entwistle, V. et al , 2001) conducted 27 a Cochrane Collaboration systematic review of interventions to promote a 28 patient-centred approach in clinical consultations. The inclusion criteria were 29 RCTs, controlled before and after studies. The interventions promoted patient- 30 centred care in clinical health care consultations, not in social support or 31 social care. Other exclusion criteria were studies that considered cultural, 32 disability, sexuality or other sensitivity training only for health care providers; 33 evaluating training in psychotherapy or counselling for health care providers; Medicines concordance: full guideline DRAFT (July 2008) page 83 of 373 DRAFT FOR CONSULTATION 1 studies that trained health care providers to deliver a specific, secondary 2 intervention. The outcome measures consultation processes; other health 3 care behaviour; health status and wellbeing including physiological measures; 4 patient and/or carers’ satisfaction with care. Interventions were grouped into 5 the intensity of patient centredness and teaching tactics (weak, medium and 6 strong). 5260 titles and abstracts found, 135 with potential to be included, 17 7 were included. The studies were mainly conducted in North America, others 8 were from the UK, Switzerland, Norway and Trinidad and Tobago. The aims, 9 format, content of interventions and the clinical conditions on which they 10 focused showed heterogeneity. Studies were grouped into broadly similar 11 interventions: Patient-centred training compared with no training for providers 12 (11 studies*); Patient-centred training for providers plus patient-centred 13 training or materials for patients (3 studies**); Patient-centred training for 14 providers plus condition or behaviour-specific training or materials for 15 providers and patients (2 studies***); patient-centred training for providers, 16 patient-centred materials for patients plus condition- or behaviour-specific 17 materials for providers and patients compared with condition- or behaviour- 18 specific materials for both providers and patients (1 study****). The 19 participants were mainly health care primary care physicians and patients 20 were the recipients in six studies. In some of the studies the primary goal was 21 to increase patient centredness of care while, these studies tended to focus 22 on communication skills as important on their own right, others patient-centred 23 care was seen as a method to change patient behaviour or improve the health 24 outcome. Overall there is fairly strong evidence that some interventions which 25 promote patient-centred care in the consultation can lead to significant 26 increases in the patient centredness of the consultation process. 27 The following studies dealt with either specific populations or topics outside 28 our main focus: Clark (1998) paediatric doctors, Cope (1986) CCT, Langewitz 29 (1998) paediatric doctors, Lewis (1991) paediatric residents and fellows, 30 Meland (1997) was lifestyle changes for CHD risk. 31 *Cope (1986), Howe (1996), Langewitz (1998), Levinson (1993), Putnam 32 (1988), Robbins (1979), Roter (1995), Roter (1998), Smith (1995), Smith Medicines concordance: full guideline DRAFT (July 2008) page 84 of 373 DRAFT FOR CONSULTATION 1 (1998), Thom (1999).**Joos (1996), Lewis (1991), Pill (1998).***Clark (1998), 2 Meland (1997).****Kinmonth (1998). 3 Little (2004) (Little, P., Dorward, M., Warner, G. et al , 2004) conducted a 4 randomised control trial to assess the effect of leaflets in empowering patients 5 in primary care consultations. 636 patients were recruited in the study, aging 6 from 16-80 years and were attendees at one of five general practices in the 7 UK. Participants were randomised to four conditions: receipt of a general 8 leaflet, depression leaflet, both leaflets and no leaflets (control group). The 9 general leaflet which asked patients to list issues they wanted to raise and 10 explained that the doctor wanted them to ask questions, talk and discuss any 11 problems of concern to them. The depression leaflet listed symptoms of 12 depression (without labelling as such) and asking if had them and that the 13 doctor would like to discuss them. The outcomes measured were patient 14 satisfaction (the scores reflected aspects of doctor patient communication), 15 consultation time, prescribing, referral and investigation. 16 The only significant interaction was the increase in satisfaction for those who 17 received the general leaflet, the mean difference was 0.17 (95% CI 0.01 to 18 0.32, p=0.04). Consultation time and the general leaflet were significantly 19 associated with improved satisfaction, and leaflet was significantly more 20 effective when consultations were short (leaflet 0.64, 95% CI 0.19 to 1.08; 21 time 0.31, 0.0 to 0.06; interaction between both showed that consultations of 22 5, 8, and 10 mins increased satisfaction by 14%, 10% and 7%. This was also 23 shown for subscales of satisfaction – comfort from communication 1.02 (0.36 24 to 1.68), relief of distress 0.74 (0.0 to 1.49), intention to comply with 25 management decisions 0.65 (0.06 to 1.23) and rapport 0.81 (0.16 to 1.45). 26 The general leaflet increased the number of investigations by the doctor (OR 27 1.43, 1.00 to 2.05), which was unlikely to be due to chance or confounders 28 after controlling. 29 30 Wilkinson (2002) (Wilkinson, C. R. and Williams, M., 2002) conducted a 31 randomised controlled trial to investigate the relationship between providing 32 patients with an education guidebook designed to encourage and support 33 participation in the health care visit and selected patient and system outcome Medicines concordance: full guideline DRAFT (July 2008) page 85 of 373 DRAFT FOR CONSULTATION 1 measures. The study population included 277 predominantly male 2 participants, with an average age of approximately 60 years. This study was 3 conducted in the USA. 4 Patients in the intervention group were mailed the appointment guidebook 5 with instructions prior to a scheduled routine visit with their primary care 6 provider. The control group was mailed the standard letter informing them of 7 the upcoming appointments. After the visit, the patients in both groups were 8 sent a short questionnaire with instructions and postage-paid return envelope. 9 No significant differences were reported in the proportion of patients in the two 10 groups that agreed with any of the six statements relevant to primary care visit 11 effectiveness. However, significant differences were reported in the proportion 12 of patients who received preventative health care interventions of influenza 13 vaccination, pneumococcal vaccination and gender specific cancer screening. 14 15 Ross (2004) (Ross, S. E., Moore, L. A., Earnest, M. A. et al , 2004) conducted 16 a randomised controlled trial that assessed the effects of a patient accessible 17 online medical record on patient satisfaction, adherence and health status in a 18 randomised controlled trial conducted in the USA. The version used in the 19 study, System Providing Patients Access to Records Online (SPPARO) 20 provides access to clinical notes and test results, and also provides a method 21 of sending electronic messages to the clinical staff. 22 Patients included in the study were aged 18 years or older. One hundred and 23 seven were enrolled in the study. 24 Participants in the SPPARO group were given a user identification and 25 password to SPPARO and a written user guide to the system. The control 26 group continued to receive standard care in practice. Periodic messages were 27 sent by research staff to all participants and were informed about upcoming 28 surveys and encouraged to contact the research assistant if they had a 29 change of address or telephone number. They were also reminded that they 30 could contact the research assistant if they had problems using SPPARO. 31 No statistically significant results were found for self-efficacy, adherence to 32 medication, health status and patient satisfaction. General adherence to 33 medication advice showed significant improvement (however no statistical 34 significance information was provided in the paper) in the intervention group Medicines concordance: full guideline DRAFT (July 2008) page 86 of 373 DRAFT FOR CONSULTATION 1 compared with the control group. Although the number of patients who visited 2 the emergency department did not differ significantly, there was a significant 3 increase in the number of overall emergency department visit in the 4 intervention group compared to the control group (again, no statistical 5 significance information was provided). 6 7 Loh (2007) (Loh, Andreas, Simon, Daniela, Wills, Celia E. et al , 2007) 8 conducted a randomised controlled trial that investigated the effects of a 9 shared decision-making intervention in primary care of depression compared 10 to usual care on adherence, satisfaction and clinical outcomes. The study was 11 conducted in Sudbagen, Germany with primary care physicians as the unit of 12 randomisation. The sampling frame (n=148) were sent a letter, 30 accepted 13 the invitation to take part, 20 were randomly assigned to the intervention 14 group and 10 to the control group, after drop out 15 and 8 were left 15 respectively. The physicians had to recruit newly diagnosed depressive 16 patients. The intervention physicians enrolled 263 patients and the control 17 group 142. After their diagnosis the patients completed a questionnaire 18 measuring patient involvement, depression severity and socio-demographic 19 characteristics. After 6-8 weeks the patients completed a second 20 questionnaire measuring adherence and treatment outcome. At the same 21 time, the physicians rated their assessment of the patients’ adherence. The 22 shared decision-making intervention was then implemented with the 23 intervention group. The intervention was a multi-faceted program including 24 physician training, a decision board for use during the consultation given to 25 the patients after the consultation, printed patient information with specific 26 encouragement to be active in the decision-making process. The physicians in 27 the intervention group also completed modules on guideline-concordant 28 depression care which included content on enhancing skills for improving the 29 shared decision-making process. The outcomes measures were patient 30 participation, treatment adherence, patient satisfaction, consultation time and 31 clinical outcomes. There was no difference for the control group in patient 32 participation before and after, whereas the intervention group had significantly 33 higher patient participation from pre to post intervention for the doctor 34 facilitation scale (p=0.001) and there was an increase in the patient Medicines concordance: full guideline DRAFT (July 2008) page 87 of 373 DRAFT FOR CONSULTATION 1 participation scale (p=0.010). There were no significant differences in 2 treatment adherence. Patient satisfaction was significantly higher in the 3 intervention 29.8 (2.7) than the control group 27.0 (3.6), p=0.14. There were 4 no values taken for satisfaction before the intervention. There was no 5 difference between groups for length of consultation. Neither group had a 6 statistically significant reduction in depression severity from baseline to post- 7 intervention. 8 9 10 11 Medicines concordance: full guideline DRAFT (July 2008) page 88 of 373 DRAFT FOR CONSULTATION 1 4.6 Key Clinical Question: How can practitioners elicit 2 patients’ preferences for involvement in decisions 3 about medicines? Related Evidence statements (summary of Evidence into recommendations references evidence) 1. No tools designed for use in No tools for use in clinical practice were clinical practice were found. found. The GDG used the evidence from the literature and their professional opinion to develop recommendations on eliciting patients information needs. Ende (1989) 2. The Autonomy Preference Index (Ende, J., (API) and developments of the API Kazis, L., Ash, have been created to elicit patients’ A. et al , preferences for involvement in 1989); decision-making and may be Langewitz reduced to a 6-item subscale. (2006) (Langewitz, W., Nübling, M., and Weber, H., 2006); Tortolero (2006) (TortoleroLunaGuillermo, Byrd-Theresa Groff-Janet, 2006); Neame (2005) Medicines concordance: full guideline DRAFT (July 2008) page 89 of 373 DRAFT FOR CONSULTATION (Neame, Rebecca, Hammond, Alison, and Deighton, Christopher, 2005); Braman (2004) (Braman, Amie C., 2004); Doherty (2005) (Doherty, C. and Doherty, W., 2005); Schneider (2007) (Schneider, A., Wensing, M., Quinzler, R. et al , 2007); Hill (2006) (Hill, S. A. and Laugharne, R., 2006) Cox (2007) 3. A brief pre-consultation (Cox, K., questionnaire may be used to elicit Britten, N., patients’ preferences for Hooper, R. et involvement in decisions about Medicines concordance: full guideline DRAFT (July 2008) page 90 of 373 DRAFT FOR CONSULTATION al , 2007) medicines. Caress (1997) 4. A set of 5 sort cards can be used (Caress, A., to elicit the patients’ preferred role 1997) and perceived role within the consultation. Doherty 5. Two of these tools have been (2005) used within routine clinical settings. (Doherty, C. and Doherty, W., 2005) within a hospital and Cox (2007) (Cox, K., Britten, N., Hooper, R. et al , 2007) before and after a general practice consultation. 1 Medicines concordance: full guideline DRAFT (July 2008) page 91 of 373 DRAFT FOR CONSULTATION Methods of the evidence review 1 4.6.1 2 This paper includes a narrative summary of the included evidence, structured 3 according to the category of the intervention, following the agreed reviewing 4 protocol: 5 Types of studies – A previous search included only randomized controlled 6 trials (RCTs) and systematic reviews (SRs). No studies which met the criteria 7 were found. We widened the search to include any type of studies to find 8 relevant information to meet our inclusion criteria. 9 Types of participants - people prescribed medication for a medical condition. 10 Duration of studies – No time limit specified for the studies. 11 Types of interventions - Any intervention (tool) which elicits patient 12 preferences for involvement in decisions about medicines. The tools had to be 13 brief enough to be utilised within a consultation between the patient and 14 practitioner. Therefore long questionnaires were excluded as they would not 15 be manageable. 16 Types of outcome measures – Any outcome relating to the use of the tool 17 was acceptable as we were looking for a tool which could be utilized within a 18 consultation, rather than looking for specific clinical outcomes. 19 4.6.2 20 No tools designed for use in clinical practice were found. 21 The tools in this review are research tools and not clinical tools designed for 22 use in a consultation. We have included these research studies to illustrate 23 brief questionnaires that have been used in research settings indicating the 24 content of the questionnaires to inform the GDG. We have not reported in 25 detail on development or validation of these questionnaires. 26 Ende (1989) (Ende, J., Kazis, L., Ash, A. et al , 1989) constructed an 27 instrument to measure patient preferences for making medical decisions and 28 their desire to be informed. The instrument, named the Autonomy Preference Evidence review Medicines concordance: full guideline DRAFT (July 2008) page 92 of 373 DRAFT FOR CONSULTATION 1 Index (API) was in a questionnaire format and the scales were developed by a 2 group of 13 clinicians, medical sociologists, and ethicists who were interested 3 in patient autonomy. Items on the scales were tested for reliability and validity. 4 The final API consisted of an 8-item scale on information-seeking and a 15- 5 item scale on decision-making. Within this 15-item scale was a 6-item sub- 6 scale which related to general items. The other 9-items were related to three 7 clinical vignettes and were too disease-specific to be useful for this question. 8 The 6-item scale (part A) meets our criteria (see above) but is a research tool 9 rather than a clinical tool. 10 11 The Decision making preference scale (Part A): (responses on a 5-point Likert 12 scale ranging from ‘strongly disagree’ to ‘strongly agree’. The higher the score 13 on the scale, the more patients wished to participate in the decision-making. 14 1.* The important medical decisions should be made by your doctor, not by 15 you. 16 2. You should go along with your doctor’s advice even if you disagree with it. 17 3.* When hospitalized, you should not be making decisions about your own 18 care. 19 4. You should feel free to make decisions about everyday medical problems. 20 5.* If you were sick, as your illness became worse you would want your doctor 21 to take greater control. 22 23 24 6. You should decide how frequently you need a check-up. 25 The API was used within a variety of studies: Langewitz (2006) (Langewitz, 26 W., Nübling, M., and Weber, H., 2006); Tortolero (2006) (Tortolero-Luna- 27 Guillermo, Byrd-Theresa Groff-Janet, 2006); Neame (2005) (Neame, 28 Rebecca, Hammond, Alison, and Deighton, Christopher, 2005); Braman 29 (2004) (Braman, Amie C., 2004); Schneider (2007) (Schneider, A., 30 Wensing, M., Quinzler, R. et al , 2007) and Hill (2006) (Hill, S. A. and 31 Laugharne, R., 2006). All of the studies used the 6-item subscale as 32 illustrated above, except for Langewitz (2006) (Langewitz, W., Nübling, M., 33 and Weber, H., 2006) who adapted the instrument into two questions and Hill 34 (2006) (Hill, S. A. and Laugharne, R., 2006) slightly adapted the questionnaire * Scoring for these items was reversed, and goes from 5 to 1, rather than 1 to 5. Medicines concordance: full guideline DRAFT (July 2008) page 93 of 373 DRAFT FOR CONSULTATION 1 to apply to psychiatric patients. Most of the studies posted their questionnaires 2 to the participants rather than having them completed in a clinical setting. 3 4 Langewitz (2006) (Langewitz, W., Nübling, M., and Weber, H., 2006) 5 As part of the questionnaire, Langewitz (2006) (Langewitz, W., Nübling, M., 6 and Weber, H., 2006) adapted the API to 4 point Likert scale: fully agree, 7 slightly agree, slightly disagree, fully disagree. How much do you agree with 8 the following statements: 9 convinced of his ideas (Follow physician’s advice) 10 11 12 One should stick to the physician’s advice even if one is not fully It should completely be left to physicians to decide on a patient’s treatment (Physician should decide) 13 This was conducted at the University of Basel in NW Switzerland and sent to 14 the patients after discharge from hospital. 15 16 Degner and Sloan (1992) (Degner, L, 2008) adapted the API and used the 17 questions on sort cards for patients with cancer. Their questions were adapted 18 by Caress (1997) (Caress, A., 1997), Doherty (2005) (Doherty, C. and 19 Doherty, W., 2005) and Cox (2007) {6698). 20 Caress (1997) {1155} conducted a cross-sectional study at a Regional Renal 21 unit in the North of England with 462 participants from a convenience sample 22 over a 12 month period. 155 of the patients were pre-dialysis, 103 were 23 dialysis patients and 147 were transplant patients. Using a set of sort cards 24 (as used by Degner and Sloan, 1992 (Degner, L, 2008)) the patients picked a 25 single card which was closest to their preferred role in decision-making and 26 also picked a single card closest to their perceived role in decision-making. 27 Patients were also asked to give their rationale for their preferred role. 28 The 5 sort cards: 29 Active options 30 Card A: I prefer to make the final decision about which treatment I will receive. Medicines concordance: full guideline DRAFT (July 2008) page 94 of 373 DRAFT FOR CONSULTATION 1 Card B: I prefer to make the final selection of my treatment after seriously 2 considering my doctor's opinion. 3 Collaborative option 4 Card C: I prefer that my doctor and I share responsibility for deciding which 5 treatment is best for me. 6 Passive options 7 Card D: I prefer that my doctor makes the final decision about which treatment 8 will be used but seriously considers my opinion. 9 Card E: I prefer to leave all decisions regarding my treatment to my doctor. 10 The key points found from the study were that: participation preference was 11 highly individualistic, with a lot of patients wishing to remain passive. Those 12 who did prefer an active role were unlikely to attain this preference; trust in the 13 HCP can influence the preference; desire for information is not synonymous 14 with desire for participation. 15 Doherty (2005) (Doherty, C. and Doherty, W., 2005) used the adapted 16 questionnaire to use in an acute hospital trust in England and was one study 17 which tried to elicit the patients’ responses within an actual clinical situation. 18 19 Cox (2007) (Cox, K., Britten, N., Hooper, R. et al , 2007) included 479 20 patients who were approached in the waiting room in general practitioner 21 surgeries to participate and then given an interview where they completed the 22 pre-consultation questionnaire. They were also administered a questionnaire 23 after the consultation. The G.P was given a questionnaire before, which 24 included their preferred role in decision making with patients and a 25 questionnaire afterwards detailing their perceptions of the decision-making 26 during each consultation. The doctors’ assessment of patients’ preference to 27 be involved in shared decision making was correct in 32% of the 28 consultations, overestimated in 45% of the consultations and underestimated 29 in 23% of the consultations. The patients’ preferences for decision making Medicines concordance: full guideline DRAFT (July 2008) page 95 of 373 DRAFT FOR CONSULTATION 1 involved: 39% wanting the G.P to share the decision, 45% wanting the G.P to 2 be main (28%) or only (17%) decision-maker and 16% wanting to be the main 3 (14%) or only (2%) decision-maker. 4 The questionnaire given to the patients at pre-consultation included the 5 following 5 statements, of which patients were asked to choose one: 6 problem. 7 8 15 I would prefer that my doctor makes the final decision about medicines I take for this problem, but seriously considers my opinion. 13 14 I would prefer that my doctor and I share responsibility for deciding about medicines I take for this problem. 11 12 I would prefer that I make the final decision about medicines I take for this problem after seriously considering my doctor’s opinion. 9 10 I would prefer that I make the decision about medicines I take for this I would prefer that my doctor makes all decisions about medicines I take for this problem. 16 Medicines concordance: full guideline DRAFT (July 2008) page 96 of 373 DRAFT FOR CONSULTATION 1 2 4.7 Key Clinical Question: What tools are available to help elicit patients’ beliefs about medicines? Related references Evidence statements Evidence into recommendations (summary of evidence) Horne (1999) (Horne, 1. No tools designed for No tools designed for use in clinical R., 2008); Menckeberg use in clinical practice practice were found although the GDG (2008) (Menckeberg, T. were found. were aware of current studies to develop T., Bouvy, M. L., Bracke, M. et al , 2008); Horne (2007) (Horne, R., Cooper, V., Gellaitry, G. et al , 2007); Brown (2005) (Brown, Charlotte, such tools, in particular studies seeking The Beliefs about Medicines (BMQ) – Specific component is a research tool which is used to elicit beliefs about medicines. to adapt the BMQ for clinical use. The GDG reviewed the research tools found but it was not appropriate to use these outside their research settings. The GDG used the information from the research studies and their professional opinion to Battista, Deena R., make recommendations in relation to Bruehlman, Richard et elicitation of patients’ beliefs about al , 2005); Khanderia medicines. (2008) (Khanderia, U., Townsend, K. A., Erickson, S. R. et al , 2008); Kumar (2008) (Kumar, K., Gordon, C., Toescu, V. et al , 2008); Kemp (2007) (KempSteven, Feely-Morgan Hay-Alastair WildHeather CooperCathryn., 2007); Aikens (2008) (Aikens, James E., Nease-Donald-E-Jr, and Klinkman, Michael S., 2008); Clifford Medicines concordance: full guideline DRAFT (July 2008) page 97 of 373 DRAFT FOR CONSULTATION (2008) {1518}; Jenkins (2003) (7606); Theunissen (2003) (Theunissen-Nicolet, C. M., 2003) Hamilton (2007) 2. One RCT that assessed (Hamilton, W., Russell, a patient self-completion D., Stabb, C. et al , agenda form on 2007) prescribing and adherence showed no significant results for a prescription, satisfaction scores or adherence. Medicines concordance: full guideline DRAFT (July 2008) page 98 of 373 DRAFT FOR CONSULTATION Methods of the evidence review 1 4.7.1 2 This paper includes a narrative summary of the included evidence, structured 3 according to the category of the intervention, following the agreed reviewing 4 protocol: 5 Types of studies – We initially included only Randomised Controlled Trials 6 (RCTs) and Systematic Reviews (SRs), however none of these types of 7 studies were found that met the criteria. We increased the current search to 8 include any type of study in order to find relevant information to meet our 9 inclusion criteria. 10 Types of participants - people prescribed medication for a medical condition. 11 Duration of studies – No time limit specified for studies. 12 Types of interventions - Any intervention (tool) which elicits patient beliefs 13 about their medicines. The tools had to be brief enough to be utilised within a 14 consultation between the patient and practitioner. Therefore long 15 questionnaires were excluded as they would not be manageable. 16 Types of outcome measures – Any outcome relating to the use of the tool 17 was acceptable as we were looking for a tool which could be utilised within a 18 consultation, rather than looking for specific clinical outcomes. 19 4.7.2 20 No tools designed to elicit patients beliefs about medicines for use in clinical 21 practice were found. The tools which we found in this review were research 22 tools. We decided to include studies which used these research tools to 23 illustrate some questions that could be asked to the patient in a consultation. 24 Most studies were in questionnaire form and so we included those which were 25 shortest. We have not reported the parts of the questionnaire which were not 26 relevant to the clinical question. 27 Horne (1999) (Horne, R., 2008) created a questionnaire which explicitly states 28 the intention of assessing patients’ beliefs about medicines. The beliefs about 29 medicines questionnaire (BMQ) included two parts – the BMQ-General, which Evidence review Medicines concordance: full guideline DRAFT (July 2008) page 99 of 373 DRAFT FOR CONSULTATION 1 assessed beliefs about medicines in general and the BMQ-Specific, which 2 looks at patients’ specific beliefs towards their medicine. The study states that 3 the two sections of the BMQ can be used together or separately. As the BMQ- 4 Specific answers the question, and we are looking for brevity within the 5 consultation, this part of the study is reported. 6 The BMQ-Specific includes two 5-item factors which assess beliefs of the 7 necessity of medicines prescribed (Specific-Necessity) and concerns about 8 medicines prescribed, based on beliefs of the danger of dependence, long- 9 term toxicity and the disruptive effects of medication (Specific-concerns). 10 The BMQ-Specific items, which are rated ‘strongly agree, agree, uncertain, 11 disagree or strongly disagree’: 12 My health, at present, depends on my medicines. 13 Having to take medicines worries me. 14 My life would be impossible without my medicines. 15 Without my medicines I would be very ill. 16 I sometimes worry about long-term effects of my medicines. 17 My medicines are a mystery to me. 18 My health in the future will depend on my medicines. 19 My medicines disrupt my life. 20 I sometimes worry about becoming too dependent on my medicines. 21 My medicines protect me from becoming worse. 22 23 The BMQ-Specific was used in many other studies to assess beliefs about 24 medicines for a range of conditions (Menckeberg, 2008 (Menckeberg, T. T., 25 Bouvy, M. L., Bracke, M. et al , 2008); Horne, 2007 (Horne, R., Cooper, V., 26 Gellaitry, G. et al , 2007); Brown 2005 (Brown, Charlotte, Battista, Deena R., 27 Bruehlman, Richard et al , 2005); Khanderia, 2008 (Khanderia, U., Townsend, 28 K. A., Erickson, S. R. et al , 2008); Kumar, 2008 (Kumar, K., Gordon, C., 29 Toescu, V. et al , 2008); Kemp, 2007 (Kemp-Steven, Feely-Morgan Hay- 30 Alastair Wild-Heather Cooper-Cathryn., 2007); Aikens 2008 (Aikens, James 31 E., Nease-Donald-E-Jr, and Klinkman, Michael S., 2008); Clifford, 2008 Medicines concordance: full guideline DRAFT (July 2008) page 100 of 373 DRAFT FOR CONSULTATION 1 {1518}; Jenkins (2003) (Jenkins, Linda, Britten, Nicky, Stevenson, Fiona et al , 2 2003); Theunissen (2003) (Theunissen-Nicolet, C. M., 2003)). 3 4 We did retrieve one study, Hamilton (2006) (Hamilton, W., Russell, D., 5 Stabb, C. et al , 2007) , which was a randomised controlled trial conducted to 6 test the effect of patient self-completion agenda forms on prescribing and 7 adherence in general practice. This RCT was considered relevant as one of 8 the items of the self-completion form was related to medication expectations 9 and it was considered a clinical tool. 10 One thousand and six hundred and ten patients at ten general practices in 11 Devon and Dorset (UK) were involved in the trial for up to 12 weeks. All 12 patients were given a letter and an envelope when attending their GP. Half the 13 group received an agenda form which they could fill out while waiting for the 14 doctor. The other half received usual care. The agenda form called the SCAF 15 (self completion agenda form) included five questions: 16 1. 17 problem you have e.g. symptoms or current illness. 18 2. Your ideas about your illness: what do you think is wrong with you? 19 3. Your concerns: have you any particular worries about your illness? 20 4. Your expectations: how do you think your problem should be treated? 21 What do you hope the doctor will do? 22 5. 23 problem? 24 The GP was handed this form when the patient went into their appointment, to 25 use in whichever they thought appropriate. There were no significant 26 differences between proportion of patients who received a prescription, or in 27 satisfaction scores or adherence to prescribed medication. What made you decide to come to see the doctor? Please describe the Medication: do you think you should receive a prescription for your 28 Medicines concordance: full guideline DRAFT (July 2008) page 101 of 373 DRAFT FOR CONSULTATION 1 2 Medicines concordance: full guideline DRAFT (July 2008) page 102 of 373 DRAFT FOR CONSULTATION 1 4.8 Key Clinical Question: What tools are available to help elicit patients’ information needs? 2 Related references Evidence statements (summary of evidence) Evidence into recommendations 1. No tools designed for use in No tools for use in clinical practice clinical practice were found. were found. The GDG used the evidence from the literature and Kinnersley (2007) 2. One systematic review found (Kinnersley, P., that 17 RCTs measured question Edwards, A., Hood, asking with 6 finding statistically K. et al , 2007) significant increases and 11 finding their professional opinion to develop recommendations on eliciting patients information needs. no effects. Kinnersley (2007) 3. One systematic review found (Kinnersley, P., that patient satisfaction in 14 RCTs Edwards, A., Hood, there were no changes and 5 there K. et al , 2007) was significant increases in satisfaction. Strydom (2001) 2. One questionnaire has been (Strydom, A., developed which can assess Forster, M., Wilkie, information needs of people with B. M. et al , 2001) learning disabilities. Agård (2004) 3. One study used 4 open-ended (Agård, Anders, questions to elicit patient’s Hermerén, Göran, information needs. and Herlitz, Johan, 2004) Duggan (1992) 4. Three studies included 5 open Medicines concordance: full guideline DRAFT (July 2008) page 103 of 373 DRAFT FOR CONSULTATION (Duggan, C. and questions from the Intrinsic Desire Bates, I., 2000); for Information (IDI) scale which Astrom (2000) can elicit information needs from (Astrom, K., patients. Carlsson, J., Bates, I. et al , 2000); Zwaenepoel (2005) (Cramer, Joyce A., 2004). Ende (1989) (Ende, 5. Seven studies use the J., Kazis, L., Ash, A. Autonomy Preference Index 8-item et al , 1989); scale to elicit information needs of Langewitz (2006) patients. (Langewitz, W., Nübling, M., and Weber, H., 2006); Tortolero (2006) (Tortolero-LunaGuillermo, ByrdTheresa Groff-Janet, 2006); Neame (2005) (Neame, Rebecca, Hammond, Alison, and Deighton, Christopher, 2005); Braman (2004) (Braman, Amie C., 2004); Doherty (2005) (Doherty, C. and Doherty, W., 2005); Schneider (2007) (Schneider, Medicines concordance: full guideline DRAFT (July 2008) page 104 of 373 DRAFT FOR CONSULTATION A., Wensing, M., Quinzler, R. et al , 2007) 1 Method of the evidence review 2 4.8.1 3 This paper includes a narrative summary of the included evidence, structured 4 according to the category of the intervention, following the agreed reviewing 5 protocol: 6 Types of studies – We initially included only randomized controlled trials 7 (RCTs) and systematic reviews (SRs) however none of these types of studies 8 were found to meet the criteria. We increased the search to include any type 9 of studies to find relevant information to meet our inclusion criteria. 10 Types of participants - people prescribed medication for a medical condition. 11 Duration of studies – No time limit specified for the studies. 12 Types of interventions - Any intervention (tool) which elicits patients’ 13 information needs. The tools had to be brief enough to be utilised within a 14 consultation between the patient and practitioner. Therefore long 15 questionnaires were excluded as they would not be manageable. 16 Types of outcome measures – Any outcome relating to the use of the tool 17 was acceptable as we were looking for a tool which could be utilized within a 18 consultation, rather than looking for specific clinical outcomes. 19 4.8.2 20 No tools designed for use in clinical practice were found. The tools in this 21 review were research tools opposed to clinical tools which could be used in a 22 consultation. We decided to include studies which used these research tools 23 to illustrate some questions that could be asked to the patient in a 24 consultation. Most studies were in questionnaire form and so we included 25 those which were shortest – and so could possibly be used in a consultation. Evidence review Medicines concordance: full guideline DRAFT (July 2008) page 105 of 373 DRAFT FOR CONSULTATION 1 We have not reported the parts of the questionnaire which were not relevant 2 and useable in a consultation. 3 One Cochrane Review included in section 5.4 addressed ways of eliciting 4 patients information needs. Kinnersley (2007) (Kinnersley, P., Edwards, A., 5 Hood, K. et al , 2007) conducted a Cochrane Collaboration systematic review 6 of the effects of interventions before consultations designed to help patients 7 address their information needs. The settings and populations varied, and 8 were from six countries, mainly the USA. They stated that it complemented 9 other Cochrane reviews by Wetzels (2007) (Wetzels, R., Harmsen, M., van, 10 Weel C. et al , 2007) and Lewin (2003). The inclusion criteria were RCTs of 11 interventions intended to help the patients, representatives or carers address 12 their information needs in a consultation. This was done by encouraging 13 question asking, to express their information needs, to overcome barriers to 14 communication and to clarify their understanding of the information provided. 15 Outcome measures were the consultation process, consultation outcomes 16 and service outcomes. 33 trials described in 35 studies met the inclusion 17 criteria. Of the studies assessing single interventions for patients 15 included 18 written materials, four were coaching. The multiple component single 19 interventions studies four had coaching and written materials. Results: 17 20 studies measured question asking with 6 finding statistically significant 21 increases and 11 studies finding no effects of the interventions compared to 22 controls. Patient participation was measured in 14 studies, it was increased in 23 8, and no effect in five studies. Patient satisfaction was measured in 23 24 studies, in 14 studies there were no changes and 5 there was increased 25 satisfaction. Patient knowledge was measured in 5 studies with reduction in 26 two studies and no changes in 3 studies. According to type of intervention, 27 comparisons between written alone and coaching alone there were similar, 28 small to moderate and statistically significant increases for both types for 29 question asking. Patient satisfaction was borderline statistically significant for 30 written materials, coaching the effect was small and statistically significant. 31 Written materials led to a small and statistically significant increase in 32 consultation length, coaching the increase was smaller but was not significant. Medicines concordance: full guideline DRAFT (July 2008) page 106 of 373 DRAFT FOR CONSULTATION 1 Interventions immediately before the consultation led to small statistically 2 significant increase in consultation length and patient satisfaction. 3 It should be noted that many of the studies were from other settings: Brown 4 (1999, 2001), Bruera (2003), Butow (1994, 2004), Davison (1997, 2002), Ford 5 (1995), Oliver (2001) were cancer studies. Finney (1990), Kim (2003), Lewis 6 (1991) were from paediatric and family planning settings. 7 8 9 Strydom (2001) (Strydom, A., Forster, M., Wilkie, B. M. et al , 2001) 10 conducted a study of a service-user questionnaire to find out gaps in 11 medication knowledge and information sources. This study specifically 12 involved finding out the views of those with learning disabilities. 21 13 participants were included who were either currently taking prescribed 14 medication (GP or specialist health services) or had taken in the recent past. It 15 should be noted that two thirds of the subjects received help with taking 16 medication. A questionnaire was designed by the authors using previously 17 published guidelines. They used structured and semi-structured sections, 18 including open questions. The questionnaire was delivered by one of the 19 research team with experience of communicating with people with learning 20 disabilities. The questionnaire was designed to find out their experiences and 21 opinions of using medicines. 22 23 A table was given in the paper to show the questions relating to medication 24 knowledge. Please note that it is unknown as to whether this was exhaustive. 25 The open questions were not reported in the paper. 26 Can you read the medication label? (Yes, no) 27 What is written on the label? (Don’t know, medication name, my name, Chemist’s name, dose, other) 28 29 Approximate name, description). 30 31 32 What is your medication called? (Don’t know, Brand or generic name, What are you taking medication for? (Don’t know, knew indication, approximate indication). Medicines concordance: full guideline DRAFT (July 2008) page 107 of 373 DRAFT FOR CONSULTATION 1 know, Yes, plus example) 2 3 Is there anything you should not do while taking this medication? (Don’t Are there any side-effects? (Don’t know, one, two or more). 4 The resulting answers led to the framework for a structure of a patient 5 information leaflet for people with learning disabilities who take medicine for 6 psychiatric conditions. It is not clear as to whether the service users who filled 7 in the questionnaire were taking psychiatric medication or another type of 8 medication. The subjects were ‘selected for their range of experiences of 9 taking medications’. Therefore this study does not elicit patients’ information 10 needs in total, but how to elicit the knowledge gaps of those with a learning 11 disability. 12 13 Duggan (2000) (Duggan, C. and Bates, I., 2000)developed and evaluated a 14 survey tool (intrinsic desire for information) to find out Patients’ perceptions 15 and information needs in regards to their medication. It was tested for 16 reliability and by factor analysis and was used with 2 cohorts of patients in 17 East London (sample of 500). 18 19 Astrom’s (2000) (Astrom, K., Carlsson, J., Bates, I. et al , 2000) paper refined 20 and validated the IDI into a 12-item scale. They included 5 open questions 21 which were a joint construction of the project aims and questions from 22 Lindegren (1999), which was a Masters thesis at the Department of 23 Biopharmaceutics at Uppsala University. The open questions were 24 transcribed at the bedside of 299 patients in the wards of three medical 25 hospitals in London. Astrom (2000) concluded that the desire for information 26 may be more complicated and involve an emotional or behavioural 27 component, which was not included. It should be noted that this is a desire for 28 information which may differ from information needs. 29 30 The 12-item scale was deemed too long to meet our inclusion criteria, 31 however some of the open questions may be of relevance. 32 33 The IDI (for reference only): Medicines concordance: full guideline DRAFT (July 2008) page 108 of 373 DRAFT FOR CONSULTATION 1 Part 1 – Demographic details. 2 Part 2 – Questionnaire items (scored from strongly agree through strongly 3 disagree on a 5-point Likert scale). 4 1. I always speak to my pharmacist when I want information about my medicines 5 6 2. Sometimes I feel a little inhibited when I ask 7 for information…they might think I should 8 know already. 9 3. If there is anything I need to know, it’s most convenient to ask at the surgery. 10 11 4. It’s not really my place to ask for information, they have enough to do. 12 13 5. The people at the hospital can easily give me information when I go for my appointment. 14 15 6. I need as much information about my medicines as possible. 16 17 7. Too much knowledge is a bad thing. 18 8. You can never know enough about these things. 19 20 9. I don’t need any more knowledge about my medicines/illness. 21 22 10. I read about my medicines/illness as much as 23 possible. 11. What you don’t know (with respect to 24 medicines/illness) doesn’t hurt you. 25 26 12. I find information about my medicines/illness 27 confusing 28 29 Open questions: 30 13. What kind of information about your 31 medicines do you want? Why? 32 33 14. How do you want your information to be presented (written, oral, both, other)? Why? Medicines concordance: full guideline DRAFT (July 2008) page 109 of 373 DRAFT FOR CONSULTATION 15. Who would you like to give you information 1 about your medicines? Why? 2 3 16. When would it be best to have the information 4 about your medicine presented (at hospital, at 5 home, at the community pharmacy, at the 6 GP’s)? Why? 17. Would you like to sit down and talk about your 7 medicines with a pharmacist at the hospital? 8 9 10 Zwaenepoel (2005) (Cramer, Joyce A., 2004) used the IDI scale and 5 open 11 questions in a survey of the need for information of 279 psychiatric in-patients 12 in Flanders. 13 14 Ende (1989) (Ende, J., Kazis, L., Ash, A. et al , 1989) created the Autonomy 15 Preference Index (API) which, as well as a decision-making preference scale 16 had an eight-item information-seeking preference scale. The instrument was 17 in questionnaire format and the scales were developed by a group of 13 18 clinicians, medical sociologists, and ethicists who were interested in patient 19 autonomy. Items on the scales were tested for reliability and validity. The final 20 API consisted of an 8-item scale on information-seeking scale (ISS) and a 15- 21 item scale on decision-making. 22 The 8-item ISS consisted of the following items for information-seeking 23 preference, with responses on a five-point Likert scale from ‘strongly disagree’ 24 to ‘strongly agree’: 25 26 27 28 1. As you become sicker you should be told more and more about your illness. 2. You should understand completely what is happening inside your body as a result of your illness. 29 3. Even if the news is bad, you should be well informed. 30 4. Your doctor should explain the purpose of your laboratory tests. 31 5. *You should be given information only when you ask for it. 32 6. It is important for you to know all the side effects of your medication. 33 7. Information about your illness is as important to you as treatment. Medicines concordance: full guideline DRAFT (July 2008) page 110 of 373 DRAFT FOR CONSULTATION 1 8. When there is more than one method to treat a problem, you should be told about each one. 2 3 * Scoring for this item is reversed and goes from 5 to 1, rather than 1 to 5. 4 5 Five further studies used the Autonomy Preference Index to elicit patients 6 information needs (Langewitz, 2006 (Langewitz, W., Nübling, M., and Weber, 7 H., 2006); Tortolero, 2006 (Tortolero-Luna-Guillermo, Byrd-Theresa Groff- 8 Janet, 2006); Neame, 2005 (Neame, Rebecca, Hammond, Alison, and 9 Deighton, Christopher, 2005); Braman, 2004 (Braman, Amie C., 2004); 10 Schneider, 2007 (Doherty, C. and Doherty, W., 2005); Hill, 2006 (Hill, S. A. 11 and Laugharne, R., 2006). All of the studies used the full length (8-item) 12 information-preference scale except for Langewitz (2006) (Langewitz, W., 13 Nübling, M., and Weber, H., 2006) who incorporated only one question from 14 the API to target information needs: ‘Even when the news is bad the patient 15 must be informed (information)’. Hill (2006) slightly altered the API questions 16 for psychiatric patients to use. 17 18 Agard (2004) conducted a qualitative analysis of semi-structured interviews in 19 Gothenburg, Sweden on 40 patients 60 years and over who were receiving 20 treatment after a heart failure diagnosis. The semi-structured qualitative 21 interview had 4 open-ended questions as an interview guide. The questions 22 were: 23 1. What is your opinion about the medical information that you have been 24 given? 25 2. What kind of information is lacking? 26 3. What information have you been given about heart failure? 27 4. What is your attitude toward receiving prognostic information? 28 They were also encouraged to speak about the questions and to raise other 29 issues related to them to ensure their major personal concerns really 30 emerged. To avoid respondents feeling ignorant or embarrassed about not 31 being able to adequately answer questions relating knowledge they were Medicines concordance: full guideline DRAFT (July 2008) page 111 of 373 DRAFT FOR CONSULTATION 1 asked first about the information they had been given, rather than asking 2 directly about their knowledge of diagnosis, treatment and prognosis. 3 4 Many patients had a limited understanding of their disease but said they were 5 still satisfied with the information they received. Some were indifferent to, 6 accept or be unaware of their low level of knowledge. 7 They concluded that 'to inform the patient adequately, physicians and nurses 8 should determine the patient's level of knowledge and explore why those 9 patients who have a limited understanding do not assimilate or request 10 information. The information they provide should also be adapted to the 11 patient's capacity, wishes and emotional reactions.’ 12 13 14 15 Medicines concordance: full guideline DRAFT (July 2008) page 112 of 373 DRAFT FOR CONSULTATION 1 2 4.9 Key Clinical Question: How can information about 3 medicines be provided for patients in order to enhance 4 SDM in regard to medicines? Related references Trevena (2006) (Trevena, Evidence statements Evidence into (summary of evidence) recommendations 1. A systematic review of Information provided to patients Lyndal J., Davey, Heather systematic reviews and RCTs about treatments increases their M., Barratt, Alexandra et found that communicating with understanding whatever the al , 2006) patients about evidence does format (verbal, written or video) increase their understanding particularly if the information is regardless of what tools used. structured or interactive and There was a greater effect if especially when tailored to the information was structured individual. The GDG considered (either written, verbal or video) that health care professionals or interactive, especially if need to be aware that individuals tailored to the individual. will vary in the amount and type of information they require and in Trevena (2006) (Trevena, 2. One systematic review of Lyndal J., Davey, Heather systematic reviews and RCTs M., Barratt, Alexandra et found that probabilistic al , 2006) information is best represented as event rates, rather than words, probabilities, or summarised as effect measures such as relative risk reduction. Illustrations, such as cartoons or graphs appear to aid understanding. how they can best access that information. It was the professional opinion of the GDG that undue emphasis is currently placed on use of leaflets and written information and there is inadequate access to pictorial and graphic information. Examples of useful websites were presented to the GDG showing information presented in a variety of ways and the GDG believed it important to widen Medicines concordance: full guideline DRAFT (July 2008) page 113 of 373 DRAFT FOR CONSULTATION Wills (2003) (Wills, C. E. 3. One systematic review of knowledge and access to such and Holmes, Rovner M., information formats concluded resources. 2003) that decision support/aids can address patient information needs for shared decision making. They enable patients to better understand treatment options, including probability information. 1 2 Evidence review 3 4.9.1 4 This paper includes a narrative summary of the included evidence, structured 5 according to the category of the intervention, following the agreed reviewing 6 protocol: 7 Types of studies: Systematic Reviews of Randomised Controlled Trials 8 (RCTs) or Randomised Controlled Trials of interventions involving shared 9 decision-making in the clinical context. 10 Types of participants: people prescribed medication for a medical condition. 11 Duration of studies: no time limit specified. 12 Types of interventions: any interventions involving shared decision making 13 in a consultation between a health care professional and patient. 14 Types of outcome measures: Patient-centred communication in the 15 consultation; Consultation process outcomes: patient involvement, question 16 asking, preparedness; Patient care outcomes: satisfaction, knowledge, self- 17 efficacy. Type of interventions involved and type of information. 18 19 It should be noted that the remit is for conditions with prescribed medication 20 and this excludes conditions which require chemotherapy or screening. All Medicines concordance: full guideline DRAFT (July 2008) page 114 of 373 DRAFT FOR CONSULTATION 1 RCTs are within this remit, however many of the systematic reviews included 2 populations outside the remit, this is noted where applicable. 3 4 Trevena (2006) (Trevena, Lyndal J., Davey, Heather M., Barratt, Alexandra et 5 al , 2006) conducted a systematic review of RCTs and review of reviews 6 structured around three aspects of communication with patients about 7 evidence: patients’ preferences and actions; research evidence; and the 8 clinical state and circumstances. These were then translated into three main 9 questions: 10 understanding of evidence? 11 12 15 What are the most effective formats to represent probabilistic information to improve patient understanding of evidence? 13 14 What are the most effective communication tools to improve patient What are the most effective strategies to elicit patient preferences/beliefs/values relating to evidence? 16 The authors excluded studies that did not address and question; were about 17 patient education; were focused on skills and behaviour outcomes without 18 attempting to increase understanding or knowledge; were concerned with 19 counselling as a therapeutic intervention; or were specific to communication 20 regarding clinical trial participation. 21 Overlap between the trials included in the systematic reviews and those 22 identified independently was verified and duplicated studies were excluded. 23 Ten systematic reviews of RCTs and 30 additional RCTs were retrieved. 24 The review concluded that communicating with patients about evidence does 25 increase their understanding regardless of the tools used. The authors also 26 found that there was a greater effect if information was structured (either 27 written, verbal or video) or interactive (computer, touch screen, question 28 prompts) and particularly if the information was tailored to the individual. 29 Probabilistic information was found to be best represented as even rates in 30 relevant groups of people, rather than words, probabilities or summarized as 31 effect measures such as relative risk reduction. Written information was 32 reported to be more effective if illustrations and graphs were used. Medicines concordance: full guideline DRAFT (July 2008) page 115 of 373 DRAFT FOR CONSULTATION 1 The authors did however remark that one of the difficulties in generalising 2 from the literature was that the trials were conducted in a wide variety of 3 clinical settings using a range of clinical problems and outcomes. 4 5 Wills (2003) (Wills, C. E. and Holmes, Rovner M., 2003) conducted a 6 systematic review of patient health information provision and use for treatment 7 decision-making. It included research from the past 10 years focusing on 8 testing different formats of information presentation for patient decision- 9 making. The three types of formats looked at were probability presentations, 10 graphic formats and words vs numbers. They found two studies where 11 participants preferred presentation of medication in terms of relative risk rather 12 than absolute risk format. They found that people simplify relative risk 13 information into a simplified format of small or large risks and there is a 14 tendency to seriously under or overestimate their personal risks for health 15 outcomes. There is a need to tailor the format of risk communication to the 16 individual’s level of numeracy. In routine clinical encounters information 17 should be presented balanced, in both positive and negative frames. Graphics 18 can improve the understanding of numerical probability information. However 19 some people may dislike some types of displays or misunderstand them. 20 Consistent finding of individual differences in preferences for probability 21 information in words, numbers of both formats implies a need for routine 22 individualized assessments of patient preferences for format. In conclusion, 23 decision support/aids can address patient information needs for shared 24 decision making. They enable patients to better understand treatment options, 25 including probability information. 26 27 Medicines concordance: full guideline DRAFT (July 2008) page 116 of 373 DRAFT FOR CONSULTATION 1 2 4.10 Key Clinical Question: What information about 3 medicines should be provided for patients in order to 4 enhance SDM in regards to medicine? Recommendations Evidence into recommendations including The GDG considered that the provision of information to patients is to facilitate informed choice about medicines, and achieve a clear picture of the benefits and risks. The information that should be provided to a patient is dependent on what that patient needs to make a decision and prescriptive list can therefore not be generated. Patient Information Leaflets (PILS) provided with medicines often do not help in providing information about medicines and in any case are only received after the medicine is dispensed. The GDG did consider that some broad areas of information that patients might require and used the evidence from expert reviews, in particular those with patient involvement to inform this. They also considered that sources such as MHRA leaflet might be useful for patients. Health care professionals that questions from patients are often not declared and patients are not always able to ask the questions that concern them. Medicines concordance: full guideline DRAFT (July 2008) page 117 of 373 DRAFT FOR CONSULTATION 1 4.10.1 Evidence review 2 The evidence review is a narrative review. The GDG requested review of 3 current national guidance and reports with a particular emphasis on those 4 where patients views and perspectives were given priority. 5 4.10.1.1 Summary of ‘Always read the leaflet (2005) Report of the 6 Committee on Safety of Medicines Working Group on Patient 7 Information’1 8 The Working Group on Patient Information received advice from patients and 9 experts on the quality of Patient Information Leaflets (PILs) and how to 10 improve them. Patient organisations identified the following with regards to 11 PILs: 12 The quality is variable and language is complex with too much jargon 13 The leaflet is often too busy and the print too small 14 Leaflets are too negative and do not mention enough of the benefits 15 The PIL should complement discussion with the prescriber, ideally available in the consultation 16 17 organisations for further advice could be given 18 19 Helpline numbers and website addresses for further information should be mentioned 20 21 One PIL can not meet everyone’s needs so information on patient Comparative information and information about lifestyle issues can help in decision making 22 23 24 Expert views regarding PILs: 25 Too much use of jargon 26 The use of capital letters was eye-catching but hard to read 27 Inappropriate punctuation can obscure the message 28 Text in boxes may be skipped over 29 Euphemisms are not helpful when referring to serious side effects 30 Messages should be consistent Medicines concordance: full guideline DRAFT (July 2008) page 118 of 373 DRAFT FOR CONSULTATION 1 Language should be clear and unambiguous 2 3 Research2 showed that patients prioritise four key points of information about 4 a medicine – side effects, dos and don’ts, what it does and how to take it – but 5 different people prefer different orders of priority. 6 The Working Group recommended the following for an improved readability 7 guideline: 8 Usability – PILs should be clear and understandable to the reader. This 9 incorporates writing style, typeface, design and layout, headings, use of 10 colour, use of symbols and pictograms. The use of templates so that 11 information is presented consistently would be useful. (see annex 6). PILS 12 should not be too long and complex. 13 User involvement in PILs – user testing of patient information is 14 recommended. This should be done under controlled conditions and meeting 15 certain stipulations. User testing of content and impact is important. The 16 production of PILs by Companies often occur at the end of the medicine 17 development process, with little thought of involving patients in writing and 18 testing the information. Views should be at all stages of development. 19 Communicating Risk 20 It is fundamental when making an informed decision to understand and weigh 21 up the risk and benefits of a treatment. The working group suggested: 22 and effectiveness of using the medicine. 23 24 Use of headlines to summarise the most important messages for safety Information on all the side effects is required by law but must be 25 presented logically and include a description of side effects, estimating 26 frequency and advice on necessary actions. 27 Inclusion of the potential benefits to provide balance is important. Medicines concordance: full guideline DRAFT (July 2008) page 119 of 373 DRAFT FOR CONSULTATION 1 Provide information about the harmful effect itself; the probability of it 2 occurring and how to minimise risk and what actions to take if a 3 problem arises 4 Put most important information first, include information on benefit, use 5 the right words and use numbers to convey risk. Also a supplementary 6 leaflet of risks and benefits in addition to PILs would be useful. 7 8 Trust in the information source is also important. Harms and benefits should 9 be side by side and drug side effects must legally be provided. Care should be 10 taken to give unbiased and clear statistical information. 11 To increase trust in PILs transparency of data and certainty of risk estimates 12 may be effective. To avoid unnecessary concerns the use of clear information 13 on a true scale and nature of such risks are important, such as using 14 analogies and alternative risk scales to show rarity of risk; describe baseline 15 risk and increased risk; provide further information sources on these risks. 16 Headlines – It is suggested that information could be portrayed in headlines 17 which should include: why the patient should take the product; the maximum 18 dose or duration of treatment; potential side effects/withdrawal reactions; 19 contraindications; important drug interactions; circumstances in which the 20 drug should be stopped; what to do if the medicine does not work or where to 21 find further information. Headlines should also include a firm encouragement 22 for the patient to read the rest of the leaflet. 23 Balance - It is important to be balanced and convey information on benefits 24 as well as risks in order for the information to be credible. The PIL should 25 therefore include the potential benefits of taking the medicine. Research 26 shows that short factual statements on benefits help weigh the risks and 27 benefits. It must also be compatible with the Summary of Product 28 Characteristics, and be useful to the patient but not promotional. 29 Information to give a balanced account would include: 30 Why it is important to treat the disease Medicines concordance: full guideline DRAFT (July 2008) page 120 of 373 DRAFT FOR CONSULTATION 1 Whether the treatment is for short term or chronic use 2 Whether the medicine is being used to treat the underlying disease or for controlling symptoms 3 4 Whether the effects will last after medication stopped 5 Where it is to treat two or more discrete indications, all should be succinctly described as above 6 7 Where to obtain more information on the condition 8 9 10 11 Side effects Better information about side effects would include: Logical order - the most important information should be first e.g. 12 situations where need to take action such as stopping medication or 13 getting medical help 14 what to do if encounter serious problems 15 Estimates of frequency should be mentioned – as the most serious side effects are also the rarest 16 17 seriousness/severity 18 19 Many side effects are dose related and so a warning statement is needed but should not alarm those prescribed high doses 20 21 Use the right wording – not just describe the side effect but convey It would be useful to have a glossary of lay terms – so there is standardised side effect lay terminology across medications. 22 23 24 Expression of risk 25 Expressing statistical risk in PILS: 26 quantifying risk using absolute numbers; 27 verbal descriptors of risk only used if accompanied by equivalent 28 statistical information; Medicines concordance: full guideline DRAFT (July 2008) page 121 of 373 DRAFT FOR CONSULTATION 1 convey uncertainty around risk estimates; frequency ranges; duration 2 of risk; frequency estimates based on spontaneous adverse drug 3 reaction data; constant denominators; 4 5 Concepts deemed inappropriate by the Working Group were: 6 NNT/NNH 7 positive framing and negative framing – too cumbersome and lengthy 8 use of diagrams – constraints in size 9 10 Supplementary information - a leaflet about risks and benefits in addition to 11 PILs would be able to go into more detail. 12 Meeting the needs of special groups of patients 13 Not everyone finds it easy to access and use information in the PIL, e.g. 14 visually impaired people, people whose English is not their first language, 15 people with poor literacy and numeracy, those with learning difficulties or 16 physical difficulties. 17 Suggestions are made and projects described to help these patient groups: 18 For health literacy: 19 A health search engine for healthcare staff and public; 20 Patient information bank for NHS trusts to print consistent information for individuals on their care and treatment; 21 22 Power questions to ask in consultations; 23 24 Poor basic skills: 25 Clearly written in plain English 26 Signpost other sources of information 27 Helplines Medicines concordance: full guideline DRAFT (July 2008) page 122 of 373 DRAFT FOR CONSULTATION 1 2 Patients with sight loss: 3 leaflets in Braille or large print 4 audio version 5 leaflets on the web 6 digital television 7 telephone helplines and automated voice systems. 8 9 10 First language not English: Provision of leaflets in other languages from the company in written or web-based format 11 12 Telephone helplines 13 The use of translator services 14 15 16 Medicines for children and young people and so leaflets should be aimed at adults 17 18 Information for Children should be communicated by parents or carers Information for young people should take into account their lifestyle of the age group and likely questions 19 20 21 22 Provision of information for carers may need training on administration techniques 23 24 25 Carers may not be in the consultation when medication prescribed and Outside power of the group but use of a telephone helpline could address some concerns 26 27 The Pharmaceutical Companies Medicines concordance: full guideline DRAFT (July 2008) page 123 of 373 DRAFT FOR CONSULTATION 1 2 Responsibilities: to the information on the PILs and other measures. 3 4 It is suggested that the pharmaceutical industry could promote access Portfolio of Information keys for pharmaceutical companies – use these 5 to help identify additional measures that would promote the 6 dissemination of information on safe use of their products to ensure 7 vulnerable group s can access it. 8 Leaflets in other formats; how to signpost these other formats; 9 translation into other languages; use of information mediators such as 10 helplines; expert sources of advice. The PIL can be a pointer to other 11 sources of information for vulnerable groups e.g. booklets, simplified 12 leaflets, magazines and websites. 13 14 The information format of the patient leaflet is very important and should be 15 clear and understandable. The information needs to be balanced, trustworthy, 16 include benefits as well as side effects, with the most important information 17 highlighted. The communication of risk should be conveyed with seriousness 18 but without alarm for the patient. Where to get extra information should be 19 mentioned, if not a separate detailed booklet given. It would be very good 20 practice to have patients test the leaflet to see it’s appropriateness. Special 21 groups of patients should be taken into account while producing PILs. 22 4.10.1.2 23 Raynor (2007) (Raynor, D. K., Blenkinsopp, A., Knapp, P. et al , 2007) Summary 24 Raynor (2007) (Raynor, D. K., Blenkinsopp, A., Knapp, P. et al , 2007) 25 conducted workshops to elicit stakeholder perceptions of the key issues 26 surrounding information presentation to patients. The workshop discussions 27 found timing of the delivery of the information important, which was often 28 presented after the medicine was prescribed. Sometimes no leaflet was 29 available at all. They found too much information to be overwhelming, harder 30 to understand, often frightening and can have too much irrelevant information. 31 Readability was the most important part of written medicines information – the 32 size of text and content, meaningful information and not jargon. Medicines concordance: full guideline DRAFT (July 2008) page 124 of 373 DRAFT FOR CONSULTATION 1 2 Other information of importance: 3 dosage and ingredients 4 when and how long to take it 5 the likelihood of it being successful 6 side-effects e.g. how common or rare they are 7 factors relevant to their personal medical condition 8 9 Also the role of medicines: 10 how to take the drug effectively 11 its potential side-effects and interactions 12 how to reduce potential harm from medicines 13 how long before the medicine will have beneficial effects 14 why it is necessary to finish the course 15 why it was recommended not to drink alcohol 16 17 18 What makes medicines information effective? Timing of delivery of the information – more effective during the consultation 19 20 Visually appealing and straightforward to read 21 No jargon 22 Basic information about what the medicine contains 23 Designed for patients or professionals 24 25 What participants feel makes medicines information valuable? 26 Looks and feels important and highlights priority information 27 Permits an informed choice 28 Is reassuring and reduces concern, conflict and anxiety about whether the medicine is the right one for them 29 30 Gives them confidence in taking medicines Medicines concordance: full guideline DRAFT (July 2008) page 125 of 373 DRAFT FOR CONSULTATION 1 4.10.1.3 The ‘Medicine use review: Understand your medicine’ NHS report4 summary 2 3 The ‘Medicine use review: Understand your medicine’ NHS report4 suggests 4 questions patients should ask about their medication. These could be adapted 5 (as below) into written information for patients: 6 What the medicine does 7 Why is it important to take the medicine 8 Other treatment options 9 When and how it should be taken 10 How long it should be taken for 11 What medicines, drinks, foods or activities the patient should be aware of when taking the medicine 12 13 What the patient should do if they do not feel well when taking it 14 How the patient can tell if it’s helping 15 How the patient can be sure it’s safe to take it 16 The possible risks and side effects 17 What to do if they get one of the side effects 18 What happens if the patient stops the medicine or takes a lower dose 19 Is there anything they can do to remember to take their medicines 20 Where to go for more information 21 “Taking Medicines” leaflet summary 22 4.10.1.4 23 The Medicines and Healthcare products Regulatory Agency (MHRA) have 24 produced a leaflet for patients called ‘Taking Medicines – some questions and 25 answers about side effects’. The leaflet has 8 questions and short answers to 26 these questions. Patients are advised that they should receive a patient 27 information leaflet with their medicines, and to ask their doctor, pharmacist or 28 NHS direct if they have further queries. The questions in the leaflet are: 29 1. What do medicines do? 30 2. Will my medicine cause side-effects? Medicines concordance: full guideline DRAFT (July 2008) page 126 of 373 DRAFT FOR CONSULTATION 1 3. What is meant by a common or rare side effect? 2 3 4. How much medicine should I take? 4 5. How can I reduce the risk of side effects? 5 6. Do side effects always come on straight away? 6 7 7. What should I do if I feel unwell after I take my medicine? 8 9 8. Will my medication affect my lifestyle? 10 11 12 1 Always Read the Leaflet: Report of the Committee on Safety of Medicines Working Group on Patient Information 13 (2005). The Stationery Office, London. 14 2 Berry, D.C. (1997) Psychol Health. As referenced in Always Read the Leaflet: Report of the Committee on Safety of 15 Medicines Working Group on Patient Information 16 17 4 The ‘Medicine use review: Understand your medicine’ (2005) NHS Report by MHRA and ASK. Medicines concordance: full guideline DRAFT (July 2008) page 127 of 373 DRAFT FOR CONSULTATION 1 2 4.11 Key clinical Question: Mental capacity narrative 3 Some concern was expressed by the GDG about the potential conflicts 4 between respecting autonomy of the patient and the duty of care felt by 5 practitioners to patients. The GDG discussed the importance of professionals 6 codes of conduct and the legal system in protecting both patients and 7 professionals. The narrative below brings together the principles discussed 8 and is a repeat of section 3.4 in chapter 3. 9 4.11.1 **Roles and responsibilities of patients and health care 10 professionals 11 Concern has been expressed by practitioners that sharing decisions with 12 patients may conflict with their duty of care to patients or their legal or ethical 13 obligations (Stevenson, F. A., 2003). While the UK General Medical Council 14 (GMC) (2001)9 considered one of the key duties of a doctor to ‘respect the 15 rights of patients to be fully involved in decisions about their care’ for many 16 clinicians there is a legitimate area of concern or indeed of conflict between 17 respect for autonomy of the patient and the duty of beneficence when a 18 clinician feels uncomfortable about a patient’s wishes. The GMC (2008)10 has 19 recently updated guidance for doctors about patient consent which makes 20 explicit the right of competent patients to make decisions about their own 21 healthcare. The guidance emphasises the need for doctors to maximize 22 opportunities and patients abilities to make decisions for themselves and that 23 doctors must respect competent patients’ decisions even if they do not agree 24 with them. Doctors do not have to provide a treatment to patients which they 25 believe is not in the patients’ interest but must under such circumstances 26 explain their reasons and other options, including a second opinion to the 27 patient. It remains important however for health care professionals to do their 28 utmost to ensure that patients’ choices are informed choices as outlined 29 above. 9 http://www.gmc-uk.org/publications/annual_reports/review_archive/report2002.pdf http://www.gmc-uk.org/publications/business_plans/Business_Plan_2008.pdf 10 Medicines concordance: full guideline DRAFT (July 2008) page 128 of 373 DRAFT FOR CONSULTATION 1 The Mental Capacity Act (2005) 11 governs the making of decisions for people 2 who lack capacity in England and Wales. Under the Act doctors are advised 3 that they must work on the presumption that every adult patient has the 4 capacity to make decisions about their care, and to decide whether to agree 5 to, or refuse, an examination, investigation or treatment. A patient is regarded 6 as lacking capacity once it is clear that, having been given all appropriate help 7 and support, they cannot understand, retain, use or weigh up the information 8 needed to make that decision, or communicate their wishes12. 9 Doctors are advised that assumptions must not be made that a patient lacks 10 capacity to make a decision solely because of their age, disability, 11 appearance, behaviour, medical condition (including mental illness), their 12 beliefs, their apparent inability to communicate, or the fact that they make a 13 decision that health professionals disagree with13.** 14 11 12 http://www.opsi.gov.uk/ACTS/acts2005/pdf/ukpga_20050009_en.pdf http://www.gmc-uk.org/publications/business_plans/Business_Plan_2008.pdf 13 http://www.gmc-uk.org/publications/business_plans/Business_Plan_2008.pdf Medicines concordance: full guideline DRAFT (July 2008) page 129 of 373 DRAFT FOR CONSULTATION 1 2 4.12 Key Clinical Question: What information about shared 3 decision making and adherence should be recorded in 4 patients’ notes? Related Evidence references statements Evidence into recommendations (summary of evidence) . The GDG considered it important that a record is kept of discussions between health care professionals and patients about medicines. Prescribing and taking of medicines is a long term dynamic process which may involve multiple interactions between health care professionals and patients. Good record keeping aids continuity of care by providing information for healthcare professionals to review the discussion they or other health care professionals have had with patients about their medicines. The GDG made recommendations based on professional opinion. 5 6 Medicines concordance: full guideline DRAFT (July 2008) page 130 of 373 DRAFT FOR CONSULTATION 1 4.13 Key Clinical Question: What tools are available to 2 support the patient in reaching an informed decision? 3 How effective are these tools? 4 Related References Evidence Statements (summary Evidence into recommendations of evidence) including O’Connor (2003) 1. One systematic review found The literature review found a (O'Connor, A. M., that there a variety of decision aids number of systematic reviews Stacey, D., Entwistle, used in studies - audio guides, CD- concerning decision aids and their V. et al , 2003) ROMs, web-based, interactive use. The results of the trials video-disc, lecture and handouts. primarily related to decisional conflict, satisfaction, involvement i O’Connor (2003) 2. One systematic review found decision and participation with little (O'Connor, A. M., that decision aids led to greater effect on health outcomes overall. Stacey, D., Entwistle, knowledge, realistic expectations, The GDG considered the evidence V. et al , 2003) lower decisional conflict from supportive of the importance of feeling informed, more active in information in a variety of formats decision making and less patients but did not feel it indecision after the intervention. appropriate to make specific recommendations regarding O’Connor (2003) 3. Decision aids reduced decisional (O'Connor, A. M., conflict (1 systematic review, 4 decisions aids. Stacey, D., Entwistle, RCTs). V. et al , 2003); Montgomery (2003) (Montgomery, A. A., Fahey, T., and Peters, T. J., 2003); Weymiller (2007) (Weymiller, A. J., Medicines concordance: full guideline DRAFT (July 2008) page 131 of 373 DRAFT FOR CONSULTATION Montori, V. M., Jones, L. A. et al , 2007); Thomson (2007) (Thomson, R. G., Eccles, M. P., Steen, I. N. et al , 2007); Oakley (2006) (Oakley, S. and Walley, T., 2006) *; O’Connor (2003) 4. Decision aids improved patient (O'Connor, A. M., knowledge (1 systematic review, 3 Stacey, D., Entwistle, RCTs). V. et al , 2003); Montgomery (2003) (Montgomery, A. A., Fahey, T., and Peters, T. J., 2003); Weymiller (2007) (Weymiller, A. J., Montori, V. M., Jones, L. A. et al , 2007); Thomson (2007) (Thomson, R. G., Eccles, M. P., Steen, I. N. et al , 2007) O’Connor (2003) 5. One systematic review found (O'Connor, A. M., that simpler decision aids Stacey, D., Entwistle, compared to detailed decision aids V. et al , 2003) showed a significant improvement in knowledge, more realistic expectations and greater Medicines concordance: full guideline DRAFT (July 2008) page 132 of 373 DRAFT FOR CONSULTATION agreement between values and choices. Montgomery (2003) 6. One RCT found that decision {2257} analysis decreased decisional conflict more than a video/leaflet intervention. Fraenkel (2007) 7. One RCT found that an (Fraenkel, Liana, interactive computer tool which Rabidou, Nicole, generated personalised feedback Wittink, Dick et al , significantly improved decisional 2007) self-efficacy and preparedness to participate in decision-making, with greatest benefit for older adults. Oakley (2006) 8. One RCT found that a workshop (Oakley, S. and plus a decision aid (identifying own Walley, T., 2006) risk and pros and cons) and worksheet did not significantly improve adherence and although were initially satisfied with the information on medicines this was non-significant by end of trial. Hamann (2007) 9. One RCT found that a decision (Hamann, Johannes, aid and planned talk with doctor Cohen, Rudolf, reduced hospitalisation for Leucht, Stefan et al , schizophrenic outpatients. However 2007) those who showed a higher preference for autonomy and better knowledge showed a significantly higher re-hospitalisation rate. Medicines concordance: full guideline DRAFT (July 2008) page 133 of 373 DRAFT FOR CONSULTATION 1 *This was compared to baseline, as was not studied if it differed from the 2 control group. 3 4.13.1 Methods of evidence review 4 This paper includes a narrative summary of the included evidence, structured 5 according to the category of the intervention, following the agreed reviewing 6 protocol: 7 Types of studies: Systematic Reviews or Randomized controlled trials 8 (RCTs) of tools to help the patient reach a decision (decision aids). 9 Types of participants: people prescribed medication for a medical condition 10 faced with a decision. 11 Duration of studies: No time limit specified. 12 Types of interventions: any interventions which aid the patient in making an 13 informed decision. 14 Types of outcome measures: types of decision aid; patient outcomes: 15 decisional conflict, patient knowledge, self-efficacy. 16 17 It should be noted that the remit of the guideline is for conditions with 18 prescribed medication and this excludes conditions which require 19 chemotherapy or screening. All RCTs are within this remit, however many of 20 the systematic reviews included populations outside the remit, this is noted 21 where applicable. 22 4.13.2 Evidence review 23 O’Connor (2003) (O'Connor, A. M., Stacey, D., Entwistle, V. et al , 2003) 24 conducted a Cochrane Collaboration systematic review of decision aids for 25 people facing health treatment or screening. They included studies that were 26 RCTs; involving those over age of 14 making decisions about screening or 27 treatment options for themselves, a child, or significant other; including 28 decision aids and looking at whether the decision aids achieved their 29 objectives. They found 131 decision aids developed in the within the previous Medicines concordance: full guideline DRAFT (July 2008) page 134 of 373 DRAFT FOR CONSULTATION 1 five years, 94 were web-based, 14 paper-based and 12 videos with print 2 resources, eight were audio-guided print resources, two were CD-ROMs and 3 one a web-based with workbook. Most of the decision aids were intended for 4 use before counselling. Decision aids were better in terms of greater 5 knowledge, realistic expectations, lower decisional conflict related to feeling 6 informed, increased proportion of people active in decision making and 7 reduced proportion of people who remained undecided post intervention. 8 Simpler decision aids were proven to have more significant improvement than 9 more detailed decision aids in knowledge, more realistic expectations and 10 greater agreement between values and choices. There was no improvement 11 compared to comparisons for affecting satisfaction with decision making, 12 anxiety, and health outcomes. 13 Only a few of the studies in this systematic review were relevant to the 14 guideline as the majority of studies included were surgery, screening, or other 15 populations not included in the medicines concordance remit. The studies that 16 were relative to the guideline were seven trials of hormone replacement 17 therapy* (audioguides, booklet and interactive videodisc, mix of lecture and 18 handouts); two trials involving Ischaemic heart disease**, interactive videodisc 19 and a videocassette; and one study of atrial fibrillation treatment*** 20 (audioguide deciding whether to change from aspirin to warfarin). All studies 21 except five of the HRT trials compared a decision aid with usual care. The 22 other five studies compared a multiple element design with a simple decision 23 aid. 24 *McBride 2002, Murray HRT 2001, O’Connor, 1998-RCT, Dodin 2001, 25 O’Connor Wells 1999, Rothert 1997, Rostrom 2002.**Morgan 1997, Bernstein 26 1998.***Man-Son-Hing 1999. 27 Montgomery (2003) (Montgomery, A. A., Fahey, T., and Peters, T. J., 2003) 28 conducted a factorial randomized control trial to evaluate two interventions to 29 help hypertensive patients decide whether to start drug therapy to reduce 30 blood pressure. This was carried out in 21 general practices in SW England 31 with 217 patients aged 32 to 80 years (mean age 59 years) newly diagnosed 32 with hypertension. Patients were allocated to decision analysis or no decision Medicines concordance: full guideline DRAFT (July 2008) page 135 of 373 DRAFT FOR CONSULTATION 1 analysis, they were then further randomised to video/leaflet or no video/leaflet 2 groups. The decision analysis in this case was a decision tree for 3 hypertension which used a computerised self-completed interview to assess 4 patients’ utilities with minimal input from researcher and absolute 5 cardiovascular risk calculated. The video was informational about high blood 6 pressure. The information booklet was four fact sheets from the British 7 Hypertension Society. The primary outcome was the total score on the 8 decisional conflict scale, a questionnaire measuring how uncertain about the 9 course of action to take and factors which could be changed that lead to the 10 uncertainty. Secondary outcomes were subscales of the Decisional Conflict 11 Scale and intentions about starting treatment, state anxiety, knowledge of 12 hypertension, actual treatment decision. 13 Both interventions successfully reduced patients’ total decisional conflict at 14 follow-up. Decision analysis decreased the decisional conflict more than the 15 video/leaflet. Total decisional conflict mean for decision analysis was 27.6 16 (SD=12.1), no decision analysis 38.9(18.3) adjusted difference 95%CI -9.4 (- 17 13.0 to -5.8) p<0.001; video/leaflet 30.3 (13.4) and no video/leaflet was 18 36.8(18.8), 95% CI -4.2(-7.8 to -0.6), p=0.021. The Decisional conflict 19 subscales showed a clear reduction in three of the five subscales - 20 uninformed 23.7 (11.8) compared to no decision analysis 40.7 (23.1) adjusted 21 difference 95% CI -15.7 (-20.2 to -11.2), unclear values 28.4 (14.7) vs 43.8 22 (24.3) adjusted difference -13.1 (-18.0 to -8.1) and unsupported 24.4 (13.4 vs 23 34.8 (18.3) adjusted difference -8.7 (-12.8 to -4.7) and some evidence for 24 reduction in uncertainty and no evidence for decision quality. The video/leaflet 25 intervention showed no evidence in these last two subscales and there was 26 only clear evidence on the uninformed subscale. For the intention to start 27 treatment when followed up the adjusted risk ration (95% CI): Yes versus 28 unsure 1.19 (0.59 to 2.40) for decision analysis and 1.80 (0.89 to 3.63) for the 29 video/leaflet. No versus unsure 3.15 (0.91 to 10.98) and 0.52 (0.15 to 1.77) 30 respectively. The overall p values were 0.09 and 0.17 respectively. Actual 31 prescription of medication was not different for either intervention or controls. 32 There was a suggestion (p=0.055) that anxiety may be reduced by decision 33 analysis although the evidence there was weak and no evidence of this for the Medicines concordance: full guideline DRAFT (July 2008) page 136 of 373 DRAFT FOR CONSULTATION 1 video/leaflet intervention. Both interventions significantly increased knowledge 2 of hypertension. Those who received both interventions had the lowest 3 decisional conflict (27.1 compared with 28.2 and 33.3 and 44.2 for decision 4 analysis only, video/leaflet and control). They had a high knowledge score – 5 the same as video/leaflet. Within the regression models there was a 6 significant (antagonistic) interaction between decision analysis and 7 video/leaflet, so the effect of each was reduced by the presence of the other 8 (interaction coefficient 12.5, 95% CI 5.4 to 19.5, p=0.001 for decisional conflict 9 and -9.1, 95% CI-16.3 to -1.9, p=0.013 for knowledge. This study was 10 followed up in 2005 by Emmett, who found that there was no evidence of any 11 difference in blood pressure, cardiovascular disease risk for either intervention 12 or between them. There were also no effects on medication prescribing, self- 13 reported adherence, consulting behaviour or management changes. 14 Weymiller (2007) (Weymiller, A. J., Montori, V. M., Jones, L. A. et al , 2007) 15 conducted an RCT study of the effect of a decision aid on statin drug 16 treatment decision making. The study was conducted in a diabetes referral 17 clinic in Minnesota, USA. 98 participants were included with a mean age of 18 64 (12) for the decision aid group and 66(8) for the control group. Participants 19 were randomized to either receive usual care plus a standard pamphlet on 20 cholesterol management or a statin choice decision aid. The decision aid 21 included name, cardiovascular risk factors, an estimated level of 22 cardiovascular risk (3 levels) and the absolute risk reduction with statins and 23 the potential disadvantages of taking them. A question prompted patients to 24 express whether they were ready to make a decision and if they wanted to 25 take statins, discuss the issues with their clinician or other. After the 26 consultation the participants were given a questionnaire to complete. The 27 outcomes of interest were improvement in patient knowledge and reduction in 28 decisional conflict. 74% would recommend the decision aid to others 29 compared to 53% of control patients recommending the pamphlet, (OR 2.6, 30 95% CI 0.8-8.0), 68% would want to receive similar support for future 31 decisions compared to 58% of control patients (OR 1.5, 95% CI 0.6-3.8). 32 Those receiving the decision aid had higher knowledge scores than the 33 control group and those allocated to receive the intervention during the visit Medicines concordance: full guideline DRAFT (July 2008) page 137 of 373 DRAFT FOR CONSULTATION 1 achieved better knowledge than those who received before the consultation. 2 The intervention group had significantly less decisional conflict afterwards 3 than the control group, and at 3 months (although not statistically significant. 4 Those in the DA group felt more informed. 30% of the DA group (6/7 were 5 from the DA during the visit group) decided to start statin therapy immediately 6 after, compared to 21% of the control group. At 3 months 63% of the DA 7 group and 63% of the control group reported taking statins (OR, 1.4, 95%CI, 8 0.8-2.4). Overall, there was no difference in adherence to patient choice at 3 9 months. 10 Thomson (2007) (Thomson, R. G., Eccles, M. P., Steen, I. N. et al , 2007) 11 conducted a randomized controlled trial of a decision aid for anti-thrombotic 12 treatment of patients with atrial fibrillation. 109 patients aged over 60 years 13 from 40 general practices in the Northeast of England. The intervention 14 involved the doctors in the clinic delivering either the decision aid or guidelines 15 to the patient. The decision aid was a computerized aid which presented the 16 individualized benefits and potential harms of warfarin treatment and 17 participants were invited to weigh up the advantages and disadvantages of 18 treatment before coming to a shared decision with the doctor. This involved 19 personalized risk assessment using the Framingham equation for stroke risk 20 and the presentation used graphical and numerical formats followed by a 21 shared decision-making component. The evidence-based guidelines group 22 applied decision analysis derived guidelines according to the participants’ risk 23 factor profile and the recommendation made directly to the participant by the 24 clinic doctor. The primary outcome measure was decision conflict scale 25 immediately after the consultation. Secondary outcomes were anxiety, 26 knowledge and decision-making preferences. The computerized decision aid 27 group had lower decision conflict immediately after the clinic (mean -0.18, 28 95%CI -0.34 to -0.01) and mean -0.15 (-0.37 to 0.06) at three month follow- 29 up. Both groups had less decision conflict after the consultation but the 30 difference between groups was significant at 5% level. Subscales suggested 31 this was due to feeling better informed and clearer of their personal values for 32 the risks and benefits of alternative options. The reduction in anxiety fell 33 significantly but was not different between groups. Knowledge scores Medicines concordance: full guideline DRAFT (July 2008) page 138 of 373 DRAFT FOR CONSULTATION 1 improved slightly after the consultation but at three months were back at 2 baseline level. Participants in the decision aid group were less likely to start 3 warfarin than those in the guideline arm (39/53, 73.6%) compared to 4 guidelines (50/56, 81.7%), RR 0.82, 95% CI 0.68 to 0.99, this was however 5 almost completely due to participants not already on warfarin, here the 6 differences was 4/6, 25% compared to guidelines 15/16, 93.8%, RR 0.27, 7 95%CI 0.11 to 0.63. There was no difference in health outcomes 3 months 8 after the clinic. 9 Fraenkel (2007) (Fraenkel, Liana, Rabidou, Nicole, Wittink, Dick et al , 2007) 10 conducted a randomized control trial which tested the efficacy of a computer 11 tool to improve informed decision-making for patients with knee pain in an 12 outpatient clinic. The trial was conducted in a primary care outpatient clinic in 13 the USA. 87 Participants over the age of 60 were randomised to receive an 14 Arthritis Foundation information pamphlet (control group) or to perform an 15 Adaptive Conjoint Analysis (ACA)(intervention group). The ACA is an 16 interactive computer tool which could generate immediate feedback to the 17 participant and helps them construct treatment preferences by means of 18 tradeoffs by rating tasks. The treatment characteristics in the ACA task 19 included route of administration, likelihood of expected benefit, and risk of 20 adverse effects. Questionnaires were given to assess outcomes. Decisional 21 self-efficacy and preparedness to participate in decision-making remained 22 significantly higher in the intervention group than the control group after 23 controlling for race and health status. Arthritis self-efficacy was of borderline 24 significance. Outcomes by age and education suggest older adults may be the 25 most likely to benefit. 98% of the participants thought the ACA task was very 26 easy/easy to do. The patients in the intervention group had greater self- 27 confidence in their ability, felt more prepared to participate in shared decision- 28 making and felt they had greater self-efficacy over arthritis than the control 29 group. 30 Oakley (2006) (Oakley, S. and Walley, T., 2006) assessed the effectiveness 31 of a decision aid on patient adherence to oral bisphosphonate medication. 32 They conducted an RCT with postmenopausal women over the age of 65 who Medicines concordance: full guideline DRAFT (July 2008) page 139 of 373 DRAFT FOR CONSULTATION 1 were diagnosed with osteoporosis. 33 participants were included in the study, 2 with a mean age of 77 years (range 61-90). The intervention group attended 3 an osteoporosis workshop, where they were given a decision aid and group 4 discussion on a worksheet. This included working through the decision- 5 making process and identifying their own lifetime risk of hip fracture, and 6 personal and family health issues which may influence their decision in 7 addition to self-care options already using or willing to try. They were to 8 consider pros and cons and their personal values regarding therapy and noted 9 any questions for the GP on a worksheet. Two weeks later they returned with 10 their worksheets for a follow up with a GP osteoporosis specialist. 11 Questionnaires were then administered to establish compliance. The 12 difference in adherence improvement was not significant. Satisfaction with 13 information about medicines improved initially in the intervention group but 14 reduced by the final questionnaire to non-significance. The decisional conflict 15 scale showed a reduction in decisional conflict from baseline measures (total 16 DCS score pre-intervention, median 2.5 (1.8-3.4) v 2.0 (1.0-2.4) post- 17 intervention, p<0.001, however this cannot be compared to the control group 18 as they were not assessed for this measure. The decision aid improved their 19 ability to make a decision about which treatment was best and to discuss their 20 medication with the GP but had no obvious effect on adherence. 21 Hamann (2007) {3478} conducted an RCT to assess whether shared decision 22 making in antipsychotic drug choice would influence long-term outcome. 86 23 patients with a diagnosis of schizophrenia were included and followed up, 24 58% female with mean age of 38 years (SD 11.4), mean duration of illness 9.2 25 years (SD 8.5). The study was conducted in Germany. The intervention group 26 received a decision aid program and the control group received usual care. 27 The decision aid was a 16 page booklet with the pros and cons of oral versus 28 depot formulation, first versus second generation antipsychotics, 29 psychoeducation and a type of socio-therapeutic intervention. 24 hours after 30 working through the booklet with the trained nurses the patients consulted 31 with the psychiatrists on further treatment. The patients then filled out 32 questionnaires relating to patient involvement, satisfaction and 33 psychopathology. Patients were followed up at 6 and 18 months after Medicines concordance: full guideline DRAFT (July 2008) page 140 of 373 DRAFT FOR CONSULTATION 1 discharge from hospital. Univariate analysis found no significant differences 2 between groups. When multivariate analysis was conducted to control for the 3 re-hospitalisation rate it showed that there was a positive trend for the 4 decision aid and planned talk in reducing rehospitalisation. Higher 5 participation preferences (OR= 1.06, p=0.03) and better knowledge (OR 6 =1.23, p=0.03) rates significantly predicted rehospitalisation. No other effects 7 were shown. 8 Medicines concordance: full guideline DRAFT (July 2008) page 141 of 373 DRAFT FOR CONSULTATION 1 2 4.14 Key Clinical Question: How can a practitioner elicit 3 whether a patient agrees with the prescription 4 recommended by the practitioner? Related Evidence statements references (summary of evidence) Evidence into recommendations 1. No evidence was found No tools designed for use in clinical practice on specific clinical tools were found. The GDG used the information that can aid the from the research studies and their practitioner in eliciting professional opinion to make whether patient agrees recommendations in relation to elicitation of with the recommended agreement with decision to prescribe. prescription 5 6 *e.g. relative value place on different outcomes, trade-offs between benefits and harms, uncertainties in Evidence, likely impact if implements, feasibility, patient perspective Medicines concordance: full guideline DRAFT (July 2008) page 142 of 373 DRAFT FOR CONSULTATION Methods of the evidence review 1 4.14.1 2 This paper includes a narrative summary of the included evidence, structured 3 according to the category of the intervention, following the agreed reviewing 4 protocol: 5 Types of studies – no restrictions on study design. 6 Types of participants - people prescribed medication for a medical condition. 7 Duration of studies - No time limit specified. 8 Types of interventions - Any interventions intended to assess tools that can 9 aid a practitioner in eliciting whether a patient agrees with recommended 10 prescription 11 Types of outcome measures – Agreement with prescription, patient 12 satisfaction with issues prescription. 13 4.14.2 Evidence review 14 We did not find any study that assessed a clinical tool that could aid a 15 practitioner in eliciting whether a patient agrees with the recommended 16 prescription. We did find one prospective observational study (Bikowski 2001) 17 (Bikowski, R. M., Ripsin, C. M., and Lorraine, V. L., 2001) that aimed to 18 characterize the degree of disparity between Physicians’ perceptions of older 19 patients’ medication regime and patients’ perceptions of their regime within a 20 community family medicine residency setting. 21 Eligible patients were age 65 and older, non-institutionalised, visiting the clinic 22 on the index day for a routine visit, had seen the index physician at least three 23 times in the past calendar year, and, by brief review of the medication flow 24 sheet, were taking at least four prescriptions medications. Study sample 25 comprised 50 physician-patient pairs. 26 Physicians were given the patients chart’s, with a request to complete a 27 questionnaire that asked for information on all prescriptions and 28 nonprescription medications, with dosages and frequencies of administration. Medicines concordance: full guideline DRAFT (July 2008) page 143 of 373 DRAFT FOR CONSULTATION 1 Patients were interviewed at home by first year medical students who 2 received specific training for the study. 3 Percentage congruence- defined as agreement between physician and patient 4 regarding all prescriptions medications, dosages and frequency, was 5 calculated for each pair. Complete congruence was showed for 14% of the 50 6 physician-patient pairs; 74% had at least one medication that either the 7 physician was unaware the patient was taking or the physician thought the 8 patients was taking but that was not part of the patients regime; 12% had 9 dose and/or frequency differences, however they agreed upon the 10 medications in the regime. 11 Antihypertensives were the most commonly prescribed medication, 12 accounting for 36% of the total. The highest congruence was found for 13 diabetic and other endocrine drugs. Pain medications and gastrointestinal 14 medications showed the lowest congruence 15 Medicines concordance: full guideline DRAFT (July 2008) page 144 of 373 DRAFT FOR CONSULTATION 1 4.15 Key Clinical Question: What aspects of consultation style increase patient involvement in decision-making? 2 3 Related References Evidence Statements (summary Evidence into recommendations of evidence) including McKinstry (2006) 1. One high quality systematic The GDG were aware that there is (McKinstry, B., review found that there is anecdotal evidence that Ashcroft, R. E., Car, insufficient evidence to conclude practitioners and patients report J. et al , 2006) that any intervention may increase that the quality of the doctor- or decrease trust in physicians. patient relationship is important in decision –making. The quality of the practitioner-patient relationship was reported as being important in Van Dam (2003) 2. One systematic review of RCTs some studies of patients medicine (van-Dam, Henk A., found that supporting patient taking behaviour as outlined in 2003) participation in diabetes care and chapter 3. The quality of a self-care behaviour (i.e. by practitioner –patient relationship is assistant-guided patient however likely to be influenced by preparation for visits to doctors, a number of factors that relate to empowering group education, previous consultations and group consultations, or automated problem under discussion. The telephone management) is more consultation skill of an individual effective than changing provider practitioner is however also likely consultation style in improving to be important regardless of patient self-care and diabetes length of professional-patient outcomes. relationship. It was considered tha Edwards (2004) 3. One RCT reported significant (Edwards, A., Elwyn, effects of the research clinic group G., Hood, K. et al , (which provided more consultation 2004) time) in confidence in decision and expectation to adhere to chosen the level of trust between practitioner and patient may be a key factor in this relationship and the GDG wished to review whethe this could be specifically increased in practitioner-patient encounters. Medicines concordance: full guideline DRAFT (July 2008) page 145 of 373 DRAFT FOR CONSULTATION treatments. The evidence from a recent systematic review suggested that there is insufficient evidence to Cohen (2004) 4. Two studies from the same recommend any specific (Cohen, D., Longo, randomised controlled trial found intervention. The GDG felt that M. F., Hood, K. et al , that training GPs in SDM or consultation style should be 2004); Edwards combined with risk communication tailored to individual patients to (2004) (Edwards, A., yielded conflicting results in the allow full communication. Elwyn, G., Hood, K. probability of a prescription being et al , 2004) issued to patients. Cohen (2004) 5. Two studies from the same (Cohen, D., Longo, randomised controlled trial found M. F., Hood, K. et al , that training GPs in SDM or 2004); Edwards combined with risk communication (2004) (Edwards, A., yielded no effect on the probability Elwyn, G., Hood, K. of investigations, referrals or follow- et al , 2004) up GP visits for any of the conditions. Savage (1990) 6. One RCT found that a directing (Savage, R. and style of consultation yielded Armstrong, D., 1990) significantly higher levels of satisfaction on almost all the outcome measures compared to a sharing style. This was particularly relevant for patients with physical problems. Shields (2005) 7. One RCT reported a statistically (Shields, C. G., significant likelihood of a physician Epstein, R. M., to promote collaboration in Medicines concordance: full guideline DRAFT (July 2008) page 146 of 373 DRAFT FOR CONSULTATION Fiscella, K. et al , treatment decision-making with 2005) patients rather than with companions of the patients. Shields (2005) 8. One RCT reported a statistically (Shields, C. G., significant responsiveness of a Epstein, R. M., physician to explore the disease Fiscella, K. et al , and illness when the issues were 2005) raised by the patient compared with the companion of the patient. 1 2 4.15.1 Methods of the evidence review 3 This paper includes a narrative summary of the included evidence, structured 4 according to the category of the intervention, following the agreed reviewing 5 protocol: 6 Types of studies: Systematic Reviews or Randomised controlled trials 7 (RCTs) that focus on aspects of consultation style that may increase patient 8 involvement in decision-making. 9 Types of participants: people prescribed medication for a medical condition 10 faced with a decision. 11 Duration of studies: No time limit specified. 12 Types of interventions: any interventions which assess which aspects of the 13 consultation style may increase patient involvement in decision-making. 14 Types of outcome measures: Patient-centred communication in the 15 consultation; Consultation process outcomes: patient involvement, question 16 asking, preparedness; Patient care outcomes: satisfaction, knowledge, self- 17 efficacy, type of information. 18 Medicines concordance: full guideline DRAFT (July 2008) page 147 of 373 DRAFT FOR CONSULTATION 1 It should be noted that the remit is for conditions with prescribed medication 2 and this excludes conditions which require chemotherapy or screening. All 3 RCTs are within this remit, however many of the systematic reviews included 4 populations outside the remit, this is noted where applicable. 5 6 4.15.2 Evidence review 7 Our searches retrieved two systematic review and 4 RCTs that were 8 considered relevant to the key clinical question. 9 All the studies looked at the patient-provider interaction, either by exploring 10 the impact of different provider styles or by focusing on patient behaviour 11 changes. 12 13 McKinstry (2006) (McKinstry, B., Ashcroft, R. E., Car, J. et al , 2006) 14 conducted a Cochrane review of interventions to improve the trust of patients 15 in their doctors. They searched 10 databases including Cochrane Central 16 Register of CTs, Medline and Embase. Inclusion criteria for studies was 17 RCTs, CCTS, controlled before and after studies and interrupted time series 18 studies. The interventions were any that influenced patients trust in their 19 doctors, or where trust was an outcome of an intervention. The participants 20 were Doctors; adults or children using healthcare or those related to them. 21 Outcome measures were increase or decrease in patients’ trust and the 22 components of trust; other healthcare behaviours; Health status and well- 23 being; Use of resources; Satisfaction with care; Perception of doctors’ 24 communication skills; perception of doctors’ humanistic attributes; perception 25 regarding patients’ trust; perceptions of doctors trustworthiness. Studies were 26 excluded if they did not measure change in trust or not the right type of study. 27 Two authors independently assessed whether the titles and abstracts were 28 relevant and four authors assessed the retrieved articles for inclusion. Two 29 authors assessed the quality of each study according to EPOC criteria. A 30 multidisciplinary advisory group was set up to assess whether there was 31 anything that had been left out of the review. 2099 titles and abstracts were 32 found, five met all the criteria, but two of these referred to the same study and 33 one had insufficient data points before and after the intervention, therefore Medicines concordance: full guideline DRAFT (July 2008) page 148 of 373 DRAFT FOR CONSULTATION 1 leaving three studies. Two of the studies had a primary aim of the impact of 2 the intervention on patient trust. Thom (1999) coached doctors in behaviours 3 known to be associated with trust and Hall (2002) looked at the impact of 4 disclosing financial incentives physicians receive for compliance with 5 managed health care protocols on the trust patients had in physicians. In the 6 third study (Thompson 2001) trust was a secondary outcome and compared 7 the impact of three different types of induction visit for new patients of an 8 HMO to those who received no intervention, the trust was in any health care 9 professional. The review detailed the quality of studies including allocation 10 concealment, blinding and protection from contamination. On assessment of 11 study quality all three being RCTs provided baseline measures and within and 12 between group differences for measures. Thom (1999) used computer 13 allocation to groups but it was not clear if the researcher was blind to this 14 allocation. The interviewer was blind to the status of the physician but it was 15 unclear if the patients were blinded. Hall (2002) conducted a stratified random 16 sample study and used a computer for allocation with no input from 17 researchers. The interviewers were blind to the patients’ status but the 18 patients were aware of their own status. Thompson (2001) did not report how 19 the randomisation was applied. Patients were aware of their status but it is not 20 clear if interviewers were blind to the status of the interviewees. The study by 21 Thom (1999) on showed no effect on trust (74.4 for the intervention and 76.2 22 for the control group, non-significant). Satisfaction or humaneness scores 23 were non-significant also. Hall found a small increase in trust for both groups 24 from baseline when adjusted this was a 1.4% increase in physician trust 25 (p<0.05). Thompson (2001) found the trust in the health plan health 26 professionals rose significantly following the enrolment visit with health 27 personnel compared to control group p<0.001). The author concluded that 28 there is insufficient evidence to conclude that any intervention may increase or 29 decrease trust in physicians. 30 31 Van Dam (2003) (van-Dam, Henk A., 2003) developed a systematic review of 32 RCTs that looked at the effects of interventions on provider-patient interaction 33 on patient diabetes health behaviour, patient self-care, delivered diabetes Medicines concordance: full guideline DRAFT (July 2008) page 149 of 373 DRAFT FOR CONSULTATION 1 care and health outcomes, and to disentangle those that are the most 2 effective. Eight studies were included after a rigorous methodological quality 3 assessment, and these showed different interventions on different levels of 4 the provider-patient interaction in diabetes care. Four studies focused on 5 provided consulting behaviour modifications (studies 1-4), and four studies 6 focused directly on patient behaviour change (studies 5-8). 7 All studies were conducted in practical diabetes care, three in hospital 8 outpatient clinics and five in general practices. 9 The main findings suggest that the most effective interventions are those with 10 a direct approach to support patient participation (i.e. by assistant-guided 11 patient preparation for visits to doctors, empowering group education, group 12 consultations, or automated telephone management) in diabetes care and 13 self-care behaviour, while interventions which focus on change of provider 14 behaviour were less effective. Thus, the authors advocate a shift from the 15 traditional medical model to a more patient centred, patient participation and 16 empowerment paradigm of delivery of diabetes care. 17 The authors pointed out that the review did present some limitations, 18 illustrated by the small number of reviewed paper; the differences between the 19 studies; and the focus on RCTs. 20 21 Cohen (2004) (Cohen, D., Longo, M. F., Hood, K. et al , 2004) and Edwards 22 (2004) (Edwards, A., Elwyn, G., Hood, K. et al , 2004) conducted a cluster 23 randomized crossover trial with the aim to explore the costs of training GPs in 24 developing SDM competences and in the use of risk communication (RC) aids 25 and to evaluate the effects of such training on a range of service resource 26 variables. Edwards (2004) published the main trial results that focused on the 27 doctor patient interaction, patient outcomes and satisfaction with the decision. 28 Within each cluster, patients were also allocated randomly to consult with the 29 doctor at one of three points in the study. The study comprised three phases. 30 Phase 1 was pre-training. Phase 2 included training for half of the GPs and 31 the other half in RC. In phase 3, each GP received training in the other 32 element making them fully trained in both. The authors argued that in this 33 way, the design offered the greatest potential to gain understanding about the 34 effects of each form of training alone and in combination and if the sequence Medicines concordance: full guideline DRAFT (July 2008) page 150 of 373 DRAFT FOR CONSULTATION 1 of skill acquisition was important. A further randomisation allocated patients to 2 attend either in usual surgery time or in a research clinic- audiotaped, 3 including fewer interruptions and more time for each consultation (up to 15 4 minutes each). 5 SDM training involved GPs attending two workshops where standardized and 6 previously piloted skill development process was used. SDM competences 7 were described and demonstrated by means of consultation simulation and 8 pre-prepared scenarios involving the four study conditions. RC also involved 9 attendance at 2 workshops, and the aids consisted of tabulated data and 10 visuals displays of risk estimates for the four study conditions. Patients with 11 one of four conditions (menorrhagia, atrial fibrillation, menopausal symptoms 12 or prostatism) were invited by their GP to attend a “review consultation” to 13 discuss their continuing treatment. Twenty GPs from 20 different practices in 14 South Wales were recruited. Costs of training for both RC and SDM included 15 time of trainers, of those being trained and of the simulated patients used as 16 part of the training exercise. Information on prescriptions, investigations and 17 referrals was obtained from questionnaires completed by each clinician at the 18 review consultation. 19 Main results published by Cohen (2004) indicated that Training in SDM or 20 combined with RC significantly affected the probability of a prescription being 21 issue to women with menopausal symptoms and menorrhagia (despite RC 22 alone not having any effect). However, there was no significant change in 23 prescribing for patients with prostatism or atrial fibrillation. There was also no 24 effect on the probability of investigations, referrals or follow-up GP visits for 25 any of the conditions. Training cost was £1218 per GP, resulting in an 26 increase of cost of consultation by £2.89. 27 Edwards (2004) reported significant effects of the research clinic (which 28 provided more time) in confidence in decision (p<0.01) and expectation to 29 adhere to chosen treatments (p<0.05). Anxiety scores approached statistical 30 significance for the RC intervention, as did expectation to adhere to chosen 31 treatment for both interventions. No statistically significant effects of the risk 32 communication or SDM interventions were seen on the whole range of 33 patient-based outcomes. Medicines concordance: full guideline DRAFT (July 2008) page 151 of 373 DRAFT FOR CONSULTATION 1 Cohen (2004) concluded that due to the explanatory nature of the study, no 2 assessment could be made on how training could affect the length of a 3 consultation. 4 5 Savage (1990) (Savage, R. and Armstrong, D., 1990) sought to compare the 6 effects of a directing and sharing style of consultation by a GP on patient’s 7 satisfaction with the consultation in a deprived inner city area. Patients were 8 aged 16 to 75 years of age and were randomised to receive a directing or 9 sharing style in the part of the consultation regarding treatment, advice and 10 prognosis. Three hundred and fifty nine patients were randomised, however 11 120 patients failed to complete the assessment that took place a week later. 12 There were no significant differences in the mean length of consultations 13 between the two experimental groups. Patients who had the directing style of 14 consultation reported significantly higher levels of satisfaction on almost all the 15 outcome measures, and was particularly strong for patients with physical 16 problems (excellent explanation p<0.02; excellent understanding p=0.04). 17 There was no significant difference in the responses to the directing and 18 sharing styles in longer consultations, where the main treatment was advice 19 and among patients with psychological or chronic problems. Statistical 20 significance values were not reported. This study was conducted in England. 21 22 Shields (2005) (Shields, C. G., Epstein, R. M., Fiscella, K. et al , 2005) 23 evaluated the influence of accompanied visits on physician-patient 24 communication- particularly on patients centred communication. Thirty 25 patients included in the study were aged above 65 years, were not cognitively 26 impaired and had a companion who could accompany them to their next visit. 27 Companions were not assigned a specific role during the session and 28 physicians were not asked to conduct the sessions in any particular way. 29 There were no statistically significant differences between accompanied and 30 unaccompanied visits on the number of issues that patients raised, however 31 patients did raise more issues in unaccompanied visits. No statistically 32 significant differences were observed for levels of patient-centeredness, or 33 satisfaction, even if patients who were accompanied reported being slightly 34 more satisfied. Medicines concordance: full guideline DRAFT (July 2008) page 152 of 373 DRAFT FOR CONSULTATION 1 Physicians were more likely to promote collaboration in treatment decision 2 making with patients than with companions (p<0.0001). Physicians were also 3 more responsive to issues regarding exploring the disease and illness when 4 the issues were raised by the patient compared with the companion (p<0.03). 5 Medicines concordance: full guideline DRAFT (July 2008) page 153 of 373 DRAFT FOR CONSULTATION 1 4.16 Key Clinical Question: Do interventions to increase 2 patient involvement increase length of the 3 consultation? 4 Related References Evidence Statements (summary Evidence into of evidence) recommendations including Evidence from the UK McCann (1996) 1. Two RCTs from the UK found The GDG were (McCann-Simon, that interventions to increase concerned that Weinman John, participation in the consultation interventions to increase 1996); Middleton (leaflet and agenda form) patient involvement in the (2006) (Middleton, J. significantly increased consultation consultation might result F., McKinley, R. K., length. in longer consultations and Gillies, C. L., and have impact on 2006) service delivery more generally. While the Little (2004) (Little, 2. Two RCTs from the UK found P., Dorward, M., interventions to increase Warner, G. et al , participation in the consultation (a 2004) prompt to elicit patient concerns and a leaflet) did not significantly McLean (2004) increase consultation length. evidence indicated that simple interventions might result in increase in consultation length, this did not always occur (McLean, Malcolm which was of some and Armstrong, surprise to the GDG. The David, 2004) GDG considered it important to reassure clinicians that increasing Middleton (2006) 3. In one RCT that reported a (Middleton, J. F., significant increase in consultation patient involvement may not affect consultation Medicines concordance: full guideline DRAFT (July 2008) page 154 of 373 DRAFT FOR CONSULTATION McKinley, R. K., and length, the increase was more length but that even if it Gillies, C. L., 2006) pronounced when using an agenda did any extra time form than when using the agenda invested in the treatment form in combination with an process may have educational intervention. When benefits later on. using the educational intervention alone no difference was found. Little (2004) (Little, 4. One RCT reported that the use P., Dorward, M., of a general leaflet in the Warner, G. et al , consultation was significantly more 2004) effective in increasing satisfaction when consultations were shorter. Edwards (2004) 5. Adherence was increased in (Edwards, A., Elwyn, clinics where more time was G., Hood, K. et al , available than usual surgery times. 2004) Evidence including the rest of the world Kinnersley (2007) 6. Evidence from two systematic (Kinnersley, P., reviews and three RCTs suggests Edwards, A., Hood, that interventions designed to K. et al , 2007); improve patient participation in Harrington (2004) consultations does not increase (Harrington-Jane, overall length of consultations. One Noble Lorraine, review (Kinnersley 2007) found 14 2004); Cegala (2001) RCTs with no significant increase in Medicines concordance: full guideline DRAFT (July 2008) page 155 of 373 DRAFT FOR CONSULTATION in Wetzels review consultation length and 3 RCTs (2007) (Wetzels, R., with a significant increase Harmsen, M., van, (Hornberger, 1997; McCann 1996 Weel C. et al , 2007); and Middleton 2006 – 2 UK studies Loh (2007) (Loh, reported earlier). Andreas, Simon, Daniela, Wills, Celia E. et al , 2007); Hamann (2006) (Hamann, J., Langer, B., Winkler, V. et al , 2006) Hornberger (1997) in 7. In one of the RCTs that reported Kinnersley (2007) a significant increase in (Kinnersley, P., consultation length, this increase Edwards, A., Hood, was due to time spent discussing K. et al , 2007) diagnoses and physical examination Kinnersley (2007) 8. Written materials had a small (Kinnersley, P., and statistically significant increase Edwards, A., Hood, in consultation length compared to K. et al , 2007) coaching where the small increase was not significant.* Kinnersley (2007) 9. Interventions carried out (Kinnersley, P., immediately before the intervention Edwards, A., Hood, had a small and significant increase K. et al , 2007) in consultation length compared to those beforehand.** Medicines concordance: full guideline DRAFT (July 2008) page 156 of 373 DRAFT FOR CONSULTATION Kinnersley (2007) 10. There was no difference in (Kinnersley, P., consultation length between RCTs Edwards, A., Hood, with or without clinician training.*** K. et al , 2007) Hornberger in 11. One RCT (from Kinnersley Kinnersley (2007) review) reported that overall quality (Kinnersley, P., of care showed a positive Edwards, A., Hood, significance in the intervention K. et al , 2007) group which had a longer consultation time than the control group. 1 2 3 *This result is from a comparison of written materials and coaching for the consultation length of all studies which included written materials or coaching (thirteen), three of which were not relevant to the population of interest in this evidence review. 4 5 6 **This result is from a comparison of studies some time before consultation (2) and immediately before consultation (11), of which three of the immediately before consultation were not the relevant population. 7 8 9 ***This result included studies of Clinician training (2) compared to 12 studies where Clinicians were not trained. One of the Clinician training studies and three of the Clinicians not trained studies were not the relevant population Medicines concordance: full guideline DRAFT (July 2008) page 157 of 373 DRAFT FOR CONSULTATION 1 4.16.1 Methods of the evidence review 2 The aim of the literature review is to identify the most relevant, published 3 evidence to answer the key clinical questions generated by the GDG. Due to 4 time constraints, exhaustive systematic reviews (see the Methods of the 5 Cochrane Review) were not undertaken. However, the evidence reviews were 6 undertaken using systematic, transparent approaches following the Guidelines 7 Manual 2007 (www.nice.org.uk). 8 The titles and abstracts of records retrieved by the searches, suggested by 9 the GDG or submitted by stakeholders were scanned for relevance to the key 10 questions. Any potentially relevant publications were obtained in full text. 11 These were then reviewed to identify the most appropriate evidence to help 12 answer the key questions and to ensure that the recommendations are based 13 on the best available evidence. This process required four main tasks: 14 selection of relevant studies; assessment of study quality; synthesis of the 15 results; and grading of the evidence. 16 This paper includes a narrative summary of the included evidence, structured 17 according to the category of the intervention, following the agreed reviewing 18 protocol: 19 Types of studies: Systematic Reviews or Randomized controlled trials 20 (RCTs) that assess whether interventions to increase patient involvement 21 increase length in consultation. 22 Types of participants: people prescribed medication for a medical condition 23 faced with a decision. 24 Duration of studies: No time limit specified. 25 Types of interventions: any interventions which assess which aim to 26 increase patient involvement and include details of consultation length. 27 Types of outcome measures: Patient-centred communication in the 28 consultation; Consultation process outcomes: patient involvement, question Medicines concordance: full guideline DRAFT (July 2008) page 158 of 373 DRAFT FOR CONSULTATION 1 asking, preparedness; Patient care outcomes: satisfaction, knowledge, self- 2 efficacy in relation to consultation length. 3 4 It should be noted that the remit is for conditions with prescribed medication 5 and this excludes conditions which require chemotherapy or screening. All 6 RCTs are within this remit, however many of the systematic reviews included 7 populations outside the remit, this is noted where applicable 8 9 10 NOTE: this evidence review is an updated version containing systematic reviews retrieved from our searches. 11 4.16.2 Evidence review 12 This review was stratified to firstly present the RCTs research from the UK 13 and secondly present systematic reviews and RCTs which include research 14 from other areas of the world. 15 16 4.16.2.1 RCTS from the UK 17 Little (2004) (Little, P., Dorward, M., Warner, G. et al , 2004) conducted a 18 randomised control trial in the UK to assess the effect of leaflets in 19 empowering patients in primary care consultations. Participants were 20 randomised to four conditions: receipt of a general leaflet, depression leaflet, 21 both leaflets and no leaflets (control group). The general leaflet which asked 22 patients to list issues they wanted to raise and explained that the doctor 23 wanted them to ask questions, talk and discuss any problems of concern to 24 them. The depression leaflet listed symptoms of depression (without labelling 25 as such) and asking if had them and that the doctor would like to discuss 26 them. The only significant interaction was the increase in satisfaction for those 27 who received the general leaflet, the mean difference was 0.17 (95% CI 0.01 28 to 0.32, p=0.04). The general leaflet was significantly more effective when 29 consultations were shorter (leaflet 0.64, 95% CI 0.19 to 1.08; time 0.31, 0.0 to 30 0.06; interaction between both showed that consultations of 5, 8, and 10 mins 31 increased satisfaction by 14%, 10% and 7%). The leaflet overall caused a 32 small non-significant increase in consultation time. Medicines concordance: full guideline DRAFT (July 2008) page 159 of 373 DRAFT FOR CONSULTATION 1 2 Middleton (2006) (Middleton, J. F., McKinley, R. K., and Gillies, C. L., 2006) 3 conducted a randomised controlled trial in the UK of agenda forms completed 4 by the patient and doctors’ education about the agenda on the outcome of the 5 consultation. The intervention group were asked to think of a list of their 6 concerns, arrive five minutes earlier and bring spectacles and an interpreter if 7 those were required. The intervention doctors were given a one day 8 educational workshop to allow the doctors to have awareness of the patient 9 agenda model of the consultation. The model involved identifying the agenda 10 (ideas, concerns, expectations and reasoning). The doctors reflected on their 11 own agenda and negotiation of action with patients. Half of the patients in this 12 group filled in an agenda form in the preceding time before their consultation, 13 half did not. The control group was the GPs not given the educational 14 programme, and this was split into half the patients using the agenda form 15 and half not using it. The consultation length for the control group was 7.1 16 minutes (95% CI 6.5 to 7.7 minutes). The agenda form significantly increased 17 the duration of consultation by 0.9 minutes (0.3 to 1.5, p=0.004) and the 18 combined intervention by 1.9 minutes (1.0 to 2.8, p<0.001). The educational 19 intervention on its own did not significantly change the length of consultation 20 (0.7 minutes, -0.2 to 1.6 minutes). There was a significant increase in both 21 interventions for number of problems identified. The only change in patient 22 satisfaction was increase in depth of doctor-patient relationship from the 23 agenda form group. 24 25 McCann’s (1996) (McCann-Simon, Weinman John, 1996) randomised 26 controlled trial in the UK was of a brief written intervention leaflet ‘Speak for 27 Yourself’ to increase participation in the consultation read before the 28 consultation compared to those given a control leaflet. The first part asked 29 patients to identify the nature of their problems and to consider their ideas to 30 causes, treatment and effects of the problems. They had space to write down 31 ideas. The second part of the leaflet advises to state their ideas and concerns 32 about the illness to the doctor and ask questions. The intervention group had 33 significantly longer consultations (8.43 minutes, SD=2.97 versus 7.22 Medicines concordance: full guideline DRAFT (July 2008) page 160 of 373 DRAFT FOR CONSULTATION 1 minutes, SD=2.42, 95% CI -0.44 (0.08, 0.81) and they asked more questions 2 than controls. 3 4 McLean (2004) (McLean, Malcolm and Armstrong, David, 2004) conducted 5 an open randomised controlled trial to see whether a prompt to elicit patients’ 6 concerns for minor illness would be beneficial and the costs of doing so. 110 7 patients from four training semi-rural general practices in the South-East of 8 the UK took part in the study. The written prompts were ‘May I ask if you have 9 any concerns about this “….” (illness/pain) you have come about today?’ 10 followed by: ‘Anything in particular about the “…”?’ and, if still unforthcoming: 11 ‘What is it about the “…” that concerns you?’ A consultation satisfaction 12 questionnaire regarding the professional care component was given. The 13 doctor had to record the consultation length (estimated to the nearest minute 14 using a clock to note start and end of consultation) and the diagnosis made. 15 The control group received the consultation as normal. The doctor depending 16 on whether the top sheet of a randomly arranged pile of papers said ‘control’ 17 or ‘intervention’. [However it must be noted that the same doctor was 18 conducting both control and intervention and the control condition may 19 inadvertently receive a more patient-oriented consultation]. Patient satisfaction 20 was 80.9 for controls and 88.2 for intervention patients (SD=11.8), mean 21 difference 7.2 (95% CI 2.0 to 12.6). The consultation length of intervention 22 consultations was on average 1 minute longer for intervention group than 23 controls (11.0 vs. 10.0 minutes), but this was not significant. 24 25 Cohen (2004) (Cohen, D., Longo, M. F., Hood, K. et al , 2004) and Edwards 26 (2004) (Edwards, A., Elwyn, G., Hood, K. et al , 2004) conducted a cluster 27 randomized crossover trial in the UK with the aim to explore the costs of 28 training GPs in developing SDM competences and in the use of risk 29 communication (RC) aids and to evaluate the effects of such training on a 30 range of service resource variables. Edwards (2004) published the main trial 31 results that focused on the doctor patient interaction, patient outcomes and 32 satisfaction with the decision. 33 Within each cluster, patients were also allocated randomly to consult with the 34 doctor at one of three points in the study. The study comprised three phases. Medicines concordance: full guideline DRAFT (July 2008) page 161 of 373 DRAFT FOR CONSULTATION 1 Phase 1 was pre-training. Phase 2 included training for half of the GPs and 2 the other half in RC. In phase 3, each GP received training in the other 3 element making them fully trained in both. The authors argued that in this 4 way, the design offered the greatest potential to gain understanding about the 5 effects of each form of training alone and in combination and if the sequence 6 of skill acquisition was important. A further randomisation allocated patients to 7 attend either in usual surgery time or in a research clinic- audiotaped, 8 including fewer interruptions and more time for each consultation (up to 15 9 minutes each). 10 SDM training involved GPs attending two workshops where standardized and 11 previously piloted skill development process was used. SDM competences 12 were described and demonstrated by means of consultation simulation and 13 pre-prepared scenarios involving the four study conditions. RC also involved 14 attendance at 2 workshops, and the aids consisted of tabulated data and 15 visuals displays of risk estimates for the four study conditions. Patients with 16 one of four conditions (menorrhagia, atrial fibrillation, menopausal symptoms 17 or prostatism) were invited by their GP to attend a “review consultation” to 18 discuss their continuing treatment. Twenty GPs from 20 different practices in 19 South Wales were recruited. Costs of training for both RC and SDM included 20 time of trainers, of those being trained and of the simulated patients used as 21 part of the training exercise. Information on prescriptions, investigations and 22 referrals was obtained from questionnaires completed by each clinician at the 23 review consultation. 24 Edwards (2004) reported significant effects of the research clinic (which 25 provided more time) in confidence in decision (p<0.01) and expectation to 26 adhere to chosen treatments (p<0.05). Anxiety scores approached 27 significance for the RC intervention, as did expectation to adhere to chosen 28 treatment for both interventions. No significant effects of the risk 29 communication or SDM interventions were seen on the whole range of 30 patient-based outcomes. 31 However, Cohen (2004) concluded that due to the explanatory nature of the 32 study, no assessment could be made on how training could affect the length 33 of a consultation. 34 Medicines concordance: full guideline DRAFT (July 2008) page 162 of 373 DRAFT FOR CONSULTATION 1 4.16.2.2 RCTs conducted outside the UK 2 3 Kinnersley (2007) (Kinnersley, P., Edwards, A., Hood, K. et al , 2007) 4 conducted a Cochrane systematic review to find interventions which aimed to 5 increase patient involvement by enabling patients to address their information 6 needs within the consultation. Most of the RCTs were from the USA, 2 from 7 Australia, 5 from the UK and one from the Netherlands. Most of the 8 interventions were written followed by face-to-face coaching and videotape. 9 Written interventions were in booklet or checklist form. The specific 10 behaviours most encouraged were question-asking, raising concerns and 11 requesting clarification or checking understanding. 12 13 Seventeen RCTs in the Kinnersley review (2007) looked at consultation 14 length, 3 studies found a significant increase (Hornberger, 1997; McCann 15 1996 and Middleton 2006). While 14 RCTs found no effect. Bolman (2005) 16 found a decrease in the first consultation and an increase in the last 17 consultation. The meta-analysis showed a small but not significant increase in 18 consultation length (SMD 0.10 95% CI -0.05 to 0.25). 19 Fifteen RCTs reported that the use of written materials during the consultation 20 led to a small and significant increase in consultation length (SMD 0.13, 95% 21 CI 0.05 to 0.21). RCTs with coaching found a non-significant increase (SMD 22 0.07 95% CI -0.07 to 0.20). In studies where there was additional clinician 23 training there was little impact on consultation length for written and coaching 24 materials. RCTs with clinician training SMD 0.17 (95% CI 0.01 to 0.32) 25 compared to studies without clinician training SMD 0.17 (95% CI 0.10 to 0.24). 26 It should be noted that of these seventeen RCTs only eleven of these related 27 to our population of interest, the results for these are detailed below. 28 29 30 4.16.2.3 RCTs included in the Kinnersley (2007) (Kinnersley, P., Edwards, A., Hood, K. et al , 2007) review 31 32 Hornberger (1997) conducted a two-armed, randomised trial of whether a 33 self-administered previsit questionnaire enhanced awareness of patients Medicines concordance: full guideline DRAFT (July 2008) page 163 of 373 DRAFT FOR CONSULTATION 1 concerns in the USA. They completed the Patient Concerns Form while 2 waiting for their visit. This covered 25 items of concerns of five categories: 3 desire for medical information, psychosocial assistance, therapeutic listening, 4 general health advice and biomedical treatment. After the interview the 5 patients completed a postvisit questionnaire which assessed their perceptions 6 of the concerns addressed by the physician. The net effect of the intervention 7 compared to the control group was a difference of 6.8 minutes (95%CI 0.4, 8 13.3) for total time in consultation. With most of the extra time spent 9 discussing diagnoses (3.35 minutes, 95%CI 0.00, 6.72) and in performing the 10 physical examination (2.7minutes, 95% CI 0.5, 4.9). The number of diagnoses 11 increased by 30% in the intervention group compared to the control group 12 (increase of 1.7 diagnoses per visit). Those in the intervention group had 13 marginally higher satisfaction but this was not significant except for overall 14 quality of care (0.35, +/- 0.23, p,0.05). 15 16 Greenfield (1985) conducted a randomised controlled trial of an intervention 17 to increase patient involvement in their care in the US. The intervention group 18 received a treatment algorithm as a guide to help them read their medical 19 record and a behaviour-change strategy. The participants were coached in 20 appropriate question-asking and negotiation of decisions. The intervention 21 occurred in a 20 minute session before their regular consultation with their 22 GP. The control group also saw a clinic assistant just before their regular 23 appointment for a similar amount of time as the intervention group and they 24 received a standard protocol of receiving information and review of ulcer 25 disease and given copies of these materials, they did not get to see their 26 medical records. There was no significant difference between groups in length 27 of consultation after the interventions, both groups averaged 16 minutes per 28 encounter. The time of the encounter before was 16.8, SD 8.2 whereas the 29 control group was 15.1 (7.6), a difference of 1.7 (95% CI -2.92, 6.32). The 30 time of the encounter after was 15.7 (6.7) for the intervention and 16.3 (9.7) 31 for the control, -0.6 (95% CI -5.49, 4.29). However, they differed in how they 32 spent their time with the intervention patients spent more time involved in the 33 interaction than controls. 34 Medicines concordance: full guideline DRAFT (July 2008) page 164 of 373 DRAFT FOR CONSULTATION 1 Greenfield (1988) conducted a randomised controlled trial in a diabetic clinic 2 in the USA. This intervention was the same as in Greenfield (1985) but 3 delivered twice, before the initial and follow up consultations. There was no 4 change in question asking, patient satisfaction, knowledge or consultation 5 length (30.30 SD=13.80 intervention group versus 32.50 SD 13.90 for the 6 control group). Participation and the preference for active involvement 7 increased. 8 9 Maly (1999) conducted a randomised controlled trial in a family medicine 10 clinic in the US. where patients received copies of their medical record 11 progress notes and produced two main questions to ask their physician. 12 Control group received health education sheets and made suggestion lists for 13 their clinic care. The consultation length did not differ between groups. 14 15 Roter’s (1977) randomised control trial in a family medicine clinic in the US 16 involved 10 minutes with a health educator to identify questions from a 17 question asking protocol. The participants were encouraged to ask questions 18 and took a list in to the consultation. Question asking and patient satisfaction 19 increased and there was no difference between consultation length (29.90, 20 SD=12.70) vs 40.50 (92.70). 21 22 Thompson (1990a) conducted a randomised controlled trial in an obstetric 23 and gynaecologist clinic in the USA. Participants received a question prompt 24 sheet with instructions to write at least 3 questions to take to the consultation. 25 Question asking increased and there was no change in patient satisfaction 26 and consultation length 7.70 (SD=2.90) vs 8.70 (SD=4.70), 95% CI -0.26 (- 27 0.80, 0.29). 28 29 Martinali’s (2001) randomised controlled trial in the Netherlands used a 30 checklist to prepare coronary patients for visiting their cardiologist. The short 31 checklist which was to be completed at home was aimed towards structuring 32 the exchange of information in the consultation and to concentrate on those 33 issues that caused most concern to the patient. A brochure was also 34 developed with instructions for the checklist. A brochure was also given from Medicines concordance: full guideline DRAFT (July 2008) page 165 of 373 DRAFT FOR CONSULTATION 1 the Dutch Heart Foundation, which both groups received. The consultation 2 length was 12 minutes (SD 4.2) in the experimental group and 10.3 (SD 3.8) 3 in the control group, f=1.82, p=0.18. 4 5 Bolman’s (2005) randomised controlled trial in cardiology clinics in the 6 Netherlands involved a checklist of 49 questions on 10 issues (as Martinali 7 2001). This was mailed to the patient a week before each of three linked 8 consultations. There was no change in patient satisfaction. Consultation 9 length was reduced at first visit but increased at third visit (13.73, SD=3.73 vs 10 16.22, SD=5.84, 95% CI -0.49, -0.88, -0.10) 11 12 4.16.2.4 Other systematic reviews of RCTs 13 14 Harrington’s (2004) (Harrington-Jane, Noble Lorraine, 2004) systematic 15 review, which investigated how to improve communication in a consultation 16 showed that studies overall found that by involving patients there was not a 17 resultant increase in consultation length. Five out of seven studies that 18 included consultation length (and were our population of interest) found there 19 was not a significant increase in the length of consultation except for 20 Hornberger (1997) and McCann (1996). All of these studies are in the 21 Kinnersley (2007) review except for McGee (1998), (a study conducted in the 22 USA) which did not find any difference in consultation length. 23 24 The Wetzels (2007) (Wetzels, R., Harmsen, M., van, Weel C. et al , 2007) 25 systematic review, which looked at interventions to improve older patients’ 26 involvement, reported findings related to consultation length. Only one of the 27 three studies meeting the inclusion criteria of the review included consultation 28 length (Cegala 2001). In Cegala (2001) the trained patients asked more 29 medically-related questions, gained more information and provided more 30 information than control patients. They did not verify information more than 31 control patients and appointment length was not longer overall. (18.81 vs 32 22.59, p=0.46) and time engaged in talk 16.25 vs 14.41, p=0.68). This study 33 was conducted in the USA. Medicines concordance: full guideline DRAFT (July 2008) page 166 of 373 DRAFT FOR CONSULTATION 1 4.16.2.5 RCTs (not included in any systematic reviews) 2 Loh (2007) (Loh, Andreas, Simon, Daniela, Wills, Celia E. et al , 2007) 3 investigated the effects of a shared decision-making intervention in primary 4 care of depression compared to usual care on adherence, satisfaction and 5 clinical outcomes. The study was conducted in Sudbagen, Germany with 6 primary care physicians as the unit of randomisation. The sampling frame 7 (n=148) were sent a letter, 30 accepted the invitation to take part, 20 were 8 randomly assigned to the intervention group and 10 to the control group, after 9 drop out 15 and 8 were left respectively. The physicians had to recruit newly 10 diagnosed depressive patients. The intervention physicians enrolled 263 11 patients and the control group 142. After their diagnosis the patients 12 completed a questionnaire measuring patient involvement, depression 13 severity and socio-demographic characteristics. After 6-8 weeks the patients 14 completed a second questionnaire measuring adherence and treatment 15 outcome. At the same time, the physicians rated their assessment of the 16 patients’ adherence. The shared decision-making intervention was then 17 implemented with the intervention group. The intervention was a multi-faceted 18 program including physician training, a decision board for use during the 19 consultation given to the patients after the consultation, printed patient 20 information with specific encouragement to be active in the decision-making 21 process. The physicians in the intervention group also completed modules on 22 guideline-concordant depression care which included content on enhancing 23 skills for improving the shared decision-making process. The outcomes 24 measures were patient participation, treatment adherence, patient 25 satisfaction, consultation time and clinical outcomes. There was no difference 26 for the control group in patient participation before and after, whereas the 27 intervention group had significantly higher patient participation from pre to 28 post intervention for the doctor facilitation scale (p=0.001) and there was an 29 increase in the patient participation scale (p=0.010). There were no significant 30 differences in treatment adherence. Patient satisfaction was significantly 31 higher in the intervention 29.8 (2.7) than the control group 27.0 (3.6), p=0.14. 32 There were no values taken for satisfaction before the intervention. There was 33 no difference between groups for length of consultation 29.2 (10.7) versus Medicines concordance: full guideline DRAFT (July 2008) page 167 of 373 DRAFT FOR CONSULTATION 1 26.7 (12.5). Neither group had a significant reduction in depression severity 2 from baseline to post-intervention. 3 4 Hamann (2006) (Hamann, J., Langer, B., Winkler, V. et al , 2006) conducted 5 a randomised controlled trial which aimed to assess an intervention for shared 6 decision-making in patients with acute schizophrenia. 107 patients from 12 7 acute psychiatric wards of two hospitals in Germany were included in the 8 study. Before the consultation participants were given a talk on their treatment 9 options and to prepare them for their GP consultation. A 16 page booklet 10 decision aid covering the pros and cons of oral vs depot formulation, first vs 11 second generation antipsychotics, psycho-education, and type of socio- 12 therapeutic intervention. Trained nurses assisted the patients to work through 13 the booklet and gave answers to any information requests. They had to write 14 down their experiences with antipsychotic medication and to indicate their 15 preferences. They met with physicians within 24 hours of working through the 16 decision aid. The control group received routine care. There was no difference 17 reported in the time spent in individual consultations as reported by the 18 psychiatrists - mean 64 min/weeks for the intervention group compared to 60 19 min/weeks for the control group, p>0.05. 20 21 22 23 Medicines concordance: full guideline DRAFT (July 2008) page 168 of 373 DRAFT FOR CONSULTATION 1 5 Patients’ experience of medicine-taking 2 5.1 Recommendations *listed at the start of the guideline according to chapter 3 4 5.2 Introduction 5 If health care professionals are to facilitate patient involvement in decisions 6 about medicines it is helpful and necessary to understand how patients 7 approach the taking of medicines, in particular the ongoing appraisal of 8 medicines that continues after a consultation. Investigation into why patients 9 do not take medicines as prescribed indicates that the decision to take 10 medicines and the continuing taking of medicines should be considered as a 11 complex behaviour (Horne, R, Weinman, J., Barber, N. et al , 2005). 12 13 5.3 Key Clinical Question: what are the barriers and facilitators for individuals in medicine-taking 14 Related Evidence statements Evidence into recommendations references (summary of evidence) Pound 1. Patients wish to minimise The GDG discussed the appropriate (2005)(Pound, P., medication intake where research methodology to provide Britten, N., possible. They may wish to do evidence of patients actual medicine Morgan, M. et al , this to decrease adverse taking behaviours. The GDG 2005); Munro effects and potential for considered it important to provide (2007) (Munro, addiction, to make the regimen health care professionals with Salla A., Lewin, more acceptable or for financial evidence of how patients actually use Simon A., Smith, reasons. Patients may decide medicines to sensitize professionals to Helen J. et al , to use prescribed medicine to issues that may be relevant for 2007); Mills (2006) alleviate symptoms or discussion with individual patients. (Mills, Edward J., strategically, to replace or The GDG accepted the use of Nachega, Jean B., supplement medicines qualitative evidence for this and Bangsberg, David sometimes or all the time with considered the Pound synthesis Medicines concordance: full guideline DRAFT (July 2008) page 169 of 373 DRAFT FOR CONSULTATION R. et al , 2006); non-pharmacological provided the type of evidence they Carrick (2004) treatments were looking for. To augment the (Carrick, Rachael, Mitchell, Annie, Powell, Richard A. et al , 2004); Deegan (2005) (Deegan, P. E., 2005); Lewis (2006) (Lewis, M. P., Colbert, A., Erlen, J. et al , 2006); Cooper (2002) (Cooper, V., Buick, D., Pound synthesis searches were made Patients will commonly evaluate prescribed medicines by trying out the medicines and weighing up the costs and benefits. They will consider adverse effects and acceptability of regimen. They may stop the medicine and see what happens and obtain information from non-medical for evidence not included in the Pound review and for any other systematic reviews. The description of patient behaviours and factors influencing patients medicine taking behaviours were used to inform the recommendations. The GDG were given access to research in progress supporting the influence of patient beliefs on actual adherence. sources and observe the effect of medicines on others. Horne, R. et al , Both subjective and objective 2002); Ogedegbe indicators may be used to (2004) (Ogedegbe, evaluate medicines. Gbenga, Harrison, Melanie, Robbins, Patients do not generally Laura et al , 2004); disclose their beliefs and Lukoscheck (2003) change of regimen to HCPs. (Lukoschek, Petra., 2003); Dowell (2002) (Dowell, Jon, Jones, Anni, and Snadden, David, 2002); Wilson (2002) (Wilson, Patients may not be able to recognize the difference between effects of medication and effects of disease and have difficulty in evaluating long term preventative medication where there are no symptoms. Holly Skodol, Hutchinson, Sally Patient on multiple medications A., and Holzemer, may make choices between Medicines concordance: full guideline DRAFT (July 2008) page 170 of 373 DRAFT FOR CONSULTATION William L., 2002); medicines. Bollini (2002) 7589}; Adam (2002) (Adam, B. D., Maticka, Tyndale E., and Cohen, J. J., 2003); Scotto (2005) (Scotto, Carrie J., 2005); Alfonso (2006) (Alfonso, V., Bermbach, N., Geller, J. et al , 2006); Sidat (2007) (Sidat, Mohsin, Fairley, Christopher, and Grierson, Jeffrey, 2007); Lewis (2006) (Lewis, M. P., Colbert, A., Erlen, J. et al , 2006); Pyne 92007) (Pyne, J. M., McSweeney, J., Kane, H. S. et al , 2006); Kikkert (2006) (Kikkert, Martijn J., Schene, Aart H., KoeterMaarten, W. J. et al , 2006); Vinter- Medicines concordance: full guideline DRAFT (July 2008) page 171 of 373 DRAFT FOR CONSULTATION Repalust (2004) (Vinter, Repalust N., Petricek, G., and Katic, M., 2004); Campero (2007) (Campero, Lourdes, Herrera, Cristina, Kendall, Tamil et al , 2007); Reid (2006) (Reid, M., Clark, A., Murdoch, D. L. et al , 2006); Elliott (2007) (Elliott, Rachel A., Ross, Degnan Dennis, Adams, Alyce S. et al , 2007); Aronson (2005) (Aronson, B., 2005); PopaLisseanu (2004) (Garcia-PopaLisseanu, M. G., Greisinger, A., Richardson, M. et al , 2005); Erwin (1999) (Erwin, J. and Peters, B., 1999); Mutchler (2007) (Mutchler, Jan E., Bacigalupe, Gonzalo, Coppin, Medicines concordance: full guideline DRAFT (July 2008) page 172 of 373 DRAFT FOR CONSULTATION Antonia et al , 2007); Lawton (2005) (Lawton, J., Ahmad, N., Hallowell, N. et al , 2005); Morgan (2005) (MorganMyfanwy, Figueroa-MuñozJose, 2005) 1 Methods of the evidence review 2 5.3.1 3 Searches of the literature revealed a large number of studies that set out to 4 explore patients’ experience of medicine taking. The majority of these studies 5 are qualitative studies. One of the current challenges in qualitative research is 6 how to bring together the findings from individual qualitative studies. One 7 approach is to present a narrative of these studies describing their findings 8 individually. More recently the field of qualitative synthesis has attempted to 9 synthesise the findings from different studies into a common set of findings 10 that includes the findings from individual studies but that may also provide 11 additional levels of understanding that may not be apparent when each study 12 is looked at individually. A synthesis of medicine-taking has already been 13 conducted using a meta-ethnography (Pound, P., Britten, N., Morgan, M. et al, 14 2005). The Pound (2005) (Pound, P., Britten, N., Morgan, M. et al , 2005) 15 review is methodologically sound and systematic synthesis developed by a 16 panel of experts within the field of medicines concordance. Following 17 discussion with the GDG it was agree that rather than conduct an alternative 18 summary or synthesis of the qualitative studies of patients experience of 19 medicine-taking we used the synthesis as the basis for our evidence review. 20 The findings of the synthesis were updated by searching for evidence 21 published since the review was conducted and a narrative summary of these 22 presented. The GDG requested additional search for systematic reviews. Medicines concordance: full guideline DRAFT (July 2008) page 173 of 373 DRAFT FOR CONSULTATION Evidence review 1 5.3.2 2 Pound (2005) (Pound, P., Britten, N., Morgan, M. et al , 2005) aimed to 3 conduct a synthesis of qualitative studies of lay experience of medicine taking. 4 The study was part of a Health Technology Assessment project to evaluate 5 meta-ethnography as a method of synthesizing qualitative research studies of 6 health and health care. This narrative has used the paper published in Social 7 Science and Medicine (Pound, P., Britten, N., Morgan, M. et al , 2005). The 8 authors of the HTA report kindly allowed us to read a draft copy of their report 9 for the HTA which primarily concerned the detail of their methodology. Studies 10 using both qualitative methods of data collection and qualitative methods of 11 data analysis and published in English were included. Studies published 12 between January 1992 and December 2002 were eligible for inclusion. 13 Medline, Embase, Cinahl, Web of Science, Psychinfo and Zetoc were 14 searched. The electronic search was supplemented by extensive 15 handsearching. Papers were appraised for quality using a version of CASP 16 1988 criteria. Thirty-eight papers were included in the synthesis. Papers were 17 organised into medicine groups. The initial synthesis brought together papers 18 looking at similar medicine/disease groups and these findings were then 19 synthesised. The medicine groups were antiretroviral therapy, 20 antihypertensives, psychotropic medicine, proton-pump inhibitors, asthma, 21 miscellaneous medicines and medicines in general. 22 The paper reports a summary of the findings of the individual studies and the 23 results of the synthesis of these studies. The authors developed the concept 24 of ‘resistance’ to medicines to describe lay peoples response to medicines. 25 One of the main conclusions of the synthesis is that people do not take 26 medicines as prescribed because of patient concerns about medicines 27 themselves. Their interpretation is that non-adherence is not necessarily a 28 result of failures from the doctors, patients or systems, but because of 29 concerns about the medicines themselves. Drawing on issues such as patient 30 reports of testing for adverse effects; worries about dependence; the potential 31 harm from taking medicines on a long term basis; and issues with disclosure 32 and stigma, the authors’ conclude that many patients ‘resist’ the taking of 33 medicines. Patients can be described as ‘acceptors‘ of medicines, some Medicines concordance: full guideline DRAFT (July 2008) page 174 of 373 DRAFT FOR CONSULTATION 1 uncritically, others following their own enquires and experimentation. People’s 2 medicine taking may change with different medicines and the illness in 3 question, which illustrates the relative nature of the concept of “resistance”. 4 Those that may “resist” a certain type of medicines may “accept” a different 5 medicine. 6 The findings of the synthesis fell into three parts (1) the way people evaluate 7 their medicines and the difficulties they encounter in doing this (2) the 8 interaction between medicines and patient identity (3) the ways people take 9 their medicines. The themes that make up each part are listed under each 10 heading. 11 5.3.3 12 encountered by people in evaluating medicines The evaluation of medicines and the difficulties 13 14 Trying out the medicines and weighing up the costs and benefits 15 The most common way of evaluating medicines was to try it out and weigh up 16 the benefits of taking it against the cost of doing so. The majority of the 17 studies focused more on the problems associated with medication and less on 18 the benefits of medication, but it was clear that people had hope that their 19 medicines would help with the symptoms; avoid relapse and hospitalization; 20 for minimization of disease progression or for prevention of future illness. 21 Adverse effects 22 Adverse effects were a key criterion in the evaluation of medicines. This was 23 found in several studies including those about treatments for cancer, 24 rheumatoid arthritis, asthma, diabetes, schizophrenia and digestive disorders. 25 Adverse effects were very prominent in studies of patients taking anti-HIV 26 medication. The frequency and nature of adverse effects experienced by 27 these patients, particularly as adverse effects were severe and unpredictable, 28 resulted in distrust and fear of the medicines. These adverse effects affected 29 social activities, relationships and work. Medicines concordance: full guideline DRAFT (July 2008) page 175 of 373 DRAFT FOR CONSULTATION 1 Acceptability of regimen 2 People reported the evaluation of the suggested regime in terms of how it 3 fitted in with daily schedules and life in general. The frequency of doses and 4 number of pills was also a challenge, as also taste, smell, size and shape of 5 the pills. Regimes that required disruption of schedules resulted in patients 6 reporting that they were no longer in control of their lives and people varied as 7 to whether they fitted the world around their regime or resisted the demands 8 of the regime and missed doses. 9 Weighing and balancing 10 A process of weighing and balancing was carried out by people where the 11 advantages of treatment in terms of symptoms or effect on disease was 12 balanced by side effects and disruption. Adverse effects and disruption would 13 lead people to question if it was worth taking medication or not.. 14 Stopping the medicine and seeing what happens 15 In some cases, patients test the medicines by either changing the doses or 16 stopping the medicine as to observe what happens. Some authors suggest 17 that this process can either be explicit or subconscious and tends to occur 18 more frequently in long term conditions. 19 Observing others, obtaining information 20 Patients used a variety of sources of information with some patients relying 21 more on their observations of how other people dealt with medicines such as 22 HIV medicines rather than the advice/information given by the doctor. People 23 use a variety of sources of information (e.g. internet, books, support groups) 24 as well as that provided from their GPs 25 Objective and subjective indicators 26 People used both objective and subjective indicators to evaluate efficacy of 27 medication. In the studies in the synthesis blood pressure monitoring 28 appeared to be used widely as a means of evaluating efficacy of 29 antihypertensive medication. The perception of symptom minimization could 30 be as important as objective indicators such as an increase in T-cell count in 31 the case of HIV patients. Medicines concordance: full guideline DRAFT (July 2008) page 176 of 373 DRAFT FOR CONSULTATION 1 Gender differences in evaluating medicines 2 Some studies suggested that women with HIV do not belong to certain social 3 networks like gay men or injecting drug users and thus could be less informed 4 about the medicines. Also, some women were reported to show scepticism 5 towards the drugs used for HIV with arguments that the clinical trials had not 6 been conducted with women. 7 Difficulties with evaluating medicines 8 In some studies it was noted that patients could not distinguish between the 9 effects of their illness from the effects of the medicine. This could even lead to 10 patients rejecting treatments mistakenly. Other authors pointed out that 11 evaluation can be dependent on the individuals understanding of how a 12 medicines works and its function and this can be difficult for certain people 13 due to lack of information and understanding. This is particularly relevant in 14 the case of preventative medicine, as there are no immediate symptoms that 15 can used as indicators of efficacy. 16 Worries about medicines that lay testing and evaluation cannot resolve 17 Fear of dependency and tolerance were pointed out as issues for patients. 18 Fear of addiction was reported when taking psychotropic medicines and 19 general concern over taking medicines on a long term basis was present with 20 hypertensives. Some authors noted that those who worried over the long term 21 effects of medicines were those most likely to change their regimen to the 22 lowest possible dose. 23 5.3.4 24 Non-acceptance 25 Since taking medicine is equated with having an illness people may not take 26 medicine if they do not accept their illness. This was particularly strong in the 27 asthma studies in the synthesis. The relationship with acceptance of disease 28 was however complex with some patients using medicines to keep their 29 problem under control and downplay its significance. In the case of 30 neuroleptic medication and HIV medication an acceptance of the diagnosis 31 was crucial in determining whether the patient would take their medicine as 32 prescribed. Medication was seen as a reminder of illness Medicines and identity Medicines concordance: full guideline DRAFT (July 2008) page 177 of 373 DRAFT FOR CONSULTATION 1 Disclosure and stigma 2 People with potentially stigmatizing illnesses such as HIV and mental health 3 problems were particularly concerned that their medications marked them out 4 as individuals with those particular health problems. This could lead to 5 avoiding taking medicine in public and either postponing or foregoing their 6 medicine intake. Some people were reported to not initiate treatment which 7 would reveal them as having HIV. Patients with mental health problems also 8 reported feeling stigmatized by their intake of medicines, and some felt 9 ashamed. Ways people take their medication 10 5.3.5 11 Motivation to minimize intake 12 The majority of studies illustrate how people wish to minimize their intake of 13 medicines with patients spontaneously reporting dislike of drugs to 14 researchers. This was also true in the case of medicines commonly reported 15 as being overused by patients, such as benzodiazepines. 16 To decrease adverse effects and addiction 17 People often reduce or skip doses, or take tablets separately rather that all at 18 once. Others may pause their medicine intake as a way of cleansing their 19 body and minimize toxicity. 20 To make the regimen more acceptable 21 This section related to how people modified their regimen in order to fit with 22 their daily schedule, alleging that to have some clinical gain, complete 23 adherence was not required. Other argued that the optimum regimen was not 24 known anyway, and that strict adherence was not possible. 25 For financial reasons 26 Some people were reported to have decreased their doses as they could not 27 afford the prescribed amounts. 28 Using medicine symptomatically 29 Some patients were reported to use medicines accordingly to symptoms 30 displayed. Patients associated symptoms with their medical problem and 31 when these symptoms indicated that the problem was not controlled they Medicines concordance: full guideline DRAFT (July 2008) page 178 of 373 DRAFT FOR CONSULTATION 1 would take their medicines. An example was symptoms of tiredness or 2 weakness in treatment for high blood pressure. People with rheumatoid 3 arthritis modified their doses according to their symptoms as did patients on 4 neuroleptics. 5 Using medicine strategically 6 People adjusted their dose or did not take their medication when planning to 7 drink alcohol as they feared possible interactions. This was also reported for 8 those on neuroleptic medicines and antihypertensive medication. PPIs were 9 also altered according to what people planned to eat. 10 Replacing or supplementing medicines with non-pharmacological treatments 11 In varying ways, people often complemented their treatment with home 12 remedies. On a more specific level, some patients who worried about the 13 harmful effects of medicines would sometime take a break and use natural 14 remedies for a certain period. 15 Doctor-patient communication about regimen modifications 16 Some patients reported being scolded by their doctors if making their own 17 decisions about their care. Rather than confronting them, they would then 18 switch doctors. Also, many patients do not reveal their beliefs to the doctors, 19 but once outside the surgery and in control, people would then change or 20 refute their regimen. Some authors noted that patients who had changed their 21 regimen would not disclose this to the HCPs due to previous experience of 22 coercion or recognition of their powerless position. Authors argue that as 23 patients self-regulate anyway, doctors need to recognize this and even 24 support them in the process, as well as helping people feel in control. 25 Imposed compliance 26 One issue that was found exclusively in studies relating to mental health was 27 that of ‘imposed compliance’. Patients with mental health problems felt 28 surveyed for signs of ill health and under pressure or even coercion from 29 friends, relatives and health care professionals to take medicines. Patients felt 30 that medicine was used to make them acceptable to society and was part of Medicines concordance: full guideline DRAFT (July 2008) page 179 of 373 DRAFT FOR CONSULTATION 1 an unwritten social contract that required the taking of medicines to allow 2 patients to be acceptable to the community. 3 5.4 Update of qualitative evidence synthesis - Pound 2005 synthesis 4 Methodology of update 5 5.4.1 6 The aim is to update the synthesis of qualitative evidence of medicine taking 7 in a similar methodological approach as when updating a Cochrane Review. 8 The titles and abstracts of records retrieved by the searches were scanned 9 and any potentially relevant publications obtained in full text. The studies were 10 reviewed to identify the most appropriate evidence to help answer the key 11 questions and to ensure that the recommendations are based on the best 12 available evidence. Qualitative synthesis is considered a process where the 13 analysis of a number of qualitative studies may result in new findings not 14 contained in the individual studies. It was felt therefore inappropriate to do 15 additional synthesis. This update is a narrative review which discusses the 16 studies found in the update particularly where the findings add to the existing 17 synthesis. 18 Types of studies - studies with both qualitative methods of data collection 19 and analysis published in English. 20 Types of participants - people aged above 16 years prescribed medication 21 for a medical condition. (3 studies included in the synthesis relate to children 22 and/or adolescents). 23 Duration of studies - no limit on duration of studies. 24 Possible challenges - One of the first challenges of the process of 25 developing a systematic review on qualitative evidence is how to find the 26 evidence. Qualitative evidence is catalogued on a wide range of databases or 27 sometimes not at all and indexes and search filters require substantial 28 improvement if they are to be rigorous and systematic. Secondly, is the lack of 29 agreement of appropriate methods for appraising the quality of qualitative 30 evidence. We will use the same criteria as those used in the original Medicines concordance: full guideline DRAFT (July 2008) page 180 of 373 DRAFT FOR CONSULTATION 1 synthesis. The third challenge relates to the sheer complexity of synthesizing 2 evidence that can have varying approaches to data collection and analysis 3 and issues surrounding generalisability of the findings which is of utmost 4 importance to the evidence based medicine (EBM) discipline. Generalisability 5 may conflict with the underpinning philosophical framework of some 6 qualitative research. 7 5.4.2 8 Update searches 9 Our update searches produced 381 references. Based on abstract Evidence review 10 information, eighty five studies were ordered. Further exclusions were made if 11 the study turned out not to be relevant for the topic, or did not have qualitative 12 data collection and analysis. CASP criteria were used to assess the quality of 13 the studies. The details of all the studies are in Appendix C. Forty-five studies 14 were included in the update. The studies covered a wide range of medical 15 conditions and patient groups but did cluster around long term conditions. 16 Twelve were concerned with HIV medication, 8 with medication for psychiatric 17 conditions and 4 for patients with diabetes. Five studies focused on patients 18 from low income and/or ethnic groups. A shift could be seen in that most 19 recent studies discussed the issue of patient-health care professional 20 communication more explicitly than those included in the Pound synthesis. 21 Many papers, as in the Pound synthesis, accepted a medical paradigm that 22 medicine-taking was a good thing and some sought to understand patients’ 23 beliefs and experience with a view to improving adherence. Patients’ 24 readiness to negotiate with health care professionals and to disclose their 25 medicine taking behaviour to health care professionals was discussed (Bane, 26 Catherine, Hughes, Carmel M., Cupples, Margaret E. et al , 2007) as was the 27 challenges facing health care professionals in supporting the integration of 28 patients needs and preferences (Hayes, Risa P., Bowman, Lee, Monahan, 29 Patrick O. et al , 2006). In general the findings fitted well into the categories 30 elaborated in the Pound synthesis although studies often developed their own 31 terminology and categories. As described in the Pound synthesis many 32 studies did not reference each other. The findings in this update are described 33 under the themes as described in the Pound synthesis i.e. (1) the evaluation Medicines concordance: full guideline DRAFT (July 2008) page 181 of 373 DRAFT FOR CONSULTATION 1 of medicines and the difficulties people encountered in this, (2) medicines and 2 identity and (3) the ways people take their medication. Medicines concordance: full guideline DRAFT (July 2008) page 182 of 373 DRAFT FOR CONSULTATION 1 2 5.4.2.1 The evaluation of medicines and the difficulties encountered by people in evaluating medicines 3 The findings of the update searches were similar to the findings of the Pound 4 synthesis in how patients evaluated medicines. The themes of trying out the 5 medicines and weighing up the costs and benefits were present as well as the 6 importance of adverse effects and the acceptability of the regimen. The 7 studies elaborated different terminology. Carrick (2004) (Carrick, Rachael, 8 Mitchell, Annie, Powell, Richard A. et al , 2004) developed a core concept of 9 ‘well being’. The study was an interview study of 25 adults taking antipsychotic 10 medication. The findings were that patients sought to maximize ‘well-being’ 11 which was normality of function, feeling and appearance. ‘Well being’ was 12 defined personally by patients and was their goal in taking treatment. The 13 achievement of well-being was a net effect of symptoms and side effects. 14 Some patients preferred the effects of their disease to the side effects of 15 treatment. This was achieved by an interplay of evaluating treatment, 16 managing treatment and patients’ understanding of the situation. Patients 17 considered medication in the context of their beliefs about their illness and its 18 causes. While the maximizing of ‘well-being’ was relevant for all patients 19 interviewed there was a spectrum of behaviour in relation to how active the 20 patient was in engaging with their doctors and talking through their views 21 about medication. Deegan (2005) (Deegan, P. E., 2005) again in the field of 22 psychiatric problems developed a concept of ‘personal medicine’. She 23 interviewed 29 patients with psychiatric problems and considered that 24 psychiatric medications are considered in relation to ‘personal medicine’ i.e. 25 non-pharmaceutical activities that gave meaning and purpose to life and that 26 serve to raise self-esteem, decrease symptoms and avoid unwanted 27 outcomes. Examples of personal medicine were the ability to work, and to 28 parent appropriately and to engage in social activities. Her analysis was that 29 medicines that conflict with patients’ ‘personal medicine’ are unlikely to be 30 used by patients. 31 Balancing the benefits of medication against the side effects of treatment was 32 a theme also in studies of patients on medication for HIV. Lewis (2006) 33 (Lewis, M. P., Colbert, A., Erlen, J. et al , 2006) interviewed patients who were Medicines concordance: full guideline DRAFT (July 2008) page 183 of 373 DRAFT FOR CONSULTATION 1 100% adherent with HAART treatment and found that they performed a trade 2 off between the benefits and side effects of medication. In this sample of 3 adherent patients the interviewees reported that they did not have many other 4 options for treatment and HAART was important in keeping them well. 5 Cooper (2002) (Cooper, V., Buick, D., Horne, R. et al , 2002) interviewed 26 6 patients who had declined HAART treatment and found that they used their 7 own interpretations (which often differed from professional interpretations) of 8 indicators such as CD4 counts to inform their decision, preferred non- 9 pharmacological treatments and also found the lack of symptoms an issue in 10 considering treatment. The paper uses the concept of patients perceived 11 ‘personal necessity’ of treatment as a factor in their decision. Patients also 12 had concerns about medicines from previous personal experiences or from 13 seeing and talking to others taking HIV medicines. 14 The methods by which people evaluate medicines and the difficulties 15 experienced by patients in doing so was present in the studies. Ogedegbe 16 (2004) (Ogedegbe, Gbenga, Harrison, Melanie, Robbins, Laura et al , 2004) 17 interviewed 106 African –American patients in urban primary care clinics. 18 Patients had difficulties in evaluating treatment due to the lack of symptoms of 19 raised blood pressure. Many used their own indicators to consider if and when 20 they should take any treatment. . Lukoschek (2003) (Lukoschek, Petra., 21 2003) interviewed 92 African-American patients about their beliefs and 22 attitudes to hypertension and anti-hypertensive medication. Patients held 23 differing understandings of high blood pressure and hypertension. Patients’ 24 beliefs about problem influenced their approach to treatments including diet, 25 exercise and medication. Patients’ weighed beliefs about advantages of 26 medicines against the side effects and many patients preferred herbal and 27 alternative remedies. Dowell (2002) (Dowell, Jon, Jones, Anni, and Snadden, 28 David, 2002) interviewed patients from 3 GP surgeries in Scotland about 29 experiences of medicines. The patients were selected as a group of non- 30 adherent patients. For this group particular problems were again found for 31 non-symptomatic conditions and patients evaluated illness and medicines 32 using information from their own personal experience and that from their own 33 social circle. Patients with HIV interviewed for the study by Wilson (2002) Medicines concordance: full guideline DRAFT (July 2008) page 184 of 373 DRAFT FOR CONSULTATION 1 (Wilson, Holly Skodol, Hutchinson, Sally A., and Holzemer, William L., 2002) 2 found it difficult to understand and assess their medication as they did not 3 know whether any symptoms were related to their disease or the medication. 4 Bollini (2004) (Bollini, P., Tibaldi, G., Testa, C. et al , 2004) in a study of 5 patients taking anti-depressant medication indicated that patients would test 6 treatment by stopping once they felt better to see what would happen and if 7 they really needed the medication. 8 9 For HIV patients to derive benefit from HAART medication, high adherence to 10 the prescribed regime is required. The acceptability of the regimen and fitting 11 it into schedules was a significant issues in all studies which examined 12 patients experience of taking HAART medication. Adam (2002) (Adam, B. D., 13 Maticka, Tyndale E., and Cohen, J. J., 2003) interviewed patients who were 14 taking HAART but found the required schedule difficult and altered the dosing 15 regime and associated eating rules to fit the regime into their schedules. This 16 paper concludes that the nature of HAART medication and its regime should 17 be seen as the problem with this medication and non-adherence not seen as 18 a patient problem. 19 20 5.4.2.2 Medicines and identity 21 The meta-ethnography synthesis indicated that for many patients the taking 22 of medicines interacted with issues of identity. Non –acceptance of diagnosis 23 and issues around disclosure and stigma were significant issues in studies 24 found in Pound synthesis. Medicines challenged patients to consider 25 themselves as someone with a disease or could provide external evidence of 26 a stigmatizing disease. These issues recurred in the studies included in the 27 update. Scotto (2005) (Scotto, Carrie J., 2005) interviewed 14 patients who 28 had required hospitalization for a relapse of heart failure symptoms. In this 29 sample the acceptance of the diagnosis of heart failure resulted in an altered 30 self-image for patients but this acceptance and its integration into patients’ 31 lives was an important part of managing medication. Behaviours to support 32 adherence worked when the illness and its management could be integrated 33 into ordinary life. Alfonso (2006) (Alfonso, V., Bermbach, N., Geller, J. et al , Medicines concordance: full guideline DRAFT (July 2008) page 185 of 373 DRAFT FOR CONSULTATION 1 2006) interviewed 15 people who were HIV positive who were not taking 2 medication and explored their reasons not to take medicine. Many had prior 3 experience of taking HAART. Not taking medication allowed some to deny 4 that they were HIV positive. For some either taking HIV medication per se or 5 the occurrence of side effects of this treatment risked exposing their HIV 6 status. Many already felt isolated and separate and did not want to exacerbate 7 this. Similar issues of taking HAART were reported in an interview study by 8 Sidat (2007) (Sidat, Mohsin, Fairley, Christopher, and Grierson, Jeffrey, 2007) 9 where patients delayed in starting treatment while dealing with issues of 10 identity and denial. Lewis (2006) (Lewis, M. P., Colbert, A., Erlen, J. et al , 11 2006) interviewed 13 patients who were known to be 100% adherent to 12 HAART treatment and a prominent theme in this sample was transcending 13 their identity as someone with HIV which for the patients was associated with 14 feelings of self-blame and moving on from that to take control of their health 15 and its’ maintenance. In an interview study by Wilson and colleagues (Wilson, 16 Holly Skodol, Hutchinson, Sally A., and Holzemer, William L., 2002) issues of 17 identity are more dynamic and are part of an ongoing appraisal of medicines 18 and medicine taking. Pyne 2007 (Pyne, J. M., McSweeney, J., Kane, H. S. et 19 al , 2006) explored explanatory models of schizophrenia and treatments for 20 schizophrenia held by professionals and patients to provide both provider and 21 patient perspectives. There were significant differences between providers 22 and patients in models of disease, its causes and required treatments. Stigma 23 was identified as a problem by patients but not providers. Kikkert (2006) 24 (Kikkert, Martijn J., Schene, Aart H., Koeter-Maarten, W. J. et al , 2006) 25 explored views of patients with schizophrenia, their carers and professionals 26 on medication for schizophrenia. A significant barrier for this patient group 27 was the social consequences of extra-pyramidal side effects of medication. 28 Patients taking antidepressant medication also reported difficulty in accepting 29 the diagnosis and therefore the need for treatment Bollini (2004) (Bollini, P., 30 Tibaldi, G., Testa, C. et al , 2004)) Patients with diabetes complained of a 31 stigma of been seen with and using treatment for diabetes in a study by 32 Vinter-Repalust (2004) (Vinter, Repalust N., Petricek, G., and Katic, M., 33 2004). For these patients coming to terms with the identify of being a patient 34 with diabetes was a significant challenge. Ogedegbe (2004) (Ogedegbe, Medicines concordance: full guideline DRAFT (July 2008) page 186 of 373 DRAFT FOR CONSULTATION 1 Gbenga, Harrison, Melanie, Robbins, Laura et al , 2004) found that African – 2 American patients reported difficulty coming to terms with a diagnosis of 3 hypertension as a significant problem in taking anti-hypertensive medication. 4 5 5.4.2.3 Ways people take their medication 6 The ways in which patients take their medicines were similar in the update 7 studies as in the Pound synthesis. Patients changed their medicines and used 8 medicines in strategic ways but did not necessarily disclose this to health care 9 professionals Deegan (2005) (Deegan, P. E., 2005)). In the sample 10 interviewed by Ogedegbe (2004) (Ogedegbe, Gbenga, Harrison, Melanie, 11 Robbins, Laura et al , 2004) the cost of prescriptions and the effort involved in 12 getting prescriptions reordered and refilled meant people did not take their 13 medication continuously. Studies of patients prescribed HIV medication 14 similarly indicated patients making adjustments to their regimes in keeping 15 with their own beliefs and experiences and not reporting these to health care 16 professionals Campero (2007) (Campero, Lourdes, Herrera, Cristina, 17 Kendall, Tamil et al , 2007). 18 19 In general patients would not report their treatment modifications to health 20 care professionals unless they saw themselves as expert patients. The 21 ubiquity of patients’ alteration of their regimes is indicated by the findings of 22 Aronson (2005) (Aronson, B., 2005). This was a small study of 11 patients 23 who were described as being completely adherent to medication. They were 24 all prescribed short term courses of antibiotics and all patients took all the 25 medication. These patients however did alter the timing of doses to fit in with 26 their schedules, doses were forgotten and then taken when remembered. 27 Wilson 2002 (Wilson, Holly Skodol, Hutchinson, Sally A., and Holzemer, 28 William L., 2002) describes HIV patients making decisions about how they 29 take their medication almost on a dose-by dose basis. This was for a number 30 of reasons and in this study patients’ medicine taking behaviours are 31 described as a result of reconciling incompatibilities which included illness 32 beliefs, the difficulty of the regime and its impact of life. Patients generally 33 described themselves as adherent to health care professionals. Reid (2006) Medicines concordance: full guideline DRAFT (July 2008) page 187 of 373 DRAFT FOR CONSULTATION 1 (Reid, M., Clark, A., Murdoch, D. L. et al , 2006) describes the strategic use of 2 diuretic medication by patients with heart failure- patients changing the timing 3 of medication or omitting the dose according to social and other activities. 4 5.4.2.4 5 The findings of the qualitative studies included in the update of the Pound 6 synthesis (2005) fitted largely within the themes found in the original 7 synthesis. A finding not elaborated in the synthesis and not found in other 8 update studies was medicine-taking experience and behaviours in older adults 9 on complex regimes. Elliott (2007) (Elliott, Rachel A., Ross, Degnan Dennis, 10 Adams, Alyce S. et al , 2007) interviewed 20 patients aged 67-90 about their 11 experience of medicine-taking. Patients were all members of one HMO and 12 took 4-12 medicines. The researchers were particularly interested in how 13 patients on multiple medications make decisions about medicines. The 14 general findings from the study did not differ from themes found in synthesis – 15 patients wished to minimize medication overall, they stopped and started 16 medication to take a break from medication, to check if they were working, to 17 determine the cause of side effects and generally did not disclose this 18 behaviour to their physicians. The patients interviewed did report having made 19 choices between which medicines they would decide to take, this included 20 choices between medicines for different disease and choices within diseases. 21 When choosing between medicines for different diseases patients chose to 22 take the medicine for the disease they feared most or that which gave 23 symptomatic relief. Choices otherwise were influenced by symptom control, 24 side effects, medication cost, negative health experience, illness beliefs and 25 acceptability of medicine (i.e. taste etc). Illness beliefs dominated more 26 general factors such as influence of family, friends, health providers and the 27 media. Complexity of regime did not affect choice. Cost was a factor even 28 when not related to financial hardship, patients appeared to resent the cost of 29 medications. Choices were generally influenced by one dominant factor and 30 less likely to be a result of analysis of multiple factors. 31 Two studies reported on structural issues that interfered with patients’ ability 32 to appraise medication and to receive the information they required to do this. Additional findings from update search Medicines concordance: full guideline DRAFT (July 2008) page 188 of 373 DRAFT FOR CONSULTATION 1 Popa-Lissenau (2004) (Garcia-Popa-Lisseanu, M. G., Greisinger, A., 2 Richardson, M. et al , 2005) reported on the difficulties patients from low 3 incomes with rheumatological disorders had in physically accessing 4 appointments with professions. Mutchler (2007) (Mutchler, Jan E., 5 Bacigalupe, Gonzalo, Coppin, Antonia et al , 2007) reported how for non- 6 English speaking Latinos in the US the difficulty in engaging with health care 7 professionals results in reduced information available to those patients and 8 poorer relationships with professionals which of itself could reduce trust in 9 treatment. 10 5.4.2.5 Experience of medicine taking of minority groups in UK 11 We specifically searched for papers examining experience of patients from 12 minority groups in the UK. Overall their use and experience of medicines was 13 similar to that already described in the evidence review. One additional finding 14 in Erwin (1999) (Erwin, J. and Peters, B., 1999) was a belief by patients of 15 African origin that they had different physiology from white people and that 16 drugs used for treatment of HIV might not be appropriate for them and that 17 they were being discriminated against. This was an issue also raised by 18 African American patients interviewed. Sidat (2004) (Sidat, Mohsin, Fairley, 19 Christopher, and Grierson, Jeffrey, 2007)This patient group were also often 20 involved in church activities and some churches were against use of 21 medication. Lawton (2005) (Lawton, J., Ahmad, N., Hallowell, N. et al , 2005) 22 explored the perceptions of diabetic patients of Indian and Pakistani origin of 23 taking oral hypoglycaemic drugs. These patients beliefs and use of drugs was 24 influenced by their experience of the health system in their country of origin. 25 They distrusted this system and admired the NHS but consequently 26 considered that the drugs available in the UK were likely to be stronger and 27 more efficacious than those available in their country of origin and so they 28 reduced dose and sought to balance effect of drugs by taking in ‘strong’ foods. 29 People of African origin living in South London were interviewed about their 30 use of malaria prophylaxis Morgan (2005) (Morgan-Myfanwy, Figueroa- 31 Muñoz-Jose, 2005). One of the factors influencing use of anti-malarial was the 32 practice of leaving drugs in Africa for family members. Medicines concordance: full guideline DRAFT (July 2008) page 189 of 373 DRAFT FOR CONSULTATION 1 2 5.5 Systematic reviews of barriers and facilitators for individuals in medicine taking 3 Methods of the update 4 5.5.1 5 The GDG requested an additional search for any further systematic reviews of 6 barriers and facilitators for individuals in medicine taking. The search strategy 7 used for the Pound updates searches was applied to this review together with 8 a systematic review filter. The titles and abstracts of records retrieved by the 9 searches were scanned and any potentially relevant publications were 10 obtained in full text. Cross-referencing of all the studies was undertaken to 11 ensure that the search is as comprehensive as possible. The studies were 12 then reviewed to identify the most appropriate evidence to help answer the 13 key questions and to ensure that the recommendations are based on the best 14 available evidence. 15 Types of studies: Systematic reviews 16 Types of participants: people aged above 16 years prescribed medication for 17 a medical condition. 18 Duration of studies: no limit on duration of studies 19 5.5.2 20 The search for systematic review of barriers and facilitators of medicine taking 21 produced two eligible reviews in very diverse patient populations. No 22 systematic reviews examining statistical associations between patient 23 reported factors and actual medicine taking were found. The systematic 24 reviews that were found examined medicine taking in particular population 25 subgroups – patients with TB and patients with HIV on retroviral treatment. It 26 should be noted that although these areas overlapped with the Pound 27 synthesis there was little overlap in papers included highlighting the issues 28 raised by authors of Pound synthesis about locating qualitative literature. Evidence review Medicines concordance: full guideline DRAFT (July 2008) page 190 of 373 DRAFT FOR CONSULTATION 1 The first of these Munro (2007) (Munro, Salla A., Lewin, Simon A., Smith, 2 Helen J. et al , 2007) is a review of qualitative studies which aimed to 3 understand the factors considered important by patients, caregivers and 4 health care providers in contributing to TB medication adherence. The authors 5 used meta-ethnography as in Pound synthesis. The majority of the studies in 6 this review were conducted in developing countries. The emphasis on 7 adherence and inclusion of carers and health care providers perspectives as 8 well as the methodology resulted in an analysis which included structural 9 factors that influence patients medicine taking as well as patient factors. 10 The primary themes that emerged from the included studies were: 1) 11 Organization of treatment and care including access to care, treatment 12 requirements and relationship with the provider; 2) Interpretation of illness and 13 wellness; 3) Financial burden including impact on work, cost of treatment, 14 general poverty; 4) Knowledge attitudes and beliefs about treatment ; 5) Law 15 and immigration; 6) Personal characteristics and adherence behaviour 16 including substance abuse, gender ,religion, motivation; 7) Side effects ; 8) 17 Family, community and household influence. 18 A systematic review by Mills (2006) (Mills, Edward J., Nachega, Jean B., 19 Bangsberg, David R. et al , 2006) examined patient reported barriers and 20 facilitators to adhering to antiretroviral therapy. This analysis included 37 21 qualitative studies and 47 surveys using structured questionnaires or 22 structured interviews. Seventy two studies were conducted in developed 23 countries. Fifty six were from the US and only 3 from the UK. A systematic 24 approach was taken to extracting themes from the qualitative studies and 25 synthesizing the quantitative data and pooling the results. Briefly themes were 26 extracted from qualitative studies and the reviewers then abstracted data from 27 the quantitative surveys to determine if the same issues had been addressed 28 in the surveys. The authors used their own criteria to assess the surveys and 29 these related to the development process and face validity of the 30 questionnaire and the population surveyed. The authors abstracted 31 prevalence of issues as reported in the surveys for their statistical methods. Medicines concordance: full guideline DRAFT (July 2008) page 191 of 373 DRAFT FOR CONSULTATION 1 This technique is called meta-study and is one of a series of methods being 2 developed to bring together findings of qualitative and questionnaire studies. . 3 Barriers identified in both economic settings (developed and developing world) 4 included: fear of disclosure, concomitant substance abuse, forgetfulness, 5 suspicions of treatment, regimens that are too complicated ,number of pills 6 required, decreased quality of life, work and family responsibilities, falling 7 asleep and access to medication. Important facilitators reported by patients in 8 developed nation settings included having a sense of self work, seeing 9 positive effects of anti-retrovirals, accepting their seropositivity, understanding 10 the need for strict adherence, making use of reminder tools, and having a 11 simple regimen. In a further study of adherence rates in sub-Saharan Africa 12 and North America (Mills, Edward J., 2006) Mills comments that the most 13 prevalent barriers to adherence in sub-Saharan Africa are cost, not disclosing 14 HIV status to a loved one or fear of being stigmatized, alcohol abuse and 15 difficulty in following complex drug regimens. 16 Medicines concordance: full guideline DRAFT (July 2008) page 192 of 373 DRAFT FOR CONSULTATION 1 6 Information regarding medicines for patients 2 and practitioners when patients are discharged 3 from inpatient care 4 6.1 Recommendations *listed at the start of the guideline according to chapter 5 6 6.2 Introduction 7 Patients are frequently started on medications when in hospital as an 8 inpatient or when attending outpatient clinics. Transitions between care 9 settings have been recognised as a time when potential errors in medication 10 can occur. NICE and the National Patient Safety Agency (NPSA) have 11 recently produced joint guidance on medicines reconciliation when adult 12 patients are admitted to hospital (www.NICE.org.uk/PSG001). Literature 13 reviews suggested unintentional variances of 30-70% between medicines 14 patients were taking before admission and those prescribed on admission. 15 The GDG considered that patients have the same rights to information and 16 choice when in hospital and community care settings but acknowledged that 17 actual care does not always allow this. 18 Medicines concordance: full guideline DRAFT (July 2008) page 193 of 373 DRAFT FOR CONSULTATION 1 2 6.3 Key clinical question: What information regarding 3 medicines should be provided for patients and 4 practitioners when patients are discharged from 5 secondary care Related Evidence statements references (summary of Evidence into recommendations evidence) The GDG recognised that medications may be initiated in hospital settings in emergency situations and when patients are unwell. They recognised that in these situations discussions of details of medication therapy may not be possible. However as patients condition improves and patients are prepared for discharge they should be offered a full explanation of their medication. This explanation should allow patients to make informed choices about their continued use of medication prescribed. The experience of the GDG was of considerable confusion and lack of information provided to patients and to subsequent providers of care when patients are discharged. Difficulties arise not only in knowledge of what medications have been prescribed for patients but what information patients have been given about their illness and medicines. The GDG based these recommendations on professional opinions and information from expert sources. Medicines concordance: full guideline DRAFT (July 2008) page 194 of 373 DRAFT FOR CONSULTATION 1 Methods of the evidence review 2 6.3.1 3 The evidence review is a narrative review. The GDG requested review of 4 available guidance and reports with a particular emphasis on those where 5 patients rights to information and involvement were given priority. 6 6.3.2 7 The Academy of Royal Medical Colleges is currently preparing consensus 8 guidance on what should be included in hospital discharge summaries. This 9 guidance which will include advice on information about medication will be Evidence review 10 available late in 2008. The WHO produced a report in 2007 on Assuring 11 Accuracy of Medicines at Transitions of Care (www.jcipatientsafety/org/). The 12 emphasis on reports and guidance about medicines reconciliation is on the 13 reduction of medication errors. Patients’ rights to information and involvement 14 in decisions about medicines are not the primary concern of these reports. 15 The WHO report does however state that effective involvement of patients 16 and families in medicines reconciliation is key to reducing errors. They 17 suggest that 18 19 20 patient is in the best position to be aware of all the medications prescribed by multiple caregivers. Consideration should be given to asking patients to put all their 21 medications in a bag and bring it with them whenever going to 22 the hospital or a doctor visit. 23 Patients, family, and caregivers should be encouraged to keep 24 and maintain an accurate list of all medications, including 25 prescription and non prescription medications, herbal and 26 nutritional supplements, immunization history, and any allergic or 27 adverse medication reactions. These medication lists should be 28 updated and reviewed with the patient/family/caregiver at each 29 care encounter. 30 Patients should be taught about the risks of medications, both 31 individually and in combination, with particular attention to Medicines concordance: full guideline DRAFT (July 2008) page 195 of 373 DRAFT FOR CONSULTATION 1 patients on multiple medications prescribed by multiple 2 caregivers. 3 Medicines concordance: full guideline DRAFT (July 2008) page 196 of 373 DRAFT FOR CONSULTATION 1 2 7 Assessment of adherence 3 7.1 Recommendations *listed at the start of the guideline according to chapter 4 5 7.2 Introduction 6 In a systematic review of communication between health care professionals 7 and patients Cox (2004) (Cox, F, Stevenson, N, Britten, N et al , 2004) states 8 that for concordance to occur patients and professionals need to have two- 9 way discussions in which they exchange information and views about 10 medicines. Clearly any initial decision to prescribe medication needs to 11 include a two-way discussion about medicines. However many patients take 12 medications over long periods of time and discussions about these 13 medications need to also consider the patients experience of taking the 14 medication. This includes an assessment or discussion about whether or not 15 the patient is taking the medication and if they are doing this exactly as 16 prescribed or in some other way. 17 18 A number of ways of assessing adherence have been developed. These can 19 generally be described as direct methods or indirect methods. Direct methods 20 are examinations of blood, urine or other bodily fluids for the presence of the 21 drug or a metabolite. Indirect methods do not measure the presence of the 22 drug but use methods such as self report from patients, pill counts, 23 prescription reordering, pharmacy refill records, electronic drug monitoring 24 and therapeutic effect to form an assessment of adherence. In the context of 25 routine clinical practice and of involving patients in decisions about medicines 26 the GDG decided that measures such as self-report were most likely to be off 27 use to health care professionals. The GDG therefore wished to consider the 28 advantages and disadvantages self report in routine clinical practice. Medicines concordance: full guideline DRAFT (July 2008) page 197 of 373 DRAFT FOR CONSULTATION 1 2 7.3 Key Clinical Question: What are the advantages and 3 disadvantages of self-report in assessing patient’s 4 adherence? Related references Evidence statements Evidence into recommendations (summary of evidence) The GDG did not consider that Advantages direct measures of adherence Hawkshead (2007) 1. Self-reporting is the most (Hawkshead, J. and simple and inexpensive method Krousel-Wood, M. of measuring adherence. A., 2007); Gagne (2005) (GagnéCamille, Godin Gaston, 2005); Paterson (2002) (Paterson, David L., Potoski, Brian, and Capitano, Blair, 2002); Miller (2000) (Miller, L. G. and Hays, R. D., 2000); Turner (2002) (Turner, Barbara J., were relevant in routine clinical practice. Indirect methods such as therapeutic effects and prescription ordering and refills are methods which should alert prescribers and dispensers to problems of adherence. In these situations and as part of medication reviews health care professionals need to be able to discuss medicine taking with patients. The GDG made recommendations on how professionals should assess adherence using the review of advantages and disadvantages of self-report. 2002); Farmer (1999) (Farmer, K. C., 1999); Bender (1997) (1997) (Bender, B., Milgrom, H., and Rand, C., 1997); Medicines concordance: full guideline DRAFT (July 2008) page 198 of 373 DRAFT FOR CONSULTATION LaFleur (2004) (La, Fleur J., 2004); Rand (1994) (Rand, C. S. and Wise, R. A., 1994); Miller (2000) (Miller, 2. Self-reporting is quick and L. G. and Hays, R. easy to administer, avoiding the D., 2000); use of sophisticated methodology Farmer(1999) or equipment. (Farmer, K. C., 1999); Paterson (2002) (Paterson, David L., Potoski, Brian, and Capitano, Blair, 2002); Bender (Bender, B., Milgrom, H., and Rand, C., 1997); Rand (1994) (Rand, C. S. and Wise, R. A., 1994); Hawkshead (2007) 3. Self-reporting methods which (Hawkshead, J. and are validated can feasibly be Krousel-Wood, M. used in clinical settings. A., 2007); Bender(1997) (Bender, B., Milgrom, H., and Rand, C., 1997); Hawkshead (2007) 4. Self-reporting can identify (Hawkshead, J. and those who are not adherent. It is Medicines concordance: full guideline DRAFT (July 2008) page 199 of 373 DRAFT FOR CONSULTATION Krousel-Wood, M. most likely those reporting non- A., 2007); Paterson adherence are correct. (2002) (Paterson, David L., Potoski, Brian, and Capitano, Blair, 2002); Farmer (1999) (Farmer, K. C., 1999); Hecht (1998) (Hecht, F. M., 1998); Bennett Johnson (1992) (Johnson, S. B., 1992); George (2007) (George, Johnson, KongDavid, C. M., and Stewart, Kay, 2007) Hawkshead (2007) 5. Self-reporting can gather (Hawkshead, J. and social, situational and Krousel-Wood, M. behavioural factors including A., 2007); Miller revealing patterns of medication (2000) (Miller, L. G. use and what leads to and Hays, R. D., noncompliance. 2000); Rand (1994) (Rand, C. S. and Wise, R. A., 1994); Bennett Johnson (1992) (Johnson, S. B., 1992); Disadvantages George (2007) 6. Self-reporting has the problem Medicines concordance: full guideline DRAFT (July 2008) page 200 of 373 DRAFT FOR CONSULTATION (George, Johnson, of over-estimating adherence. Kong-David, C. M., and Stewart, Kay, 2007); Hawkshead (2007) (Hawkshead, J. and KrouselWood, M. A., 2007); LaFleur (2004) (La, Fleur J., 2004); Turner (2002) (Turner, Barbara J., 2002); Miller (2000) (Miller, L. G. and Hays, R. D., 2000); Hecht (1998) (Hecht, F. M., 1998); Bender (1997) (Bender, B., Milgrom, H., and Rand, C., 1997); Hawkshead (2007) 7. Inaccurate self-reporting can (Hawkshead, J. and be caused by recall bias, social Krousel-Wood, M. desirability bias and errors in A., 2007); self-observation. Gagne(2005) (Gagné-Camille, Godin Gaston, 2005); LaFleur (2004) (La, Fleur J., 2004); Turner (2002) (Turner, Barbara J., 2002); Farmer(1999) (Farmer, K. C., 1999); Bennett Medicines concordance: full guideline DRAFT (July 2008) page 201 of 373 DRAFT FOR CONSULTATION Johnson (1992) (Johnson, S. B., 1992) Paterson (2002) 8. The timeframe of the (Paterson, David L., adherence recollection can affect Potoski, Brian, and the accuracy of the recall. Capitano, Blair, Specifying the time period can 2002); Hecht (1998) help. (Hecht, F. M., 1998); Bennett Johnson (1992) (Johnson, S. B., 1992); Hawkshead (2007) 9. Wording of questions, the way (Hawkshead, J. and a question is asked and the skills Krousel-Wood, M. of the interviewer can either A., 2007); Farmer facilitate or be detrimental to (1999) (Farmer, K. gaining accurate responses. C., 1999); Hecht (1998) (Hecht, F. M., 1998); Turner (2002) 10. Being non-judgmental, giving (Turner, Barbara J., a preamble before adherence 2002); Bennett questions, and asking about Johnson (1992) specific behaviours can help (Johnson, S. B., validity. 1992); Medicines concordance: full guideline DRAFT (July 2008) page 202 of 373 DRAFT FOR CONSULTATION Methods of the evidence review 1 7.3.1 2 This paper includes a narrative summary of the included evidence, structured 3 according to the category of the intervention, following the agreed reviewing 4 protocol: 5 Types of studies – We included literature reviews and systematic reviews 6 only. 7 Types of participants - People prescribed medication for a medical 8 condition. 9 Duration of studies – No time limit was specified. 10 Types of interventions - Any interventions intended to change adherence to 11 prescribed medication which reviews studies which focus on self-report 12 advantages and disadvantages. 13 Types of outcome measures - No outcome measures specified. 14 7.3.2 15 The searches mainly returned literature reviews, rather than systematic 16 reviews, therefore details (of the various studies mentioned in these reviews) 17 were not always given and some only mentioned the studies briefly. 18 Garber (2004) (Garber, Mathew C., Nau, David P., Erickson, Steven R. et al , 19 2004) produced a systematic review on the concordance of self-report with 20 other measures of medication adherence. They searched a number of 21 databases and identified 2757 articles. The inclusion criteria were to include 22 studies where at least 2 adherence measures were used, one of which being 23 a self-report measure, the other a non self-report measure. The self-report 24 measures included questionnaires, diaries or interviews and were categorised 25 under these. They found 86 unique comparisons, mostly interviews (57%), 26 questionnaires (27%) and diaries (17%). The nonself-report measures were 27 electronic measures (35%), pill count or canister weight (26%), a plasma drug 28 concentration (20%) a claims-based measure (13%) and a clinical opinion 29 (6%). 43% of the pairings of self-report and nonself-report measures were Evidence review Medicines concordance: full guideline DRAFT (July 2008) page 203 of 373 DRAFT FOR CONSULTATION 1 highly concordant. In the majority (45/59) of those not highly concordant the 2 self-report measure showed higher adherence compared to the nonself-report 3 measure but this varied widely depending on type of self-report measure. 31% 4 of the interviews were highly concordant with nonself-report measures. Diaries 5 (71%) and questionnaires (55%) were much more likely to be highly 6 concordant with non-self report measures. It also depended on the nonself- 7 report measures. Self-reporting had higher concordance with other types of 8 nonself-report measures (58%) than electronic measures (17%). Interviews 9 showed the least concordance with electronic measures. It should be noted 10 that this is a comparison between measures which can not fully evaluate the 11 accuracy of any of the measures. 12 In summary, questionnaires and diaries were more concordant with other 13 measures. However this is a comparison with other measures which cannot 14 fully evaluate the accuracy of any of the measures. 15 George (2007) (George, Johnson, Kong-David, C. M., and Stewart, Kay, 16 2007) conducted a literature review to assess adherence of COPD patients 17 with disease management programs. They searched OVID and International 18 Pharmaceutical Abstracts. They did not report the inclusion/exclusion criteria 19 or how many studies were retrieved. 20 They found that self-reporting of missed doses (by questionnaire) 21 underestimated non-adherence compared to more objective measures e.g. 22 capsule count (Dompeling, 1992), inhaler weights (Rand, 1995) and electronic 23 monitoring (Rand, 1992; Braunstein, 1996; Simmons, 2000). Self-report was 24 shown to have moderate reliability compared to objective measures such as 25 canister weight (Rand, 1995) and electronic monitoring (Gong, 1988; Nides, 26 1993; Bosley,1995). 27 Self-reporting of non-adherence of medication for COPD has shown 28 satisfactory reliability, when compared to objective measures (Dolce, 1991; 29 Nides, 1993; Rand, 1995). Self-report is commonly criticised for 30 overestimating adherence and poor reliability yet those who report non- Medicines concordance: full guideline DRAFT (July 2008) page 204 of 373 DRAFT FOR CONSULTATION 1 adherence are likely to be telling the truth (Haynes, 1980; Inui, 1981; Choo, 2 1999; Erickson, 2001). 3 In summary, self-reporting questionnaires underestimate non-adherence but 4 have shown reliability and are usually correct for those who say they are non- 5 adherent. 6 Hawkshead (2007) (Hawkshead, J. and Krousel-Wood, M. A., 2007) 7 presented a narrative review of the advantages and limitations of methods for 8 measuring adherence in hypertensive patients. No mention is given to how 9 they searched for these studies or decided to include/exclude. They state that 10 self reporting is the simplest method for assessing medication adherence and 11 can include patient diaries, interviews during office visits and adherence- 12 specific questionnaires. ‘Several multi-item questionnaires have been 13 developed and tested in outpatient settings with the explicit aim of 14 ascertaining valid and reliable estimates of adherence to antihypertensive 15 medications’, of which many have reported high measures of validity and 16 reliability (Morisky, 1986; Kim, 2000; Shea, 1992; Hyre, 2007). There are 17 three previously validated self-reported medication-adherence instruments – 18 the Medication Adherence Survey (MAS), the Brief Medication Questionnaire 19 (BMQ) and the Medical Outcomes Study (MOS). 20 Validated self-report measures can feasibly be used in clinical settings and 21 help to identify those who are non-adherent, and intervene to increase this 22 (Harmon, 2006). The advantages they state are that self-report is simple and 23 economical; it can also gather social, situational, and behavioural factors 24 which can impact on adherence. The disadvantages are the possibility that 25 there could be recall bias, over-estimation of compliance and responses which 26 are socially acceptable. Validity can also depend on the skills of the 27 interviewer as well as the question construction and timeframe (Farmer, 1999 28 and Wang, 2004). It is suggested that self-report could be combined with 29 objective information, e.g. prescription-fill data, to improve adherence 30 measurement. Medicines concordance: full guideline DRAFT (July 2008) page 205 of 373 DRAFT FOR CONSULTATION 1 In summary, some self-reporting questionnaires have been validated and can 2 be simple and feasible to use in clinical settings and identify non-adherers. 3 However they can have biases and overestimate adherence. 4 Gagne (2005) (Gagné-Camille, Godin Gaston, 2005) reported on how to 5 improve self-report measures for non-adherence to HIV medications, with 6 particular attention to techniques that can be applied with questionnaires 7 administered in clinical practice. Questionnaires are inexpensive and 8 convenient and can be conducted in clinical and research settings. But can 9 vary in terms of accuracy. According to many authors, forgetfulness (Brooks, 10 1994; Hayes & DiMatteo, 1987; Holzemer, 1999; Rand, 2000; Svarstad, 1999) 11 and social desirability (Felkey, 1995; Gordis, 1969; Gray, 1998; Rand, 2000; 12 Svarstad, 1999) are main factors leading to inaccurate self-reporting of non- 13 adherence. Social desirable answers can depends on how much the patient 14 perceives the desirability of the behaviour to be. Those behaviours perceived 15 as undesirable are under-reported and behaviours perceived as desirable can 16 be over-reported (Cannell 1979; Fowler, 1995). There are techniques 17 suggested for minimising forgetfulness and social desirability (Cannell, 1979; 18 Fowler, 1995; Sudman & Bradburn, 1974; Sudman & Bradburn, 1982) 19 although methods to reduce these are not well-documented, are often derived 20 from clinical practice than controlled experimental studies and their reported 21 effectiveness is inconsistent. 22 Suggestions were made to reduce socially desirable answers: 23 HCPs (Eldred, 1998; Gordillo, 1999). 24 25 28 Explaining that there are no right or wrong answers (Des Jarlais, 1999; Chesney, 1990). 26 27 Assuring confidentiality and that information will not be available to How the question is asked (Ickovics, in Eldred, 1998; Chesney, 1999; Svarstad, 1999). Medicines concordance: full guideline DRAFT (July 2008) page 206 of 373 DRAFT FOR CONSULTATION 1 Wording the question to increase the likelihood of gaining certain 2 desired answers, such as non-adherence (loading the question) 3 (Sudman, 1982; Bradburn, 1982; Allaire, 1988). 4 5 Open-ended questions can avoid the pitfalls of response categories (Schwarz, 1985; Sudman, 1982). 6 Open-ended questions have been used in studies of HIV (e.g. Chesney, 7 1990) and for measuring adherence/non-adherence (e.g. Svarstad, 1999). 8 Open-ended answers have shown to be less affected by social desirability 9 than close-ended answers (Sudman, 1974). Sudman (1974) also found 10 that open-ended questions were less affected by forgetfulness and 11 remembering it happening more than it actually did. 12 Recall can be aided by: 13 meaning and one idea (Sudman, 1982); 14 15 Aided-recall techniques such as memory cues may be useful (Sudman, 1982). 18 19 Words should not have blame implications (Averitt, in Eldred, 1998). 16 17 Item wording, using familiar words and words that have only one Specifying a reference time period, especially a recent and short 20 time frame can aid forgetfulness (Brooks, 1994; Chesney, 1999; 21 Holzemer, 1999; Sudman, 1982). 22 However there is the problem of the time period being too short and not 23 accurately representing the adherence level, as adherence varies over time 24 (Chesney, 1997b; Gray, 1998; Kastrissios, 1998). This could be solved by 25 using a short period of time and administering the questionnaire a number of 26 times over the period. However, this could lead to less motivation and could 27 be costly. Shorter periods of reference could be used when administering the 28 questionnaire only once. According to episodic and semantic memory it may Medicines concordance: full guideline DRAFT (July 2008) page 207 of 373 DRAFT FOR CONSULTATION 1 be best to ask more precise information about the past few days and less 2 specific information from a longer time period. 3 In summary, self-reporting by questionnaire can have biases such as social 4 desirable responses and recall bias. These biases can be minimised using 5 certain techniques. 6 LaFleur (2004) (La, Fleur J., 2004) conducted a brief narrative review of 7 methods to measure compliance with medication regimens. No search or 8 inclusion/exclusion criteria were given. They state that self-report is the most 9 popular method for assessing compliance as it is inexpensive but is often 10 unreliable (Myers, 1998). Self-report can include patient interviews or self- 11 report surveys. When compared to objective measures e.g. electronic 12 monitoring devices or drug level monitoring of compliance self-reporting has 13 shown to over-report compliance over 50% of the time (Spector, 1986; Gordis, 14 1969; Waterhouse, 1993; Straka, 1997). It is also often inaccurate for those 15 reporting non-compliance with medication-taking. In Kwon (2003) a 16 comparison of self-reporting of antidepressant use with prescription claims 17 showed a 20% difference in those reporting non-adherence to 18 antidepressants. The reasons for any discrepancies with other measures 19 could be that patients do not understand regimens, know indications for their 20 medicine, or not report behaviours perceived as not socially-acceptable, or 21 forgetting of non-compliance. No references were given for these assertions. 22 In summary, self-report by interviews or surveys can be inexpensive but can 23 be unreliable and over-report compliance. Those who report non-compliance 24 can also be inaccurate. There could be biases such as social desirability, 25 recall and not understanding medication regimes. 26 Turner (2002) (Turner, Barbara J., 2002) reviewed literature to compare 27 various measures of adherence to Antiretroviral Therapy. This was a narrative 28 review with no details of search/inclusion criteria. They state that self-reports 29 are less complex but that there can be problems with recall over long time 30 periods. Many studies use self-report over the past 4 days but additional Medicines concordance: full guideline DRAFT (July 2008) page 208 of 373 DRAFT FOR CONSULTATION 1 questions may be needed, e.g. about weekends, as this tends to be a difficult 2 time for adherence. 3 All types of self-reporting overestimate adherence compared to other 4 measures (Arnsten, 2001; Golin, 1999; Melbourne, 1999). Even those who 5 report missing doses tend to overestimate adherence compared to other 6 measures (Wagner, 2000). Social desirability biases can contribute. Those 7 who report problems with adherence usually have poorer adherence with 8 other measures (Haynes, 1980). Those who report non-adherence appear 9 responsive to interventions, and are important to identify (Haynes, 1980). 10 The validity can be increased with a preamble before questions about 11 adherence in order to reassure patients that information will not be held 12 against them and that non-adherence is common. Audio computer-assisted 13 self-interviewing is suggested for more sensitive topics (Metzger, 2000; 14 Gribble, 2000). 15 In summary, all types of self-report overestimate adherence, even with those 16 who report non-adherence and biases such as social desirability can occur. 17 Certain techniques could be used to minimise these biases. 18 Paterson (2002) (Paterson, David L., Potoski, Brian, and Capitano, Blair, 19 2002) conducted a brief narrative review to ascertain how adherence to 20 antiretroviral medicine should be measured. The methods reported were 21 electronic monitoring, pill counts, pill recognition, review of pharmacy records, 22 patient self-report, biological parameters, therapeutic drug monitoring and 23 provider prediction of adherence. They noted that how a question is asked 24 can influence self-report of adherence (i.e. in face-to face inquiry or patient- 25 completed questionnaires). A non-judgemental stance can help and this can 26 be achieved by a preamble before the questions to show that they are not 27 being judged and are looking for honest answers (Turner, 2001). 28 Another disadvantage of self-report (face-to-face interview) is that periods 29 shorter than 7 days are not long enough to determine the percentage of 30 adherence likely, however some patients may not correctly report adherence Medicines concordance: full guideline DRAFT (July 2008) page 209 of 373 DRAFT FOR CONSULTATION 1 for 7 day periods. They state that additional questions may be necessary to 2 counteract this e.g. about adherence at the weekend. 3 One method to counteract problems with gaining honest answers is computer- 4 assisted self-interviewing (Bangsberg, 2001) or diary. Diaries hold advantage 5 as they can be inexpensive and accurate. Their disadvantage is that some 6 may complete them retrospectively or not at all. 7 Paterson (2002) asserts that self-report is ‘likely to be the simplest means of 8 assessing adherence’ and so the reliability is important to assess. Adherence 9 was found to be ‘considerably higher’ than that measured by electronic 10 monitoring or pill count (Liu, 2001). Self-report overestimates adherence. It is 11 most useful in those who admit to being poor adherers (Murri 2000). They 12 conclude that electronic monitoring devices are the closest to a gold standard 13 in adherence measurement. 14 In summary, various self-reporting measures were reported and interviews 15 may be too late for recall or may be too early to gain useful adherence 16 information. Diaries are inexpensive and can be more accurate as there is no 17 recall bias however they may not be completed or completed retrospectively. 18 Self-report can overestimate adherence but can identify those who report non- 19 adherence. 20 Miller (2000) (Miller, L. G. and Hays, R. D., 2000) reviewed current literature 21 on measuring adherence to Antiretroviral Medications in clinical trials. They 22 report that the simplest method of measuring adherence is self-report. But 23 there is no standardised instrument. Self-reported surveys are quick and avoid 24 sophisticated methodology or equipment and are inexpensive compared to 25 other methods of measurement. They have limitations, such as significantly 26 exceeding adherence measured by other objective methods (Bond, 1991; 27 Stratka, 1997; Cramer, 1991). HIV studies also confirm this (Golin, 1999; 28 Arnsten, 2000; Paterson, 1999; Bangsberg, 1999). Interviews and surveys 29 often promote socially acceptable responses (DiMatteo, 1982). Less adherent 30 patients report higher adherence than they actually had (Bond, 1991). 31 Memory can also affect the accuracy of reporting adherence. Most surveys Medicines concordance: full guideline DRAFT (July 2008) page 210 of 373 DRAFT FOR CONSULTATION 1 use broad response categories to report the proportion of pills taken, thus 2 small degrees of nonadherence is hard to distinguish with self-report. The 3 information is useful, but accuracy is limited and biased towards higher 4 adherence. 5 However, self-reported non-adherence has been associated with worse 6 virologic outcomes (Demasi, 1999; Bangsberg, 1999; Duong, 1999; Murri, 7 1999; Le Moing, 1999) and as an independent predictor of clinical response to 8 HAART when controlling objective virologic and immunologic markers 9 (Montaner, 1999). Therefore it can provide information that explains variation 10 in clinical response to antiretroviral therapy which is not explained by other 11 clinical factors. 12 In summary, self-report surveys are simple and inexpensive but can 13 overestimate adherence. Interviews and surveys can have social desirability 14 and recall biases. Also as categories are large, small degrees of non- 15 adherence are hard to detect. There is no standardised instrument. However it 16 can explain variation in clinical responses to ART. 17 Farmer (1999) (Farmer, K. C., 1999) conducted a review of methods for 18 measuring and monitoring medication regimen adherence in clinical trials and 19 clinical practice. They searched Medline for the years 1990 to 1999 and 20 retrieved 2630 articles regarding patient compliance. They found that forms of 21 self-report included questioning/interrogation and the use of diaries and 22 survey instruments. They tabulated the various methods for assessing 23 adherence and their advantages and disadvantages. Patient interviews are 24 easy to use and inexpensive but the patient can be influenced by question 25 construction and interviewer’s skill. Adherence questionnaires are easy to 26 administer (on site, mail, telephone), can be validated and may explain patient 27 behaviour. However there is a lack of continuous data and the accuracy is 28 instrument dependent. 29 Patient interviews are considered the most unreliable for assessing adherence 30 (Grymonpre, 1998; Matsui, 1994; Craig, 1985; Straka, 1997; Park, 1964; Inui, 31 1981; Gordis, 1969). Those who report non-adherence are usually correct, Medicines concordance: full guideline DRAFT (July 2008) page 211 of 373 DRAFT FOR CONSULTATION 1 whereas those who say they are adherent may not be (Cramer, 1991). 2 However it can depend on the method used and how it is used. Assessing 3 self-reporting is difficult mainly because there are so many methods. The 4 interviewer’s skill and the construction of the questions can affect the 5 accuracy and validity of self-report. The relationship and communication 6 between the HCP and patient have shown to significantly affect compliance 7 (Davis, 1969). Highest compliance was found with those who joked, laughed 8 and sought suggestions from their GP. The wording of questions can affect 9 the response, and implications of blame can encourage biased responses 10 (Ross, 1991). Some answers are socially desirable and concealed their real 11 behaviour (Sherbourne, 1992). It is hard to assess studies of interviews as the 12 way they are asked could bias the result. Stewart (1987) looked at 2 13 compliance questions in an interview to assess medication-taking behaviour. 14 Comparing the results to pill counts, the questions had a specificity of 69.8% 15 and sensitivity of 80%, therefore an overall 74.5% accuracy. The time frame 16 used for recall can differ, some researchers do not specify, others are 7-10 17 days and some are a month (Grymonpre, 1998; Dirks, 1982; Straka, 1997). 18 To correct these problems some researchers have tried to construct a 19 standardised questionnaire for measuring adherence. For example Morisky 20 (1986) developed a 4-item questionnaire specific to medication regimen 21 adherence. It was assessed on unidimensionality and reliability and 22 concurrent validity with blood pressure control. The instrument’s sensitivity 23 was 81% and specificity 44%. It was not found to be efficient at predicting 24 poor adherence (Morisky, 1986). 25 In summary, a few methods of self-report were looked at. Interviews are 26 simple and inexpensive, but can depend on the interviewer. Questionnaires 27 can be administered in a variety of methods, but are considered the most 28 unreliable. Those who say they are non-adherent are usually correct but many 29 who say they were adherent may not be. 30 Bender (1997) (Bender, B., Milgrom, H., and Rand, C., 1997) conducted a 31 literature review to assess non-adherence in asthmatic patients. A search of 32 Medline was made from 1990 to 1997 of all pertinent articles, preferably Medicines concordance: full guideline DRAFT (July 2008) page 212 of 373 DRAFT FOR CONSULTATION 1 controlled studies. Self-report measures can be collected by interview, diaries 2 and questionnaires but o validated adherence-specific questionnaire is 3 commonly used as they are often too specific. Self-report measures are 4 simple, inexpensive and usually brief and so they are commonly used to 5 measure adherence. Especially in the clinical setting they are the best 6 measure for collecting information of beliefs, attitudes and experiences with 7 medication regimes. Accuracy with other measures is highly variable. Spector 8 (1986), Coutts (1992) and Gibson (1995) compared asthmatics self-reporting 9 of inhaler usage with electronic medication monitoring devices and they 10 showed that asthma diaries usually overestimate adherence. Demands of the 11 setting can influence the usefulness and reliability of the information gained 12 from self-reporting. These can be a desire to please on the part of the patient 13 and HCP skill and sensitivity in eliciting self-reports. When collected well it can 14 give good insight into patients’ problems with adherence. And as they are 15 unlikely to identify themselves as nonadherers, this helps identify the 16 nonadherers (Coutts, 1992; Spector, 1986; Dolce, 1991; Morisky, 1990). 17 In summary, self-report measures are simple, inexpensive, brief and the best 18 way of collecting information in the clinical setting. However diaries 19 overestimate adherence and the demands of the setting can influence the 20 usefulness and reliability of the measure. 21 Rand (1994) (Rand, C. S. and Wise, R. A., 1994) reported in a narrative 22 review on measuring adherence to asthma medication regimens. They did not 23 state search or inclusion criteria. 24 They state that self-report is the most inexpensive and quick way of 25 measuring adherence (Soutter, 1974). The possible advantage of diary cards 26 is that they can measure adherence across time and can reveal patterns 27 between the disease exacerbation and compliance with the medication. As 28 there are many drugs used within asthma prescribing, it can help to see the 29 adherence of certain drugs rather than just overall. It can also specifically 30 assess overuse, inappropriate use or erratic use of medications as well as 31 triggering events for the need for medication e.g. in Kesten (1991). Asthma 32 diaries may share commonalities but there is no standardised diary as such in Medicines concordance: full guideline DRAFT (July 2008) page 213 of 373 DRAFT FOR CONSULTATION 1 research. A disadvantage of asthma diaries can be they may be complex and 2 time-consuming. Also criteria of acceptable adherence may differ from patient 3 to patient. One way to evaluate the level of adherence is to use trained, 4 masked, medical personnel to score the compliance. It is preferable to 5 develop standardised compliance criteria for all raters and train them by a 6 standardised protocol and make sure there is interrater reliability. 7 Many studies have used questionnaires to collect clinic or follow-up data of 8 patient adherence (Bailey, 1987; Kinsman, 1980; Dolce, 1991), mainly 9 designed for a particular research project. Many have adherence questions 10 within a larger questionnaire, such as the 76-item Revised Asthma Problem 11 Behaviour Checklist for adults. Rand (1994) points out that both asthma 12 diaries and self-report are the most common for assessing asthma medication 13 adherence but that these instruments, because they are not standardised or 14 not published so they rarely have validity and reliability assessed. Except for 15 the Medication Adherence Scale and Inhaler Adherence Scale (Kinsman, 16 1980; Dolce, 1991; Bailey, 1990), which are six-item scales based on 17 Morisky’s work (1990). This instrument was found to have a Chronbach’s 18 alpha of 0.76 and 0.69 and was concordant with outcome measures in the 19 UAB adult asthma study. 20 The limitations of self-report have been mentioned by many authors (Masur, 21 1981; Mawhinney, 1991; Cramer, 1989; Rand, 1992). When compared to 22 objective measures it varies highly on the degree of accuracy (Gordis, 1966; 23 Mattar, 1974). Diary self-reports were compared to electronic medication 24 monitoring device to measure adherence to asthmatic medication by Spector 25 (1986). The findings were that all patients self-reported using the inhaler on 26 certain days, whereas the measured medication suggested just over half did 27 so. Adding a diary can add more complexity to the patient regime than there 28 all ready is. It has been shown that the greater the complexity of a regime the 29 lower the compliance (Masur, 1981). Some participants alter their records of 30 medication use to appear compliant (Mawhinney, 1991; Rand, 1992). This can 31 be improved if they also have reporting by the family/partner of the patient 32 (Paulson, 1977). Medicines concordance: full guideline DRAFT (July 2008) page 214 of 373 DRAFT FOR CONSULTATION 1 Self-reporting can also depend on the individual patient or practitioner. For 2 example elderly patients may have memory impairment, especially when 3 taking many medications and not report accurately. Long-term usage may be 4 forgotten but able to recall recent usage. The skill and sensitivity of the HCP 5 can also play a role in how much information is given and the reliability of it. 6 When collected carefully it could be very god insight into the problems of a 7 patient’s adherence. Also it is unlikely that patients will represent themselves 8 as non-adherers (Gordis, 1976) so it will identify non-adherers correctly. 9 In summary, self-report is inexpensive and quick and diaries can measure 10 adherence across time and reveal any patterns and assess overuse of 11 medication. However there is no standardised diary and it can be complex 12 and time consuming. If there is no standardised questionnaire of diary then no 13 validity or reliability are assessed. Therefore there is variation on accuracy, 14 depends on the individual or practitioner. 15 Bennett Johnson (1992) (Johnson, S. B., 1992) conducted a narrative 16 literature review of adherence measurement in diabetes management. No 17 search or inclusion criteria was given. 18 They point out that self-report of regimen adherence are often mistrusted. 19 Patients may say one thing but do something completely different, often 20 because of what they think the doctor wants to hear. However noncompliance 21 self-reporting appears more valid than self-reporting of compliance (Diehl, 22 1987). Asking about specific behaviours can lead to better adherence data 23 (Cerkoney, 1980; Cox, 1984; Shlenk, 1984; Brownlee-Duffeck, 1987; Hanson, 24 1987; Hanson, 1987; Hanson, 1987; Hanson, 1988; Hanson, 1990). There 25 have only been a few that have looked at the reliability of these reports 26 (Hanson, 1987 and Hanson, 1988). If asked to report their specific behaviours 27 over a certain time period, the data can be good quality (Glasgow, 1987; 28 Johnson 1986). Multiple interviews are recommended to ensure 29 representation of adherence behaviours. Medicines concordance: full guideline DRAFT (July 2008) page 215 of 373 DRAFT FOR CONSULTATION 1 One disadvantage with self-reporting is problems of memory recall. Where 2 possible a significant other should additionally be interviewed regarding the 3 patient’s behaviour. 4 The advantages of self-report are numerous, as reliable information can be 5 obtained; interviews can be done over the telephone making them accessible; 6 the patient does not have to do very much apart from give their time for an 7 interview. They however do need trained interviewers, or with multiple 8 interviews and multiple patients the process can take a lot of time and effort. 9 No references were made for these assertions. 10 In summary, self-reporting of non-compliance is likely to be more valid, 11 whereas compliance reporting is not valid. They can ask about specific 12 behaviours and find out about what leads to non-compliance. It is easy for the 13 patients to do and interviews can be done by phonecall. However there are 14 biases with recall and people may say one thing but do another and there can 15 be errors in reporting e.g. self-observation skills. 16 Dunbar (1989) (Dunbar, J., Dunning, E. J., and Dwyer, K., 1989) reviewed the 17 methods to assess adherence to arthritis medication with a review that 18 included ‘16 representative studies of compliance’. No inclusion/exclusion 19 criteria or search details were given. 20 They noted that a major problem is the accuracy of reporting, with poor 21 compliance usually underreported. One issue is the memory decay when 22 assessing adherence (Farr, 1987). Effects, such as not realising the 23 diminishment of higher adherence levels has occurred and moving past 24 events forward in perception can all lead to inaccuracy. Motivational factors 25 are also important, errors in reporting can be due to self-observation skills, 26 especially when the compliance behaviour is itself variable. Misconceptions of 27 the regimen may lead to errors through inaccurately labelling events compliant 28 or noncompliant. 29 Self report has advantages in that it can identify some noncompliance in a 30 cost-efficient manner and permits an in-depth study of the types of errors that Medicines concordance: full guideline DRAFT (July 2008) page 216 of 373 DRAFT FOR CONSULTATION 1 patients can make which leads to non-compliance. It also has reasonable 2 sensitivity and specificity in discriminating compliers from non-compliers. 3 In summary, self-reporting can mean poor compliance is underreported, there 4 can be recall bias and self-observation skills may be erroneous. It is cost- 5 efficient and can identify non-compliers. 6 Hecht (1998) (Hecht, F. M., 1998) reported briefly with a narrative review on 7 measures for HIV adherence in clinical practice. Sackett (1975) compared 8 self-report to pill counts. Of those that reported having less than 80% 9 adherence, 95% were found non-adherent by pill count. Those reporting that 10 they were adherent over 80% of the time, 34% were shown to be non- 11 adherent by pill count. Gilbert and Sackett’s studies, suggest that self-report is 12 more accurate than physician assessment. Thus if HCPs want to know if 13 patients are taking ART, they need to ask them rather than relying on their 14 judgement. When they say they are missing medication, believe them, as this 15 is mostly the truth. Patient self-report tends to overestimate adherence. Those 16 who report missing doses infrequently may have a significant problem of non- 17 adherence. 18 Hecht (1998) says that what matters is how HCPs ask the questions. Stating it 19 should be in a specific, non-judgmental way and one that allows them to 20 disclose non-adherence. Therefore, questions should not imply that they are 21 wrong if they do not take their medication the way they are ‘supposed to’. A 22 time period must also be specified. No references given for these 23 conjectures. 24 Self-report is more accurate than physician’s judgement alone. It tends to 25 overestimate adherence. It depends on how the questions are asked and a 26 time period must be specified. 27 28 29 Medicines concordance: full guideline DRAFT (July 2008) page 217 of 373 DRAFT FOR CONSULTATION 1 8 prescribed medication 2 3 Interventions to increase adherence to 8.1 Recommendations *listed at the start of the guideline according to chapter 4 5 8.2 Introduction 6 Adherence is defined as ‘the extent to which the patient’s behaviour matches 7 agreed recommendations from the prescriber’ (Horne, R, Weinman, J., 8 Barber, N. et al , 2005). Adherence describes patient behaviour in the actual 9 taking of medicines. This definition of adherence presumes that the patient 10 has reached some agreement with the health care professional about the 11 prescribed medication. The Guideline Development Group were interested in 12 interventions that would support patients in taking of medicines following 13 agreement with the health professional. 14 Non-adherence can be intentional or unintentional. Non-adherence is 15 unintentional if the patient does not take the medication, for example, due to 16 forgetfulness or not being able to access the medication because of problems 17 with packaging and dexterity. Non adherence is intentional when the patient 18 makes a decision to either not take the medicine as previously agreed or to 19 take it in a way other than recommended by the prescriber because of their 20 own beliefs and appraisals of the medicine and medicine taking. Both 21 intentional and unintentional non-adherence can occur regarding the amount 22 or duration of missed medication e.g. a single dose may be missed, a patient 23 may miss several days of medication or a patient may permanently stop 24 taking medication. In some patients non-adherence takes the form of the 25 patient reducing or increasing the dose of prescribed medication rather than 26 omitting it. 27 Medicines concordance: full guideline DRAFT (July 2008) page 218 of 373 DRAFT FOR CONSULTATION 1 8.3 Methods 2 The aim of the evidence review is to identify the most relevant, published 3 evidence to answer the key clinical questions generated by the GDG. Due to 4 time constraints, exhaustive systematic reviews (see the Methods of the 5 Cochrane Review) were not undertaken. However, the evidence reviews were 6 undertaken using systematic, transparent approaches following the Guidelines 7 Manual 2007.The following evidence reviews were developed as an update to 8 the 2005 Cochrane Review14 titled: “interventions for enhancing medication 9 adherence”. The Haynes review aimed to summarise the results of 10 randomised controlled trials (RCTs) of interventions to help patients follow 11 prescriptions for medications for medical problems, including mental disorders 12 but not addictions. For short-term treatments, not enough studies were 13 retrieved on any individual disease condition to allow grouping by disease. 14 The main results were: For short-term treatments, four out of nine 15 interventions (reported in eight RCTs) showed an effect on both adherence 16 and at least one clinical outcome, while one intervention reported in one RCT 17 significantly improved patient compliance, but did not enhance the clinical 18 outcome. For short-term treatments, three of the interventions were related to 19 counselling (mainly about the importance of adherence and information on the 20 medicines). For long-term treatments, 26 out of 58 interventions reported in 49 21 RCTs were associated with improvements in adherence, but only 18 22 interventions led to improvement in at least one treatment outcome. Almost all 23 of the interventions that were effective for long term care were complex, 24 including combinations of more convenient care, information, reminders, self- 25 monitoring, reinforcement, counselling, family therapy, psychological therapy, 26 crisis intervention, manual telephone follow-up, and supportive care. Even the 27 most effective interventions did not lead to large improvements in adherence 28 and treatment outcomes. Six studies showed that telling patients about 29 adverse effects of treatment did not affect their adherence. 30 These results led the authors to conclude that for short-term treatments 31 several quite simple interventions increased adherence and improved patient 14 first published in 2002 Medicines concordance: full guideline DRAFT (July 2008) page 219 of 373 DRAFT FOR CONSULTATION 1 outcomes. Current methods of improving adherence for chronic health 2 problems are mostly complex and not very effective, so that the full benefits of 3 treatment cannot be realised and that further research concerning innovations 4 to assist patients to follow medication prescriptions for long-term medical 5 disorders is required. 6 A more recent version of this Cochrane Review has now been published in 7 2008. The main results were: for short-term treatments, four out of ten 8 interventions reported in nine RCTs showed an effect on both adherence and 9 at least one clinical outcome, while one intervention reported in one RCT 10 significantly improved patient adherence, but did not enhance the clinical 11 outcome. For long-term treatments, 36 out of 81 interventions reported in 69 12 RCTs were associated with improvements in adherence, but only 25 13 interventions led to improvement in at least one treatment outcome. Almost all 14 of the interventions that were effective for long-term care were complex, 15 including combinations of more convenient care, information, reminders, self- 16 monitoring, reinforcement, counselling, family therapy, psychological therapy, 17 crisis intervention, manual telephone follow-up, and supportive care. Even the 18 most effective interventions did not lead to large improvements in adherence 19 and treatment outcomes. In this update the authors also for short-term 20 treatments where several quite simple interventions increased adherence and 21 improved patient outcomes the challenge is that there are inconsistent results 22 from study to study with less than half of studies showing benefits. 23 The evidence in the Cochrane Review was organized by disease, however as 24 this guideline is intended to be a generic document we re-arranged the 25 evidence according to the type of intervention. We also conducted additional 26 specific searches for several questions. This is described below. 27 The titles and abstracts of studies retrieved by electronic searches were 28 scanned for relevance to the topic on interventions to increase adherence. 29 Any potentially relevant publications were obtained in full text. These were 30 then reviewed to identify the most appropriate evidence and were then 31 allocated to the relevant key clinical questions. Following this, the assessment Medicines concordance: full guideline DRAFT (July 2008) page 220 of 373 DRAFT FOR CONSULTATION 1 of study quality; synthesis of the results; and grading of the evidence was 2 undertaken. 3 While the GDG were interested in any intervention that might be useful in 4 supporting adherence, the key clinical questions agreed by the GDG included 5 questions on a number of specific interventions. We initially conducted one 6 broad search in order to find evidence on interventions to increase adherence. 7 One broad search allowed us to pick up any intervention designed to increase 8 adherence and then allocate the evidence to the respective clinical question. 9 This reduced the duplication of sifting and reviewing of the evidence. This 10 search was supplemented by specific searches in areas considered important 11 by the GDG where we also included some observational studies. These were 12 areas where the broad search left considerable uncertainty but the 13 interventions were considered either potentially important in clinical practice or 14 areas where popular preconceptions may exist. 15 These were: Does change in dosing regime affect adherence to prescribed 16 medication? 17 Does drug formulation/packaging affect adherence to prescribed 18 medication? 19 Do medication reviews affect adherence to prescribed 20 medication? 21 Do prescription charges/costs affect adherence to prescribed 22 medication? 23 Do dosette boxes affect adherence to prescribed medication? 24 25 26 8.4 Evidence to recommendations : difficulties in 27 interpreting studies on interventions to improve 28 adherence 29 Given the advances of medical therapies and the increase in prescribing and 30 drug use in the last decades, it could be expected that this has been Medicines concordance: full guideline DRAFT (July 2008) page 221 of 373 DRAFT FOR CONSULTATION 1 accompanied by a greater understanding of the processes of adherence and 2 non adherence and the effectiveness of interventions to promote adherence. 3 In general however the current body of literature on adherence is of poor 4 methodological quality and was considered inadequate to answer many of the 5 GDG questions on interventions to increase adherence. Medicine taking does 6 not appear to be recognized by many researchers as a complex human 7 behaviour which should be studied using complex interventions in line with the 8 MRC framework (2000) (Medical Research Council, 2000). Specific issues 9 affecting quality of the studies appraised for interventions to promote 10 adherence are outlined below 11 (a) The content and method of delivery of the components of the intervention 12 are not well described and differ in different studies. 13 (b)The lack of distinction between content of intervention and how it was 14 delivered makes it impossible to understand the overall poor and contradictory 15 outcomes 16 (c)The majority of the studies does not assess a single intervention but 17 include multi-component interventions. In a trial the two interventions being 18 compared may each have a different set of components. This means it is not 19 possible to compare one trial with another and even when an intervention 20 works it is not possible to know which component or combination of 21 components is effective. 22 (d) Studies do not report whether the planned interventions took place as 23 intended in the study protocol. 24 (e) No standard method of assessing adherence is used. 25 (f) There is infrequent justification of the relevance of certain interventions in 26 improving adherence e.g. there is no theoretical framework informing the 27 studies and this precludes development of understanding of phenomena of 28 adherence/non-adherence. Medicines concordance: full guideline DRAFT (July 2008) page 222 of 373 DRAFT FOR CONSULTATION 1 (h) Studies had frequently small sample sizes and inadequate description of 2 settings and existing rates of adherence; generalising of these interventions to 3 routine clinical practice is therefore not possible. 4 (i) Many of the interventions are extremely complex and labour-intensive and 5 are carried out by paid research staff and thus it is unknown whether they 6 could be carried out in a non-research setting. 7 (i ) Even the most apparently effective interventions did not lead to substantial 8 improvements in adherence and treatment outcomes. 9 (k) Interventions are not targeted to causes of non-adherence – it is likely 10 therefore that some interventions may be more effective than evidence 11 suggests if directed to actual cause of non–adherence in individual patients. 12 (l) Comment is not made on patient involvement in decision to prescribe 13 medicines 14 (m) Few studies mention whether raters of outcome or adherence were blind 15 to whether subjects were in the intervention or control group 16 (n) Despite the comprehensive and detailed searching, some trials that met 17 our criteria may have been missed. The literature on patient adherence is not 18 well indexed as it sprawls across the traditional disease areas and as a result 19 of all these, there is little information available to fully understand why one 20 intervention works and other very similar ones did not. Our concern about the 21 evidence in this area is mirrored in other key reviews and trials (Haynes, R. 22 B., Ackloo, E., Sahota, N. et al , 2008). 23 The results of the reviews need to be interpreted with caution, as some of the 24 elements that have worked within some of the trials are present in other 25 studies that have not yielded significant improvements. We were interested in 26 simple interventions that might be targeted to individuals but the majority of 27 the studies are complex interventions with an extremely wide range of 28 components. Aggregation of studies was extremely difficult and has required 29 subjective judgment. 30 Medicines concordance: full guideline DRAFT (July 2008) page 223 of 373 DRAFT FOR CONSULTATION 1 2 3 8.5 Key Clinical Question: Does change in dosing regime affect adherence? Related references Evidence statements Evidence into recommendations (summary of evidence) Shi (2007) (Shi, L., 1.Systematic reviews and For general discussion of limitations Hudges, M., Yurgin, N. et RCTs show that evidence see section 7.3 al , 2007); Schroeder simplification of dosing (2004) (Urien, A. M., frequency can increase Guillen, V. F., Beltran, D. adherence to prescribed O. et al , 2004); Iskedjian medication. (2002) (Iskedjian, M., Einarson, T. R., MacKeigan, L. D. et al , 2002); Claxton (2001) (Claxton, A. J., Cramer, J., and Pierce, C., 2001) The GDG were interested in whether there was evidence to indicate that changes in dosing regime would imp adherence. The findings of the evide review were that reducing the comple of a regime can increase adherence the quality of evidence was low. The evidence from the qualitative intervie indicated that the difficulty for patient integrating the regime into their lives Baird (1984) (Baird, M. G., rather than dose complexity per se a Bentley-Taylor, M. M., the GDG recommendation is that Carruthers, S. G. et al , changes to dosing regime need to be 1984); Brown (1997) tailored to needs of individual patient (Brown, B. G., Bardsley, J., Poulin, D. et al , 1997); Girvin (1999) (Girvin, B., McDermott, B. J., and Johnston, G. D., 1999); Portsmouth 2005 [1216}; Molina (2007) (Molina, Jean Michel, Podsadecki, Thomas J., Johnson, Medicines concordance: full guideline DRAFT (July 2008) page 224 of 373 One area of interest for the GDG was use of drugs by injection particularly antipsychotic drugs and contraceptive This option could be classified as changes to dose regime or drug formulation. The GDG were clear tha using drugs by injection in this way m be an appropriate choice where patie DRAFT FOR CONSULTATION Margaret A. et al , 2007) have non-intentional adherence i.e. t forget to take their medicines. As su Schroeder (2004) (Shi, L., 2. Two Systematic reviews Hudges, M., Yurgin, N. et and four RCTs show that al , 2007); Iskedjian (2002) reducing twice-daily to (Iskedjian, M., Einarson, T. once-daily dosing may R., MacKeigan, L. D. et al , increase adherence to 2002); Portsmouth 2005; prescribed medication. Molina (2007) (Molina, this choice should be offered to patie The GDG were aware of evidence fro qualitative synthesis that patients with mental health problems can feel coer to take medication and wished it mad clear that the aim of our recommendations is to support inform Jean Michel, Podsadecki, adherence. Thomas J., Johnson, Margaret A. et al , 2007); Baird (1984) (Baird, M. G., Bentley-Taylor, M. M., Carruthers, S. G. et al , 1984); Girvin (1999) Claxton (2001) (Claxton, 3. Evidence from one low AL, Cramer, J, and Pierce, quality systematic review C, 2001); Rudd (2004) and two RCTs showed (Rudd, P., Miller, N. H., that once daily dosing Kaufman, J. et al , 2004); compared to twice daily Parienti (2007) (Parienti, did not increase Jean Jacques, Massari, adherence to prescribed Véronique, Reliquet, medication. Véronique et al , 2007). 1 Methods of the evidence review 2 8.5.1 3 This paper includes a narrative summary of the included evidence, structured 4 according to the category of the intervention, following the agreed reviewing 5 protocol: Medicines concordance: full guideline DRAFT (July 2008) page 225 of 373 DRAFT FOR CONSULTATION 1 Types of studies – We initially included only randomized controlled trials 2 (RCTs) of interventions to increase adherence. The excluded studies list from 3 the Cochrane review was checked as we have included those studies with 4 less than 80% follow-up of participants. As with the Cochrane review we found 5 only a small number of studies that fulfilled our criteria. For this evidence 6 review we excluded any randomised controlled trials that evaluated changes 7 in dosing regimes but did not assess the same medication in all comparative 8 groups. The GDG requested further search to pick up any systematic review 9 published after the Cochrane Review search cut-off. However, the included 10 systematic reviews did not follow this criterion. 11 Types of participants - people prescribed medication for a medical condition. 12 Duration of studies - six months follow up from the time of patient entry for 13 long-term regimens for the RCTs. No time limit specified for short-term 14 conditions. 15 Types of interventions - any interventions intended to change adherence to 16 prescribed medication. As the Cochrane review is presented by condition, we 17 have used the evidence extracted in that review and reconfigured it by 18 intervention. 19 Types of outcome measures – inclusion criteria (as defined in the Cochrane 20 review) were expanded by including studies that used adherence as the only 21 outcome variable as opposed to adherence and treatment outcome variables. 22 The excluded studies list of the Cochrane review was cross-referenced to 23 ensure that no potentially relevant study was missed out. 24 8.5.2 25 8.5.2.1 26 Systematic Reviews 27 One recent review of systematic reviews and empirical studies (Shi 2007) 28 (Shi, L., Hudges, M., Yurgin, N. et al , 2007) looked at the impact of dose 29 frequency on compliance and health outcomes, particularly for injectables. Evidence review Does change in dosing regime affect adherence? Medicines concordance: full guideline DRAFT (July 2008) page 226 of 373 DRAFT FOR CONSULTATION 1 Inclusion criteria were that the studies should compare different dose 2 frequencies, including injectable medications and be published in a peer 3 reviewed journal. Exclusion criteria were that the article should not focus 4 solely on dosage forms, dose administration or dose timing. 5 Full text reviews were conducted on a total of 64 empirical studies and 25 6 literature/systematic reviews. No details were given on overall methodological 7 quality of the studies. 8 Results were presented through five main areas: cardiovascular diseases, 9 diabetes, nephrology/urology, neurology/psychiatry and rheumatoid/muscle. 10 Of the 21 studies that measured compliance, 17 reported a positive impact 11 (no details of significance given) of reducing dose frequency on compliance, 12 whilst inconclusive results were seen in four. Details of the dose frequency 13 reductions contained in the studies were not provided by the review. 14 Articles not measuring compliance as the main outcome looked at efficacy 15 and other outcomes of extended-release medications in comparison to the 16 immediate-release forms. The studies also supported the general benefits of 17 reducing dosing frequency on improved quality of life or patients’ satisfaction 18 (6 studies), greater control over side effects (5 studies) and improved 19 economic outcomes using extended-release formulation (2 studies). 20 Schroeder (2004) (Urien, A. M., Guillen, V. F., Beltran, D. O. et al , 2004) 21 used the Cochrane methodology to review dosing regimes and adherence in 22 hypertensive patients. The methodological quality of the primary included 23 studies was assessed to be generally low. Many RCTs showed marked 24 heterogeneity in terms of participants, interventions and outcomes. A pooled 25 analysis was considered inappropriate as results on adherence were reported 26 in many different ways. Simplifying dosing regimens improved adherence in 7 27 of 9 studies with relative improvement in adherence increasing by 8% to 28 19.6%. All of the studies that used objective outcome measurement 29 (Medication Event Monitoring System) showed an improvement in adherence 30 through the use of once daily instead of twice daily dosing regimens, although 31 4 of these compared 2 different drugs. Only 1 study showed an increase in Medicines concordance: full guideline DRAFT (July 2008) page 227 of 373 DRAFT FOR CONSULTATION 1 adherence (905 vs. 82%; p<.01) together with a reduction in systolic blood 2 pressure of 6 mm Hg (p<.01). 3 A weak meta-analysis conducted by Iskedjian (2002) (Iskedjian, M., 4 Einarson, T. R., MacKeigan, L. D. et al , 2002), combined comparative studies 5 of different research designs including prospective trials (RCTs and cohort 6 studies), retrospective chart reviews and database analyses. Adherence was 7 defined differently in various studies and different instruments were used to 8 measure patient adherence. In this meta-analysis, all variables that could 9 affect adherence other than daily dose frequency, were assumed to be equal 10 between comparators. 11 Eight studies involving a total of 11,465 observations were included (1830 for 12 daily [QD] dosing, 4405 for twice a day dosing [BID] and 4147 for dosing >2 13 times daily [>BID] and 9655 for multiple daily dose [MDD]). The primary 14 objective was to assess adherence. The average adherence rate for QD 15 dosing (91.4%, SD=2.2%) was significantly higher than for MDD (83.2%, 16 SD=3.5%; p<0.001). The difference between adherence rates for QD dosing 17 (92.7%) and BID dosing (87.1%) was also statistically significant (p=0.026), 18 although the difference in this analysis was smaller than in the QD-versus- 19 MDD analysis (5.7% vs. 8.2%). The difference in adherence rates between 20 BID dosing (90.8%, SD 4.7%) and >BID dosing (86.3%, SD=6.7%) was not 21 significant (p=0.069). However, a subgroup analysis using a stricter definition 22 of adherence (≥90% intake) did reveal a statistically significant difference 23 between BID and >BID dosing (respective adherence rates of 76.1% and 24 67.0%, p<0.001). 25 Another systematic review published by Claxton (2001) (Claxton, A. J., 26 Cramer, J., and Pierce, C., 2001) also found that simpler, less frequent dosing 27 regimes resulted in better compliance. This systematic review appeared to 28 include several study designs. This review showed strong methodological 29 limitations particularly in terms of data analysis. The study did not give full 30 details of inclusion/exclusion criteria and thus possibly including studies that 31 compared different dosing regimes in different medicines. The results should 32 therefore be viewed with caution. Medicines concordance: full guideline DRAFT (July 2008) page 228 of 373 DRAFT FOR CONSULTATION 1 Seventy-six studies were included in the review. By combining all data it was 2 found that increasing the number of daily doses was significantly related to a 3 decline in compliance (p<0.001 among dose schedules). Comparisons 4 between dose regimens showed that compliance was significantly higher with 5 once daily regimens vs. 3 times daily (p=0.008) or 4 times daily regimens 6 (p<0.001). Compliance with twice daily dosing was significantly higher than 4 7 times daily dosing (p=0.001). There were no significant differences in 8 compliance between once daily and twice daily regimens or between twice 9 daily and 3 times daily. 10 Randomised controlled trials 11 Three RCTs Baird (1984) (Baird, M. G., Bentley-Taylor, M. M., Carruthers, 12 S. G. et al , 1984); Brown (1997) (Brown, B. G., Bardsley, J., Poulin, D. et al , 13 1997) and Girvin (1999) (Girvin, B., McDermott, B. J., and Johnston, G. D., 14 1999) from the Cochrane Review (2005) , assessed the effect of the 15 simplification of a dosing frequency. Baird (1984) (Baird, M. G., Bentley- 16 Taylor, M. M., Carruthers, S. G. et al , 1984) compared twice a day 100mg 17 Betaloc tablets to once daily 200mg Betaloc Durules, in a sample comprising 18 389 participants. Mean age of the participants was 52.7 years for the twice 19 daily group, compared to 54.5 years. Over the total study period, compliance 20 exceeded 80% in 96.4% of patients in the Durules group and 90.0% of patient 21 in the Betaloc tablets group (p=0.0591). When to 90% levels of compliance 22 were compared, overall compliance exceeded this level in 92.8% of patients 23 on Durules and in only 81.5% of patients on tablets (p=0.009). A significant 24 effect in increasing adherence was reported. Brown (1997) (Brown, B. G., 25 Bardsley, J., Poulin, D. et al , 1997) tested controlled-release niacin twice 26 daily to regular niacin, four times daily, in the treatment of hyperlipidemia and 27 coronary artery disease, in 29 male participants aged ≤65 years. Compliance 28 was 95% with the controlled- release niacin versus 85% with regular niacin 29 (p< 0.001). Girvin (1999) (Girvin, B., McDermott, B. J., and Johnston, G. D., Study information indicates that duration is less than 6 months, however this is not stated in Cochrane Review. Study with less than 6 months duration that was included in the Cochrane Review as results for blood pressure outcomes were negative. Medicines concordance: full guideline DRAFT (July 2008) page 229 of 373 DRAFT FOR CONSULTATION 1 1999) tested enalapril 20 mg once daily versus enalapril 10 mg twice daily in 2 the treatment of high blood pressure. Sample size comprised of 27 patients. 3 Mean age of participants was 62 years. Overall medication adherence was 4 improved with once-a-day dosing. The difference in percentage of doses 5 taken by pill count between the two periods was significantly in favour of the 6 once daily regimen (p<0.01), as was the percentage of doses taken as 7 measured by a pill container that recorded lid openings (MEMS) (p< 0.001). 8 One RCT Portsmouth (2005) (Portsmouth, S. D., Osorio, J., McCormick, K. 9 et al , 2005) included in the 2008 revision of the Cochrane Review assessed 10 whether virologically controlled HIV-1-infected individuals switched from a 11 twice-daily antiretroviral regimen to a once daily regimen demonstrate 12 improved adherence and quality of life while maintaining virological control. 13 Forty-three patients were included in this study, with 22 in the once daily 14 (intervention) group, and 21 in the twice daily (control) group. 15 The once daily group (intervention): the prolonged release capsule group 16 (PRC) were assigned to take d4T PRC/3TC/EFV all once-daily (24 h apart); 17 Twice daily (control group): participants in the control group were assigned to 18 continue either d4T IR/3TC/EFV or Combivirs/EFV as per their screening 19 regimen. Note: participant weighing less than 60 kg were prescribed either 30 20 mg of d4T IR or 75 mg of d4T PRC. 21 After randomization, patients allocated to the PRC (intervention) maintained 22 this high adherence, while those allocated to IR (control) showed a 23 significantly reduced adherence in ‘taking compliance’ (P=0.0237) 24 (percentage of prescribed number of doses taken), ‘correct dosing 25 compliance’ (P=0.0104) (percentage of days with correct number of doses 26 taken) and ‘timing compliance’ (P=0.028) (percentage of doses taken within 3 27 hours of the prescribed dosing intervals) at both weeks 12 and 24. 28 One RCT, Rudd (2004) (Rudd, P., Miller, N. H., Kaufman, J. et al , 2004), 29 from the Cochrane Updated Review that was included in the evidence review 30 on the effects of self-monitoring on adherence reported some results in regard 31 to once-daily regimens compared to twice-daily regimens. This RCT Medicines concordance: full guideline DRAFT (July 2008) page 230 of 373 DRAFT FOR CONSULTATION 1 assessed a system for patients to monitor their own blood pressure. Seventy 2 six patients received routine care while the intervention group (n=74) received 3 an automated blood pressure device for use at home with management by a 4 nurse care manager. The patients recorded their own blood pressure then the 5 device printed these which were mailed to the nurse care manager in order to 6 guide drug therapy. The adherence measures by a drug event monitor was 7 found to be significant (80% for the intervention group and 69% for the control 8 group, p=0.03). One of the outcomes found that once-daily regimens had 9 higher adherence 82% (SD=28%) than twice-daily 69% (SD=34%) or more 10 frequently 49% (SD=41%). None of these differences were statistically 11 significant. 12 Update searches 13 From the conducted update searches, we retrieved two RCTs that were 14 considered important as they would contribute to modifying the 15 recommendations drafted for the topic of the impact of changes of dosing 16 regimes on adherence to prescribed medication. These were Molina (2007) 17 and Parienti (2007). 18 The safety, efficacy and adherence to lopinavir/ritonavir (LPV/r) dosed QD or 19 BID in antiretroviral-naive, HIV-1-infected subjects was evaluated in an RCT 20 Molina (2007) (Molina, Jean Michel, Podsadecki, Thomas J., Johnson, 21 Margaret A. et al , 2007). A randomized, open-label, multicenter comparative 22 study was conducted through 96 weeks of treatment. A total of 190 23 antiretroviral-naive subjects with plasma HIV-1 RNA above 1000 copies/ml 24 and any CD4(+) T cell count were enrolled. Subjects were randomized (3:2) to 25 LPV/r 800/200 mg QD (n = 115) or 400/100 mg BID (n = 75). Subjects 26 received TDF 300 mg and FTC 200 mg QD. Adherence to LPV/r through 96 27 weeks was measured using MEMS((R)) monitors. Median baseline VL and 28 CD4(+) T cell count were 4.8 log(10) copies/ml and 216 cells/mm(3), 29 respectively. Prior to week 96, 37% (QD) and 39% (BID) of subjects 30 discontinued, primarily due either to adverse events (17% QD, 9% BID) or to 31 loss to follow-up or nonadherence (12% QD, 17% BID). The proportion of 32 subjects with VL <50 copies/ml (57% QD, 53% BID; p = 0.582 (ITT NC = F)), Medicines concordance: full guideline DRAFT (July 2008) page 231 of 373 DRAFT FOR CONSULTATION 1 change in CD4 count (244 cells/mm(3) QD, 264 cells/mm(3) BID; p = 0.513), 2 and evolution of resistance did not differ between groups through 96 weeks. 3 Diarrhoea (17% QD, 5% BID, p = 0.014) was the most common moderate or 4 severe, study drug-related adverse event. Adherence to LPV/r was higher for 5 the QD group than the BID group and declined over time in both groups. 6 Parienti (2007) (Parienti, Jean Jacques, Massari, Véronique, Reliquet, 7 Véronique et al , 2007) aimed to determine the effect of once-daily dosing on 8 adherence to nevirapine. This RCT was comprised of three-phase (3-month 9 observational, 4-month randomized, 5- month interventional) open-label, 10 clinical trial at four French academic medical centres during 2005-2006 11 among 62 chronically HIV-1- infected subjects with long-lasting viral 12 suppression under a twice-a- day nevirapine-based antiretroviral combination. 13 Participants were randomly assigned to switch to nevirapine 400 mg once- 14 daily (n = 31) or continue nevirapine 200 mg twice-a-day (n = 31). After the 15 randomized phase, participants had an opportunity to choose their 16 antiretroviral dosage. 17 Fifty-two patients qualified for electronic data analysis. During the randomized 18 phase, the mean adherence rate was non-significantly superior by 0.5% in 19 once-daily versus twice-a-day dosing (P = 0.68), adjusting for previous twice- 20 a-day adherence rate (P < 0.0001). Once-daily group increased days without 21 dose (odds ratio (OR) 1.7; 95% confidence interval (CI) 1.0, 2.8; P = 0.04), 22 adjusting for previous drug interruptions (P < 0.0001). In the longitudinal 23 analysis, once-daily dosing was significantly associated with at least two 24 consecutive days without dose (OR 4.4; 95% CI 1.9, 10.3; P < 0.001). The 25 authors concluded that changing from twice to once-daily nevirapine did not 26 improve adherence 27 28 Medicines concordance: full guideline DRAFT (July 2008) page 232 of 373 DRAFT FOR CONSULTATION 1 2 8.6 Key Clinical Question : Effect of prescription charges/costs on adherence to prescribed medication 3 Related Evidence statements references (summary of evidence) Hirth (2008) 1. Some UK patients may have Most of the evidence for cost as a barrier (Hirth, Richard difficulty affording medicines. to adherence comes from US. Only a few A., Greer, Scott L., Albert, Justin M. et al , 2008); Atella (2005) (Atella, Vincenzo, Schafheutle, Ellen, Noyce, Peter et al , 2005) Evidence into recommendations studies have been conducted in the UK. 2. The most common strategies for patients with problems affording medicines is to delay the dispensing of medicines, to These indicate that for some patients this is a concern. Cost concerns may also indicate doubts by the patient about the value of the prescription. not visit the GP and to lower the dose below that prescribed to When cost is a concern for patients a extend the duration of the variety of options are available each of prescription. which have advantages and disadvantages. Prescription length may be increased giving the patient longer prescription for same cost but this may reduce opportunity for review. Quite short dispensing time frames may be important for example when patients are suicidal or need careful monitoring of medication and its effects. Instalment dispensing is possible for certain drug items but in general it seems unreasonable to ask a patient to pay for each dispensing point. However the GDG was also mindful of the fact, that it might also be unreasonable for the pharmacist to make serial dispensings for a single dispensing fee. Costs and value of prescriptions should be Medicines concordance: full guideline DRAFT (July 2008) page 233 of 373 DRAFT FOR CONSULTATION considered not just at the point of prescribing but at all stages of the process 1 2 *e.g. relative value place on different outcomes, trade-offs between benefits and harms, uncertainties in Evidence, likely impact if implements, feasibility, patient perspective 3 Medicines concordance: full guideline DRAFT (July 2008) page 234 of 373 DRAFT FOR CONSULTATION Methods of the evidence review 1 8.6.1 2 This paper includes a narrative summary of the included evidence, structured 3 according to the category of the intervention, following the agreed reviewing 4 protocol: 5 Types of studies – no restrictions on study design. However, due to the 6 nature of the question, the requirement was that the studies needed to be 7 conducted in the UK. 8 Types of participants- people prescribed medication for a medical condition. 9 Duration of studies - no time limit specified. 10 Types of interventions - any interventions intended to assess the correlation 11 between prescription charges/costs and the impact on adherence to 12 prescribed medication. 13 8.6.2 14 Types of outcome measures – adherence to prescribed medication, cost 15 reducing strategies. 16 We retrieved one observational study (Hirth 2008) (Hirth, Richard A., Greer, 17 Scott L., Albert, Justin M. et al , 2008) that examined out of pocket medication 18 spending and cost-related medication nonadherence among dialysis patients 19 in twelve countries including the UK. 20 Data were gathered from 2002 to 2004 as part of the dialysis outcomes and 21 practice patterns study (DOPPS), an observational study of haemodialysis 22 practices and outcomes in twelve countries- Australia, New Zealand, Belgium, 23 Canada, France, Germany, Italy, Spain, Sweden, United Kingdom, Japan, 24 and the United States. A random sample of patients was selected, totalling 25 N=7.766. Of the selected 83 per cent who agreed to enrol and have their 26 medical records abstracted, 85 per cent of these enrolled patients also 27 completed the patient questionnaire. A total of 70 per cent of patients provided 28 both medical and questionnaire data. Local currencies were converted to US 29 Dollars. Evidence review Medicines concordance: full guideline DRAFT (July 2008) page 235 of 373 DRAFT FOR CONSULTATION 1 Questionnaires and medical record abstraction techniques were standardised 2 across countries and languages. Patient questionnaires were administered 3 soon after recruitment. They were asked about the total out-of-pocket 4 spending for prescription and over the counter (OTC) medications in the 5 previous month. They were also asked “Do you sometimes decide not to 6 purchase medications because of cost?” and to report their out-of-pocket 7 spending for haemodialysis treatments. 8 Whilst the United States reported 86 per cent of out-of-pocket spending for 9 medications, only patients in Australia/New Zealand, Belgium, and Sweden 10 were significantly more likely to face out-of-pocket spending, while those in 11 France, Japan, Spain and the UK were significantly less likely to do so. 12 Mean monthly spending for prescription and OTC medications ranged from $8 13 in the UK to $114 to the United States. Among patients with medication 14 spending, only 10 per cent faced monthly costs greater than $30 in the United 15 Kingdom, whereas 10 per cent incurred costs greater than $310 in the United 16 States. 17 Observed cost-related nonadherence, indicated by the proportion of patients 18 who reported that they sometimes did not purchase medications because of 19 cost, was significantly less than expected in France, Japan, Spain, Sweden 20 and the UK. 21 Nonadherence was associated with the percentage of patients reporting any 22 out-of-pocket spending and the average out-of-pocket cost. Although the US 23 had high out-of-pocket spending burdens, their nonadherence was still clearly 24 higher than would be expected on the basis of the percentage facing any 25 costs or the mean cost burden. On the other hand, Sweden and Belgium had 26 lower levels of nonadherence than would be expected given either measure of 27 out-of-pocket spending burden. The lowest nonadherence rates existing in 28 France, Japan, Spain and the UK were correlated with low out-of-pocket 29 spending. 30 Atella 2005 (Atella, Vincenzo, Schafheutle, Ellen, Noyce, Peter et al , 2005) 31 aimed to explore how and to what extent costs incurred by patients influence Medicines concordance: full guideline DRAFT (July 2008) page 236 of 373 DRAFT FOR CONSULTATION 1 their decision-making behaviour in accessing medicines, both in the UK and in 2 Italy. 3 Based on findings from focus groups, a questionnaire was designed to assess 4 medication cost issues. As such, several hypotheses were tested regarding 5 patients’ decision-making behaviour and how it was influenced by health and 6 sociodemographic status and the novel concept of a self-rated affordability 7 measure. Patients were eligible if they had either dyspepsia or mild 8 hypertension. They were sampled as successive patients who visited 51 9 physicians in Italy and 21 community pharmacists in the UK. Samples were 10 drawn from the areas of Manchester and Rome. Of the 550 dyspepsia and 11 600 hypertension questionnaires distributed, 122 and 153 were returned- a 12 response rate of 22.2% and 25.5%, respectively. In the UK, 296 dyspepsia 13 and 277 hypertension questionnaires were distributed, targeting dyspepsia 14 patients who bought OTC medicines, and dyspepsia and hypertension 15 patients who had to pay prescription charges; 110 dyspepsia and 134 16 hypertension questionnaires were returned, giving a response rates of 37.5% 17 and 48.4%. In both countries the majority of the respondents were not 18 exempt. 19 The self-rated affordability measure showed that 70.3 per cent of the UK 20 sample and 66.5 percent of the Italian sample had to think about the cost of 21 medicines at least sometimes. Also, 24.3 per cent and 16.3 per cent, 22 respectively said they always have to think about how much money they have 23 available to spend when they obtain medicines. According to the results, the 24 patient initiated strategy most commonly used by UK respondents with 25 affordability problems is (1) to delay the dispensing of drugs until they get 26 paid, (2) not visiting the GP to avoid incurring the cost of prescribed 27 medication and (3) reducing the dose below that prescribed to extend the 28 course of medication. 29 Affordability issues were also strong when examining the use of self- 30 medication strategies. The UK respondents were particularly cost conscious 31 when considering the price of an OTC product before buying it, or they would 32 ask for something cheaper if they could not afford a particular OTC product. Medicines concordance: full guideline DRAFT (July 2008) page 237 of 373 DRAFT FOR CONSULTATION 1 The authors point out that affordability seemed to play a more important role 2 in the UK sample than in the Italian, however they do point out that Italian 3 patients with dyspepsia were sampled only through GPs and may be those 4 more severely affected and/or less likely to be disposed towards self 5 medication. Also, OTC products are much more expensive in relation to the 6 prescription charge that they are in the UK where the prescription charge is 7 high. 8 9 10 11 12 13 14 Medicines concordance: full guideline DRAFT (July 2008) page 238 of 373 DRAFT FOR CONSULTATION 1 2 8.7 Key Clinical Question: Does drug formulation and/or packaging affect adherence? 3 Related references Evidence statements (summary Evidence into of evidence) recommendations For general discussion of Drug packaging limitations of evidence see Orton (2005) (Orton, L. and 1. Two systematic reviews do not Barnish, G., 2005) and convincingly support the use of Connor (2004) (Connor, J., unit-dose packaging to improve The GDG considered the Rafter, N., and Rodgers, adherence to prescribed evidence review did not A., 2004) medication provide convincing section 7.3 evidence that drug packaging per se increases Lee (2006)(Lee, J. K., 2. There is conflicting evidence on adherence. Health care Grace, K. A., and Taylor, A. the value of special packaging in a professionals should J., 2006); Becker multicomponent intervention in explore with patients (1986)(Becker, L. A., increasing adherence to whether the way in which a Glanz, K., and Sobel, E., prescribed medication. drug is packaged causes 1986); Henry (1999)(Henry, difficulty and respond to A. and Batey, R. G., 1999) individual problems. Schneider (2008) 3. One RCT found that the use of (Schneider, P. J., Murphy, blister packaging (Pill Calendar) J. E., and Pedersen, C. A., compared to medication in a bottle 2008) significantly increased adherence. Lee (2006)(Lee, J. K., 4. One RCT showed that special Grace, K. A., and Taylor, A. packaging in a multicomponent J., 2006) intervention given to an elderly population (≥65 years) may Medicines concordance: full guideline DRAFT (July 2008) page 239 of 373 DRAFT FOR CONSULTATION increase adherence to prescribed medication. Drug formulation Bangalore (2007) 5. One highly biased systematic For general discussion of (Bangalore, Sripal, review suggests that fixed dose limitations of evidence see Kamalakkannan, Gayathri, combination compared to free drug section 7.3 Parkar, Sanobar et al , component regimen may increase 2007) adherence to prescribed medication. The GDG considered the evidence review did not provide convincing Brown 1997(Brown, B. G., 6. One RCT showed that evidence that changes to Bardsley, J., Poulin, D. et al controlled release medication drug formulation will , 1997) along with simplified dosing improve adherence. compared to regular medication Changes to drug may increase adherence to formulation should be prescribed medication. considered with changes to dosing as a response to individual patient problems only. 1 2 *e.g. relative value place on different outcomes, trade-offs between benefits and harms, uncertainties in Evidence, likely impact if implements, feasibility, patient perspective Medicines concordance: full guideline DRAFT (July 2008) page 240 of 373 DRAFT FOR CONSULTATION Methods of the evidence review 1 8.7.1 2 This paper includes a narrative summary of the included evidence, structured 3 according to the category of the intervention, following the agreed reviewing 4 protocol: 5 Types of studies - We initially included only randomized controlled trials 6 (RCTs) of interventions to increase adherence. The excluded studies list from 7 the Cochrane review was cross-referenced as we included studies with less 8 than 80% follow-up of participants. After the GDG voiced concerns over the 9 possibility of missing out important studies by only having included a small 10 amount of studies, the search was redone to pick up any systematic review 11 published after the Cochrane Review search cut-off. 12 Types of participants - people prescribed medication for a medical condition. 13 Duration of studies - six months follow up from the time of patient entry for 14 long-term regimens for the RCTs. No time limit specified for short-term 15 conditions. 16 Types of interventions - any interventions intended to change adherence to 17 prescribed medication. As the Cochrane review is presented by condition, we 18 have used the evidence extracted in that review and reconfigured it by 19 intervention. 20 Types of outcome measures - inclusion criteria (as defined in the Cochrane 21 review) were expanded by including studies that used adherence as the only 22 outcome variable as opposed to adherence and treatment outcome variables. 23 The excluded studies list of the Cochrane review was cross-referenced to 24 ensure that no potentially relevant study was missed out. 25 8.7.2 26 8.7.2.1 27 Systematic Reviews Evidence review Effect of drug packaging on adherence Medicines concordance: full guideline DRAFT (July 2008) page 241 of 373 DRAFT FOR CONSULTATION 1 One Cochrane Review Orton (2005) (Orton, L. and Barnish, G., 2005) that 2 aimed to assess the effects of unit-dose packaged treatment and treatment 3 adherence in people with uncomplicated malaria was retrieved. Any type of 4 programme that included unit-dose packaging of antimalarial drugs packed in 5 units of a single dose was incorporated in the review. Treatment adherence 6 was a secondary outcome, however all four included studies measured it. 7 Interventions and control arm groups had to received the same antimalarial 8 drug and any other intervention. The interventions that were assessed in this 9 systematic review ranged from labelled and boxed blister packs of chloroquine 10 and primaquine tablets and capsules, simple labelled and sectioned polythene 11 bags of chloroquine tablets, tablets or capsules in paper envelopes or loose 12 and chloroquine syrup in bottles. 13 Three quasi RCTs and one cluster RCT met the inclusion criteria, and overall 14 trials were of poor methodological quality. 15 A meta-analysis of two trials (with 596 participants) showed that participant 16 reported treatment adherence was higher with blister-packed tablets 17 compared with tablets in paper envelopes (RR 1.18, 1.12 to 1.25). Two trials 18 using tablets in sectioned polythene bags as the intervention also reported an 19 increase in participant reported treatment adherence: in one study (cluster 20 RCT) it was compared with the tablets in paper envelopes whilst the other trial 21 compared it with syrup in bottles (RR 2.15, 1.76 to 2.61; 299 participants). 22 It appears that unit-dose packaging drugs (in combination with prescriber 23 training and patient information) was associated with higher participant 24 reported treatment adherence, however this conclusion is drawn from trials 25 with methodological limitations. 26 The Bulletin of the World Health Organization published a systematic review 27 in Connor (2004) (Connor, J., Rafter, N., and Rodgers, A., 2004) which 28 evaluated the evidence on fixed dose combination pills and unit-of use 29 packaging with relation to adherence. Fifteen randomized or quasi 30 randomized trials met inclusion criteria: fixed dose on cure combination pills 31 were investigated in three of these while unit-of-use packaging was studied in Medicines concordance: full guideline DRAFT (July 2008) page 242 of 373 DRAFT FOR CONSULTATION 1 12 trials. Eligible studies varied substantially in their settings, participant 2 selection, medical conditions, interventions, adherence measures and clinical 3 outcome measures, as well as in study quality. For these reasons, a meta- 4 analysis was not undertaken. The results of the trials suggested that there 5 were trends towards improved adherence which reached statistical 6 significance in seven out of thirteen trials reporting medication adherence. 7 Measures of adherence were however heterogeneous and interpretation was 8 further limited by methodological issues, particularly small sample size, short 9 duration and loss to follow up. The authors therefore concluded that 10 uncertainty remains about the size of the benefits of drug formulation and 11 packaging. 12 Randomised Controlled Trials 13 Lee (2006) (Lee, J. K., Grace, K. A., and Taylor, A. J., 2006) compared a 14 comprehensive pharmacy care program which was delivered to one group for 15 3-8 months and a second group for 3-14 months in 200 patients aged 65 16 years or over, taking 4 or more chronic medications daily with positive results. 17 The first group returned to usual care after 8 months. The care program 18 consisted of 3 elements, including individualised medication education, 19 medications dispensed using an adherence aid (blister packs) and regular 20 follow-up with clinical pharmacists every 2 months. This study was conducted 21 in the USA. 22 Mean baseline adherence overall was 61.2% (SD 13.5%) with an overall level 23 of adherence of 96.9% at 8 months of intervention. At 14 months, medication 24 adherence was 95.5% (SD 7.7%) in the continued intervention group and 25 69.1% (SD 16.4%) in the control group (p < 0.001). Proportions of people who 26 had at least 80% adherence rates were 97.4% in the intervention group and 27 21.7% in the control group (p < 0.001). 28 Henry (1999) (Henry, A. and Batey, R. G., 1999) from the Cochrane Review 29 delivered an intervention where verbal advice on medication use and possible 30 side-effects were employed along with information sheets on the treatments 31 and medication with dose-dispensing unit. Control group were only given Medicines concordance: full guideline DRAFT (July 2008) page 243 of 373 DRAFT FOR CONSULTATION 1 treatment, along with verbal advice and information sheets. A total of 119 2 patients were studies. Mean age of patients was 58 years for the control 3 group and 57 for the intervention group. Compliance in intervention group 4 patients was also encouraged by a phone call 2 days after the start of therapy. 5 This study was conducted In Australia. No significant effects on adherence 6 were reported. 7 Becker (1986) (Becker, L. A., Glanz, K., and Sobel, E., 1986) from the 8 Cochrane review delivered an intervention whereby patients aged 20 to 80 9 years were assigned to the experimental group received all their medications 10 in the special packaging format (all pills taken together were packaged in a 11 single plastic blister sealed with a foil backing on which was printed the day of 12 the week and the time of day at which each medication was to be taken). One 13 hundred and eighty patients were included in the study. Patients in the control 14 group received all of their antihypertensive medications in the conventional pill 15 vials (separate vials for each pill that were labelled with the drug name, the 16 dosage, the medication instructions, and the physician’s name). All 17 medications for both groups were provided free of charge to ensure that all 18 patients would receive their medications. This study was conducted in the 19 USA. No significant effects on adherence were reported. 20 Schneider (2008) (Schneider, P. J., Murphy, J. E., and Pedersen, C. A., 21 2008) conducted a randomised controlled trial to assess the impact of one 22 medication packaging type on adherence and treatment outcomes of older 23 patients. The study was conducted at 3 sites in Tucson and Columbus in the 24 USA. 85 participants aged 65 years or older, prescribed lisinopril 25 (antihypertensive medication) were randomised to receive daily-dose blister 26 packaged medication (pill calendar) as the intervention compared to traditional 27 bottles of loose tablets as the control group. Patients returned for refills every 28 28 days during a 12 month period where the pharmacist would record the time 29 between prescription refills for the medication and any study-related problems. 30 At 6 and 12 months after enrolling the patients visited the physician to find out 31 blood pressure management; the occurrence of morbidity in the past 6 months 32 e.g. angina, myocardial infarction and stroke; and any medical services they Medicines concordance: full guideline DRAFT (July 2008) page 244 of 373 DRAFT FOR CONSULTATION 1 had required in the past 6 months e.g. hospitalisations or emergency 2 department visits. Medical charts were reviewed by two pharmacists to gather 3 this information. The percentage of times prescriptions were refilled on time 4 (within 5 days before or after due date) were significantly higher 80.4% 5 (SD=21.2) for the intervention group than the control group, 66.1% (SD=28), 6 p=0.012. The Medication possession rate (the sum of the day’s supply for all 7 prescriptions received during the study divided by the number of days 8 between the first and last prescription dispensed) was also significantly higher 9 for the intervention group, 0.93 (SD=11.4) and 0.87 (SD 14.2) for the control 10 group, p=0.039. No differences were found between the groups for systolic 11 blood pressure and diastolic blood pressure measures. 12 13 8.7.2.2 Does drug formulation affect adherence 14 Systematic Reviews 15 The possibility of bias was assessed to be high in a systematic review by 16 Bangalore (2007) (Bangalore, Sripal, Kamalakkannan, Gayathri, Parkar, 17 Sanobar et al , 2007) which included RCTs and retrospective reviews of data 18 bases. Nine studies were combined in a meta-analysis which included three 19 RCTs and four retrospective data bases of pharmacy claims. There was 20 marked heterogeneity in the compliance measures among the studies 21 evaluated and the patient group had different conditions which were being 22 treated. In the meta-analysis a total of 11,925 patients on fixed dose 23 combination were compared against 8317 patients on free drug component 24 regimen. Fixed dose combination resulted in a 26% decrease in the risk of 25 non compliance compared with free drug component regimen (pooled RR 26 0.74 [CI 0.69-0.80]; p<0.0001). There was no evidence of heterogeneity in this 27 analysis (p=.07). A subgroup analysis of the four studies on hypertension 28 showed that fixed dose combination (pooled RR 0.76 [CI 0.71-0.81]; 29 p<0.0001) decreased the risk of medication non-compliance by 24% 30 compared with free drug combination regimens. Due to methodological 31 concerns about the conduct of the meta-analysis, the results of this study 32 should be viewed with caution. Medicines concordance: full guideline DRAFT (July 2008) page 245 of 373 DRAFT FOR CONSULTATION 1 Randomized Controlled Trials 2 Brown (1997) (Brown, B. G., Bardsley, J., Poulin, D. et al , 1997) from the 3 Cochrane Review looked at the effect of different formulations as they 4 compared regular niacin versus polygel controlled release niacin. All patients 5 received lovastatin 20 mg, colestipol 10 g, and niacin 500 mg for 12 months, 6 with dosage adjustment to target cholesterol of 150 to 175 mg/dl, and to 7 minimize side effects. Twenty-nine male participants were enrolled aged ≤65 8 years At 12 months, patients were randomly assigned to 1) continue with 9 regular niacin at a dose identical to that established during the 12 month 10 dose-finding period, or 2) change to polygel controlled-release niacin at that 11 daily dosage, but given twice rather than 4 times/day. At 20 months, groups 1) 12 and 2) were reversed (crossover). This study was conducted in the USA. 13 Adherence was significantly greater for the controlled-release preparation. Medicines concordance: full guideline DRAFT (July 2008) page 246 of 373 DRAFT FOR CONSULTATION 1 8.8 Key Clinical Question: Is there any evidence on 2 interventions that aim to minimize side-effects in order 3 to increase adherence? Related references Evidence statements Evidence into (summary of evidence) recommendations Rickles (2005) (Rickles, 1. There is conflicting For general discussion of N. M., Svarstad, B. L., evidence with regard to limitations of evidence see section Statz-Paynter, J. L. et al , whether discussing side 7.3 2005); Collier (2005) effects, as part of a multi (Collier, A. C., Ribaudo, component intervention, H., Mukherjee, A. L. et al increases adherence. Interventions relating to side effects primarily involve providing information for patients on side , 2002); Kemp (1998) effects, No single way of providing (Kemp, R, Kirov, G, information on side effects with a Everitt, B et al , 1998). view to increase adherence to prescribed medication can be recommended. This is mainly due Rathbun (2005) 2. There is conflicting (Rathbun, R. C., Farmer, evidence with regard to K. C., Stephens, J. R. et whether educating patients al , 2005); Chaplin (1998) about side effects, as part of (Chaplin, R. and Kent, A., a multi component 1998); Canto De Cetina intervention, increases The GDG considered that there (2001) (Canto-De-Cetina, adherence. are a number of ways managing to the evidence not assessing the impact of minimizing side effects independently, thus not being able to ascertain their true effect. T. E., Canto, P., and side effects to support patient Ordoñez, Luna M., 2001); adherence. These include Tuldra (2000) (Tuldrà, A., adequately informing patients Fumaz, C. R., Ferrer, M. about side effects, exploring how J. et al , 2000); Peveler patient wants to manage side (1999) (Peveler, R., effects, reducing the dose of drug George, C., Kinmonth, A. and changing the drug to an L. et al , 1999). alternative. Medicines concordance: full guideline DRAFT (July 2008) page 247 of 373 DRAFT FOR CONSULTATION Howland (1990) 3. One RCT showed that (Howland, J. S., Baker, informing patients about side M. G., and Poe, T., effects did not have a 1990). significant effect on adherence. Vivian (2002) (Vivian, E. 4. There is conflicting M., 2002); Finley (2003) evidence with regard to (Finley, P. R., Rens, H. whether giving an intervention R., Pont, J. T. et al , deliverer (e.g. a pharmacist) 2003); Katon (2002) the power to adjust a patient’s (Katon, W., Russo, J., medication and/or dosage, as Von, Korff M. et al , part of a multi component 2002). intervention, increases adherence. Chisholm (2001) 5. There is conflicting (Chisholm, M. A., Mulloy, evidence with regard to L. L., Jagadeesan, M. et whether giving an intervention al , 2001); Adler (2004) deliverer (e.g. a pharmacist) (Adler, D. A., Bungay, K. the power to make M., Wilson, I. B. et al , recommendations about the 2004). treatment to the patient’s practitioner, as part of a multi component intervention, increases adherence. 1 2 *e.g. relative value place on different outcomes, trade-offs between benefits and harms, uncertainties in Evidence, likely impact if implements, feasibility, patient perspective 3 Medicines concordance: full guideline DRAFT (July 2008) page 248 of 373 DRAFT FOR CONSULTATION Methods of the evidence review 1 8.8.1 2 This paper includes a narrative summary of the included evidence, structured 3 according to the category of the intervention, following the agreed reviewing 4 protocol: 5 Types of studies - randomised controlled trials (RCTs) of interventions to 6 increase adherence. The excluded studies list from the Cochrane review was 7 checked as we have included those studies with less than 80% follow-up of 8 participants. 9 Types of participants - people prescribed medication for a medical condition. 10 Duration of studies - six months follow up from the time of patient entry for 11 long-term regimens. No time limit specified for short-term conditions. 12 Types of interventions - any intervention intended to change adherence to 13 prescribed medication. As the Cochrane review is presented by condition, we 14 have used the evidence extracted in that review and reconfigured it by 15 intervention. 16 Types of outcome measures – inclusion criteria (as defined in the Cochrane 17 review) were expanded by including studies that used adherence as the only 18 outcome variable as opposed to adherence and treatment outcome variables. 19 The excluded studies list of the Cochrane review was cross-referenced to 20 ensure that no potentially relevant study was missed out. 21 8.8.2 22 Although a number of RCTs were found which addressed side effects we did 23 not find many where the interventions’ sole purpose was to address side 24 effects. Within the RCTs which did address side effects there was a lot of 25 variability in how they did this and to what extent side effects were a focus of 26 the intervention. What follows is a summary of the RCTs which addressed 27 side effects with information on interventions limited to those parts of the study 28 which addressed side effects. For further details please see evidence tables 29 for the RCTs retrieved from update searches, or the Cochrane review. Evidence review Medicines concordance: full guideline DRAFT (July 2008) page 249 of 373 DRAFT FOR CONSULTATION 1 8.8.2.1 Discussing side effects with patients 2 RCTs were excluded from this section if side effects were only discussed, and 3 if no further details were given on how to educate the patients in managing 4 their side-effects with a view to increase adherence. Other RCTs were 5 excluded if addressing side effects was potentially part of a multi-component 6 intervention but when the actual decision to address side effects was the 7 choice of the intervention provider (therefore meaning it was impossible to tell 8 how many patients in the intervention group had side effects addressed and 9 how many did not have this issue addressed). 10 3 RCTs addressed side effects by clearly discussing them with patients. 11 Rickles (2005) (Rickles, N. M., Svarstad, B. L., Statz-Paynter, J. L. et al , 12 2005) had pharmacists address adverse events during telephone calls in 98 13 patients with a mean age of 38 years. The pharmacist could probe or explain 14 issues not understood by the patient and make recommendations. The 15 intervention had positive effects. There was significant difference at six 16 months in adherence with the rate of missed doses significantly lower in the 17 intervention group (30.3%, SD 36.4 vs. 48.6%, SD 39.2, p = < 0.05). This 18 study was conducted in the USA. Collier (2005) (Collier, A. C., Ribaudo, H., 19 Mukherjee, A. L. et al , 2002) had nurses address participants’ medication- 20 related behaviour and barriers to adherence during telephone calls in 282 21 patients. Advice around side effects was offered. Over 24 months, rates of 22 adherence were high in both groups (>72% reported at least 95% adherence). 23 No difference was seen between groups (OR 0.86, 95% CI 0.57 to 1.29). This 24 study appears to have been conducted in the USA. Kemp (1998) (Kemp, R, 25 Kirov, G, Everitt, B et al , 1998), an RCT from the Cochrane review, as part of 26 “compliance therapy”, had 2 sessions where intervention group participants 27 focused on symptoms and the side effects. This study had 74 participants. 28 The mean age in the intervention group was 34 years (SD: 10.6) and 37 years 29 (SD: 11.9) in the control group. This study was conducted in the UK. 30 Patients receiving compliance therapy demonstrated higher adherence ratings 31 (p < 0.001) than control group patients. Medicines concordance: full guideline DRAFT (July 2008) page 250 of 373 DRAFT FOR CONSULTATION 1 8.8.2.2 Educating patients and follow up 2 5 RCTs addressed side effects through educating the patient about them and 3 then following up on this education. 4 Rathbun (2005) (Rathbun, R. C., Farmer, K. C., Stephens, J. R. et al , 2005) 5 provided patients with a mean age of 38.0 years with education about 6 adverse-event management strategies, among other things. A total of 43 7 patients were included in this study. Telephone follow-up followed to identify 8 early problems. At week 28, adherence rates were 74% (SD 31%) in the 9 intervention group and 51% (SD 41%) in the control group (p = 0.080, 10 difference between groups 23%, 95% CI: 1% to 44%). Mean decline in 11 adherence between weeks 4 and 28 were 12% (p = 0.15) in the intervention 12 group and 22% (p = 0.002) in the control group. Patients in the intervention 13 group were more likely to take their medication at the prescribed dosing 14 schedule: at 4 weeks, 69% in the intervention group vs 42% in the control 15 group (p = 0.025) and at 28 weeks, 53% in the intervention group vs 31% in 16 the control group (p = 0.046). No significant difference was seen between the 17 groups based on patient self-report (94% vs 89% intervention vs control, 18 p=0.51). This study was conducted in the USA. 19 Chaplin (1998) (Chaplin, R. and Kent, A., 1998), an RCT from the Cochrane 20 review, had intervention group participants, “…participate in a discussion 21 about the risks and benefits of neuroleptic medications based on individual 22 semi-structured educational sessions with reference to a standardised 23 information sheet. The patients were asked whether they had heard of tardive 24 dyskinesia. The common movements of TD were modelled and the patients 25 were asked whether they thought they had the condition or had seen others 26 with it. They were informed that they were receiving an antipsychotic drug and 27 were given information about extrapyramidal symptoms and TD, its risk 28 factors, prevalence, treatment, potential irreversibility and the 1% risk of TD in 29 non-antipsychotic-treated patients. They were told that gradual discontinuation 30 of antipsychotic medication was the best way to prevent the condition but if 31 done abruptly carries a high risk of relapse and of precipitating TD. It was 32 stated that the optimum maintenance treatment, taking into account its risks Medicines concordance: full guideline DRAFT (July 2008) page 251 of 373 DRAFT FOR CONSULTATION 1 and benefits, was to use the lowest dose of antipsychotic drug that would 2 keep them well. Most importantly, they were asked not to make any changes 3 to their treatment without discussion with their psychiatrist. Finally, they were 4 given the opportunity to ask questions in an informal interactive session 5 lasting 30 minutes, and were given an information sheet for reference” 6 (Haynes, R. B., Yao, X., Degani, A. et al , 2005). This study included 56 7 participants (age range not given). This study was conducted in the UK. The 8 intervention did not increase adherence relative to the control condition. 9 Canto De Cetina (2001) (Canto-De-Cetina, T. E., Canto, P., and Ordoñez, 10 Luna M., 2001), an RCT from the Cochrane review, had women in their 11 intervention group (counselling group) receive, “…a structured pretreatment 12 counselling with indications about the mode of action of DMPA, the common 13 side effects of the drug, including the possibility of irregular menstrual periods, 14 heavy bleeding, spotting, and amenorrhea. To mentally prepare users for 15 potential side effects, it was stressed that these side effects would be not 16 detrimental to their health. These indications were repeated at each follow-up 17 visit” (Haynes, R. B., Yao, X., Degani, A. et al , 2005). This study included 350 18 participants. The mean age in the counselling group was 33.9 years with a 20- 19 35 range and 34 years in the control group with also a 20-35 range. This study 20 was conducted in Mexico. There was a positive effect of the intervention in 21 terms of cumulative termination rates. 22 Tuldra (2000) (Tuldrà, A., Fumaz, C. R., Ferrer, M. J. et al , 2000), an RCT 23 from the Cochrane review, investigated a psycho-educative intervention part 24 of which involved participants being taught how to manage medication and 25 tackle problems such as forgetting, delays, side effects and changes in the 26 daily routine. During follow-up participants were provided with skills to deal 27 with minor adverse effects. This study had 116 participants. The mean age in 28 the intervention group was 39 years (SD: 10) and 38 years (SD: 7) in the 29 control group. It appears this study was conducted in Spain. “In an intention to 30 treat (ITT) analysis, no improvements were found in adherence (the p-values 31 were slightly above the 0.05 significance level). However, when a per protocol 32 analysis was conducted, the intervention resulted in improvements in Medicines concordance: full guideline DRAFT (July 2008) page 252 of 373 DRAFT FOR CONSULTATION 1 compliance to HAART at 48 weeks. The lack of statistical significance 2 observed using the ITT analysis might be a reflection of a low power to detect 3 differences due to the relatively small sample size for each arm (n = 55 for 4 intervention, n = 61 for control). The per protocol analysis is suspect in any 5 adherence study as it ignores patients who dropped out, the most severe form 6 of non-adherence.” (Haynes, R. B., Yao, X., Degani, A. et al , 2005). 7 Peveler (1999) (Peveler, R., George, C., Kinmonth, A. L. et al , 1999), an 8 RCT from the Cochrane review, compared four treatment groups, 9 “…treatment as usual, leaflet, drug counselling, or both interventions. The 10 information leaflet contained information about the drug, unwanted side 11 effects, and what to do in the event of a missing dose. Patients were given 12 drug counselling by a nurse at weeks 2 and 8, according to a written protocol. 13 Sessions included assessment of daily routine and lifestyle, attitudes to 14 treatment, and understanding of the reasons for treatment...The importance of 15 drug treatment was emphasized, and side effects and their management 16 discussed.” (Haynes, R. B., Yao, X., Degani, A. et al , 2005). This study had 17 213 participants. The mean age of participants was 45.3 years with a range of 18 21-83. This study was conducted in the UK. “The treatment leaflets had no 19 effect on adherence…This study was only 12 weeks in duration, which is 20 shorter than our usual 6 months follow-up criterion. However, because the 21 results were negative for adherence and clinical outcomes with the leaflet 22 intervention, the paper was included for this review. (Counselling about drug 23 treatment, however, did result in significant improvements in adherence and 24 clinical outcomes. Nonetheless, because the follow-up was less than six 25 months in duration, the results for counselling are not considered in the 26 conclusions of this review.)” (Haynes, R. B., Yao, X., Degani, A. et al , 2005). 27 Howland (1990) (Howland, J. S., Baker, M. G., and Poe, T., 1990), an RCT 28 from the Cochrane review, informed intervention group patients of six possible 29 side-effects of treatment with erythromycin, while control (uninformed) patients 30 were not made aware of potential side effects of treatment. This study had 98 31 participants. The mean age in the intervention group was 50 years and 48 32 years in the control group. It appears this study was conducted in the USA. Medicines concordance: full guideline DRAFT (July 2008) page 253 of 373 DRAFT FOR CONSULTATION 1 The intervention did not increase adherence relative to the control condition 2 nor did it decrease adherence. 3 4 8.8.2.3 Adjusting medication and/or dosage 5 3 RCTs addressed side effects by giving the intervention deliverer power to 6 adjust medication and/or medication dosage. 7 In a study by Vivian (2002) (Vivian, E. M., 2002) intervention patients saw 8 clinical pharmacists who could make changes in the prescribed drugs and 9 dosages and provided medication counselling centred around the discussion 10 of side effects, lifestyle and adherence (note we have made an assumption 11 here that the discussion of side effects and changes in medication are related, 12 this is not explicitly stated in the study). Fifty seven patients aged above 18 13 years were included in this study. The majority of the study population was 14 African American (77%). There were no significant differences in compliance 15 (from self report measure) between (p > 0.25, mean and SD not given for 16 adherence) or within (p = 0.07) the two groups at baseline or at the end of the 17 study. This study was conducted in the USA. 18 Finley (2003) (Finley, P. R., Rens, H. R., Pont, J. T. et al , 2003) had 19 pharmacists provide a detailed explanation of the role of antidepressants 20 (including potential therapeutic effects and adverse effects). Care managers 21 were permitted to titrate antidepressant drugs in a fashion consistent with the 22 HMO’s clinical guidelines and current recommended practices (note we have 23 made an assumption here that the discussion of side effects and changes in 24 medication are related, this is note explicitly stated in the study). . During 25 follow-up phone calls and clinic appointments pharmacists followed a 26 standardized set of questions that assessed adverse effects, among other 27 things. One hundred and twenty five patients were included in the study. The 28 majority of the study population was female ( 42 (84) % in the control group 29 and 64 (85) % in the intervention group; and the mean ages (s.d.) were 54.1 ± 30 17.3 years in the control group and 54.4± 14.1 years in the intervention group. 31 After 6 months, the intervention group demonstrated a significantly higher Medicines concordance: full guideline DRAFT (July 2008) page 254 of 373 DRAFT FOR CONSULTATION 1 drug adherence rate than that of the control group (67% vs. 48%, p = 0.038). 2 This study was conducted in the USA. 3 As part of a multifaceted intervention, Katon (2002) (Katon, W., Russo, J., 4 Von, Korff M. et al , 2002) scheduled two sessions for intervention patients 5 with a psychiatrist in a primary care clinic. The study included 228 patients 6 aged between 18 and 80 years. Mean ages were 47.2 ± 14.0 years in the 7 intervention group and 46.7 ± 13.4 years in the usual care group. When 8 severe side effects or inadequate response to treatment occurred, the 9 psychiatrist helped the patient and primary care physician alter the dosage or 10 choose an alternative medication. There were no differences between the four 11 study groups in either adherence to the care suggestions, combined or 12 individually. There were no inter-group differences in medication adherence. 13 This study was conducted in the USA. 14 8.8.2.4 15 2 RCTs addressed side effects by giving the intervention deliverer power to 16 make recommendations to other health care professionals involved in the 17 patients care. 18 Chisholm (2001) (Chisholm, M. A., Mulloy, L. L., Jagadeesan, M. et al , 2001) 19 examined an intervention which included a pharmacist taking medication 20 histories and reviewing medications with the patient, with an emphasis on 21 optimizing medication therapy to achieve compliance outcomes while 22 minimizing adverse events related to medication. Twenty four patients aged 23 between 18 and 60 years were included in the study. Majority of the study 24 population was male (75%). The clinical pharmacist also provided 25 recommendations to the nephrologists with the goal of achieving desired 26 outcomes. Counselling involved discussion of patients concerns around their 27 medication therapy and instructing them how to properly take their 28 medications. Counselling was both verbal and/or in writing. At 12 months the 29 mean compliance rate in the intervention group was 96.1% (SD 4.7%) 30 compared to 81.6% (SD 11.5%) in the control group (p < 0.0001). For 6 of the 31 12 months, higher rates of compliance were seen in the intervention group (p 32 < 0.05). Also, 75% (n = 9) of the intervention patients were compliant each Medicines concordance: full guideline DRAFT (July 2008) page 255 of 373 Recommendations to health care professionals DRAFT FOR CONSULTATION 1 month compared to 33.3% (n = 4) of the control group. This study was 2 conducted in the USA. 3 Adler (2004) (Adler, D. A., Bungay, K. M., Wilson, I. B. et al , 2004) employed 4 a pharmacist intervention which emphasised, among other things; assessing a 5 patient's medication regimen for drug-related problems (such as side effects 6 or drug interactions); monitoring drug efficacy and toxicity and educating 7 patients about depression and antidepressants. This study included 533 8 patients aged above 18 years. The mean age was 42.3 years, and the 9 majority was female. After an initial appointment with the patient, pharmacists 10 provided the patients PCP with a thorough medication history (including 11 adherence to prescribed medications and drug-related problems) and 12 whatever recommendations the pharmacist may have suggested to improve 13 the regimen. For patients using antidepressants at study entry (n = 227) there 14 were no significant differences in antidepressant usage between the 15 intervention and control groups either at 3 (90.7% vs. 87.2, p = 0.50) or 6 16 months (83.4% vs. 78.4%, p = 0.33). This study was conducted in the USA. 17 18 19 20 Medicines concordance: full guideline DRAFT (July 2008) page 256 of 373 DRAFT FOR CONSULTATION 1 8.9 Key Clinical Question: How does the way the 2 information is presented (e.g. pictorial vs. written) 3 affects adherence? Related Evidence statements references (summary of evidence) Raynor (2007) 1. One high quality systematic For general discussion of limitations of (Raynor, D. K., review of quantitative and evidence see section 7.3 Blenkinsopp, A., qualitative research on the role Knapp, P. et al , and effectiveness of written 2007) information available to patients about individual medicines stated that no robust evidence was found that the information (delivery) had an effect on patient satisfaction or compliance. Schaffer (2004) 2. One RCT showed that written (Schaffer, S. D. information alone and written and and Tian, L., verbal information resulted in 2004) greater improvements in Evidence into recommendations While there is no conclusive evidence about the effectiveness of the mode of delivery of information affects adherence, in certain cases/diseases it made a difference. Thus, it should be tailored to the individual’s choices and preferences. adherence to prescribed medication compared to verbal information alone. Segador (2005) 3. One RCT showed that verbal (Segador, J., Gil- and written information compared Guillen, V. F., to verbal information alone Orozco, D. et al , resulted in significantly greater 2005) improvements in adherence to prescribed medication. Medicines concordance: full guideline DRAFT (July 2008) page 257 of 373 DRAFT FOR CONSULTATION Atherton-Naji 4. One RCT showed that simple (2001) (Atherton- tailored information (mailed Naji , A., leaflets with written and pictorial Hamilton, R., information) did not significantly Riddle, W. et al , improve adherence to prescribed 2001) medication compared to usual care. 1 2 *e.g. relative value place on different outcomes, trade-offs between benefits and harms, uncertainties in Evidence, likely impact if implements, feasibility, patient perspective 3 Medicines concordance: full guideline DRAFT (July 2008) page 258 of 373 DRAFT FOR CONSULTATION Methods of the evidence review 1 8.9.1 2 This paper includes a narrative summary of the included evidence, structured 3 according to the category of the intervention, following the agreed reviewing 4 protocol: 5 Types of studies - randomised controlled trials (RCTs) of interventions to 6 increase adherence. The excluded studies list from the Cochrane review was 7 checked as we have included those studies with less than 80% follow-up of 8 participants. 9 Types of participants - people prescribed medication for a medical condition. 10 Duration of studies - six months follow up from the time of patient entry for 11 long-term regimens. No time limit specified for short-term conditions. 12 Types of interventions - any interventions intended to change adherence to 13 prescribed medication. As the Cochrane review is presented by condition, we 14 have used the evidence extracted in that review and reconfigured it by 15 intervention. 16 Types of outcome measures - inclusion criteria (as defined in the Cochrane 17 review) were expanded by including studies that used adherence as the only 18 outcome variable as opposed to adherence and treatment outcome variables. 19 The excluded studies list of the Cochrane review was cross-referenced to 20 ensure that no potentially relevant study was missed out. 21 8.9.2 22 A health technology assessment report of a “Systematic review of quantitative 23 and qualitative research on the role and effectiveness of written information 24 available to patients about individual medicines” (Raynor, D. K., Blenkinsopp, 25 A., Knapp, P. et al , 2007) was retrieved. This report aimed to address the role 26 and value of written information given to patients; and how effective this 27 information is in improving patient’s knowledge of their treatment and health 28 outcomes. The inclusion criteria of this review was broader than those applied 29 in our reviews, as members of the public not currently taking medication; Evidence review Medicines concordance: full guideline DRAFT (July 2008) page 259 of 373 DRAFT FOR CONSULTATION 1 general public using over the counter medicines and some addiction therapies 2 were included. Also, studies did not necessarily need to have an aim to 3 increase adherence. Despite these differences, it was felt that due to the high 4 quality of this document, it would be relevant to refer to it in this section to 5 support the evidence included in the reviews and inform the decision making 6 process. 7 Key findings of the report show that the majority of people do not value the written information they 8 receive, and 9 no robust evidence was found that the information had any effect 10 on patient satisfaction or compliance. 11 12 Most patients did not value the current package insert patient information 13 leaflets (PILS) and did not consider information written by medicine 14 manufacturers to be sufficiently independent. 15 The (PILS ) supplied had deficiencies in the content (e.g. complexity of 16 language) and layout (e.g. print size). However, it did show that patients 17 valued written information that contained condition-based details along with 18 the medicines information, in addition to alternative treatments for the 19 condition. 20 In addition, the qualitative evidence included in the report did not show that 21 patients perceive improvement of compliance as a function of PILs. This can 22 be explained by how an informed decision not to take medication is a 23 legitimate and acceptable outcome. In contrast, some health care 24 professionals viewed that the increase of compliance was one of the main PIL 25 uses. 26 The key points for improvement of written medicines information outlined by 27 the review were: 28 29 The need to involve patients in all stages of the process, as to reflect better their needs. Medicines concordance: full guideline DRAFT (July 2008) page 260 of 373 DRAFT FOR CONSULTATION To incorporate the findings from the review to improve future 1 information design and content 2 To present risk information numerically instead of verbal 3 descriptions. 4 5 8.9.2.1 Verbal vs. written vs. verbal and written vs. usual care 6 Schaffer (2004) (Schaffer, S. D. and Tian, L., 2004), a study from the 7 Cochrane review, compared two interventions and a combination of the two 8 control groups in patients with asthma aged 18-65 (n = 46) with results 9 dependant on the measure of adherence used. There were 4 groups, 10 “…standard provider education (control group); (b) audiotape alone; (c) 11 National Heart Lung and Blood Institute (NHLBI) booklet alone; and (d) 12 audiotape plus NHLBI booklet”. This study was conducted in the United 13 States. 14 “The results showed a significant increase in adherence by pharmacy-refill 15 measure (but not by self-report) for NHLBI booklet versus control, and for 16 NHLBI booklet plus audiotape versus control, but not for audiotape versus 17 control at six months”. 18 8.9.2.2 19 Segador (2005) (Segador, J., Gil-Guillen, V. F., Orozco, D. et al , 2005) 20 compared the effect of written information in addition to verbal information in 21 patients receiving antibiotic treatment for acute sore throat (n = 158) with 22 significant results. Patients in the written information group were given written 23 information at the time of their first visit to their GP. The written information 24 emphasized the importance of completing the antibiotic treatment, of 25 respecting intervals between doses and the drawbacks of an early drop-out, 26 and was given only at the time of initial consultation. The control group was 27 given verbal information only. This study was conducted in Spain. 28 The pill count average was 87.4+/-25.2% and it was higher in the intervention 29 group (93.7+/-24.5%) than in the control group (81.1+/-24.5%) (P < 0.05). Verbal and written information vs. verbal information alone Medicines concordance: full guideline DRAFT (July 2008) page 261 of 373 DRAFT FOR CONSULTATION 1 8.9.2.3 Written and/or pictorial information given vs. usual care 2 Atherton-Naji (2001) (Atherton-Naji , A., Hamilton, R., Riddle, W. et al , 2001) 3 compared an educational intervention to routine care in patients (n = 45) with 4 depression with non-significant results. Patients in the intervention group 5 received simple tailored information (mailed leaflets with written and pictorial 6 information) at 1, 6 and 16 weeks after the initial prescription. The leaflets 7 contained basic information about the condition, treatment and general 8 problems people may have with adherence to the treatment. Leaflets were 9 personalised for each patient and their specific drug. This study was 10 conducted in the UK. Over 6 months, 35.6% (n = 16) collected prescriptions at 11 each month (no significant difference between groups). The proportion 12 decreased over time from month 1 (intervention group: 95.8% (n = 23) vs. 13 control group: 100% (n = 21)) to month 6 (intervention group: 58.3% (n = 14) 14 vs. control group: 52.4% (n = 11)). 15 Medicines concordance: full guideline DRAFT (July 2008) page 262 of 373 DRAFT FOR CONSULTATION 1 2 8.10 Key Clinical Question: Do specific forms of therapy (e.g. CBT) affect adherence? Related references Evidence Evidence into recommendations statements (summary of evidence) All studies included in 1. There is some For general discussion of limitations of evidence evidence review evidence that see section 7.3 elements of CBT or psychobehavioural can help but the quality of the evidence does not allow us to generalise these results. The evidence concerning specific forms of therapy and their effect on adherence is inadequate. Conclusions from a variety of single studies with ill defined content and delivery are inconclusive. The GDG noted that definitions of CBT and other therapies were often unclear. It was also argued that some of these intervention may not be salient to medicine-taking behaviou Gray 2006 (Gray, R., 2. The majority of The available evidence is that patients make Leese, M., Bindman, evidence suggests their own appraisal of medicines based on J. et al , 2006); Wyatt that CBT approaches factors important to them and in this context the 2004 (Wyatt, G. E., do not improve behaviour is rational and coherent and as such Longshore, D., Chin, adherence relative to not appropriate for CBT D. et al , 2004); other forms of Bechdolf 2004 and treatment. 2005 (Bechdolf, A., Knost, B., Kuntermann, C. et al , Therapies which worked with patients and families addressing social and cultural issues d provide evidence of specific principles of engaging with patients that may be of value. 2004) (Bechdolf, A., Köhn, D., Knost, B. et al , 2005); Weber 2004 (Weber, R., Christen, L., Christen, Medicines concordance: full guideline DRAFT (July 2008) page 263 of 373 DRAFT FOR CONSULTATION S. et al , 2004); Antoni, 2006 (Antoni, M. H., Carrico, A. W., Duran, R. E. et al , 2006); Wagner 2006 (Wagner, G. J., Kanouse, D. E., Golinelli, D. et al , 1909); Lam, 2003 (Lam, D. H., Watkins, E. R., Hayward, P. et al , 2003); Strang (1981) (Strang, 3. One RCT showed J. S., Falloon, I.-R. H., that family therapy and Moss, H. B., increased adherence 1981) to prescribed medication when compared to individual support sessions. Xiong (1994) (Xiong, 4. Two RCTs showed W., Phillips, M. R., that family therapy Hu, X. et al , 1994); did not increase Zhang (1994) (Zhang, adherence to M., Wang, M., Li, J. et prescribed al , 1994) medication when compared to standard care. Miklowitz, 2003 5. One RCT showed (Miklowitz, D. J., that family therapy Medicines concordance: full guideline DRAFT (July 2008) page 264 of 373 DRAFT FOR CONSULTATION George, E. L., and Richards, J. A. et al , pharmacotherapy 2003); increased adherence to prescribed medication when compared to crisis management and pharmacotherapy. Razali, 2000 (Razali, 6. One RCT showed S. M., Hasanah, C., I, that culturally Khan, U. A. et al , modified family 2000) therapy increases adherence when compared to behavioural family therapy. Remien, 2005 7. One RCT showed (Remien, R. H., that couple based Stirratt, M. J., Dolezal, therapy increases C. et al , 2005) adherence compared to usual care. Ruskin, 2004 (Ruskin, 8. One RCT showed P. E., Silver-Aylaian, that telepsychiatry M., Kling, M. A. et al , did not increase 2004) adherence compared to face to face psychiatry. Kemp 1996, 1998 9. There is conflicting (Kemp, R., Hayward, evidence with Medicines concordance: full guideline DRAFT (July 2008) page 265 of 373 DRAFT FOR CONSULTATION P., Applewhaite, G. et regards to whether al , 1996) (Kemp, R, compliance therapy Kirov, G, Everitt, B et increases adherence al , 1998); O’Donnell compared to 2003 (O'Donnell, counselling Colin, Donohoe, Gary, Sharkey, Louise et al , 2003). Pradier, 2003 10. There is (Pradier, Christian, conflicting evidence Bentz, Laurence, to suggest that Spire, Bruno et al , multicomponent 2003); Van Servellen interventions mainly (2005) (van, based on Servellen, G, motivational Nyamathi, A., Carpio, principles increase F. et al , 2005) adherence Weber 2004 (Weber, 11. One RCT R., Christen, L., showed that CBT Christen, S. et al , and usual care did 2004) not increase adherence compared to usual care alone. Gray, 2006 (Gray, R., One RCT showed Leese, M., Bindman, that CBT and health J. et al , 2006). education did not increase adherence compared to health education alone. Medicines concordance: full guideline DRAFT (July 2008) page 266 of 373 DRAFT FOR CONSULTATION Bechdolf, 2004 and 12. One RCT 2005 (Bechdolf, A., showed that CBT Knost, B., and health education Kuntermann, C. et al , did not increase 2004) (Bechdolf, A., adherence compared Köhn, D., Knost, B. et to psycho-education. al , 2005). Antoni 2006 (Antoni, 13. One RCT M. H., Carrico, A. W., showed that Duran, R. E. et al , cognitive behavioural 2006) stress management in addition to antiretroviral medication adherence training did not increase adherence compared to medication adherence training alone. 1 2 *e.g. relative value place on different outcomes, trade-offs between benefits and harms, uncertainties in Evidence, likely impact if implements, feasibility, patient perspective 3 4 8.10.1 Methods of the evidence review 5 This paper includes a narrative summary of the included evidence, structured 6 according to the category of the intervention, following the agreed reviewing 7 protocol: 8 Types of studies - randomised controlled trials (RCTs) of interventions to 9 increase adherence. The excluded studies list from the Cochrane review was Medicines concordance: full guideline DRAFT (July 2008) page 267 of 373 DRAFT FOR CONSULTATION 1 checked as we have included those studies with less than 80% follow-up of 2 participants. 3 Types of participants - people prescribed medication for a medical condition. 4 Duration of studies - six months follow up from the time of patient entry for 5 long-term regimens. No time limit specified for short-term conditions. 6 Types of interventions - any interventions intended to change adherence to 7 prescribed medication. As the Cochrane review is presented by condition, we 8 have used the evidence extracted in that review and reconfigured it by 9 intervention. 10 Types of outcome measures – inclusion criteria (as defined in the Cochrane 11 review) were expanded by including studies that used adherence as the only 12 outcome variable as opposed to adherence and treatment outcome variables. 13 The excluded studies list of the Cochrane review was cross-referenced to 14 ensure that no potentially relevant study was missed out. 15 8.10.2 Evidence review 16 8.10.2.1 17 Miklowitz (2003) (Miklowitz, D. J., George, E. L., Richards, J. A. et al , 2003) 18 compared family focused therapy and pharmacotherapy with crisis 19 management and pharmacotherapy (serving as control group) in patients with 20 bi-polar disorder with positive results. The study included 101 participants with 21 ages that ranged from 18 to 62 years (mean age 35.6 ± 10.2 years). Family 22 focused therapy involved three modules: 1/ psycho-education, which involved 23 passing on information about the disorder, its aetiology, signs, symptoms and 24 also information on how to prevent relapse; 2/ communication training where, 25 through role play, skills of listening, offering feedback, and requesting 26 changes in behaviour were passed on; and 3/ problem solving skills, where 27 participants identified potential problems, came up with and evaluated various 28 solutions. Family focused therapy involved approximately 21 sessions over a 29 nine month period and was conducted with the whole of the patient’s family at 30 the patients/family’s home. This study was conducted in the USA. Medicines concordance: full guideline DRAFT (July 2008) page 268 of 373 Family Therapy DRAFT FOR CONSULTATION 1 Patients in the intervention group had higher mean drug adherence scores 2 during follow up (2.77 +/- 0.43) than patients in the control group (2.56 +/- 3 0.48, p = 0.04). This study was conducted in the USA. 4 Razali (2000) (Razali, S. M., Hasanah, C., I, Khan, U. A. et al , 2000), a RCT 5 from the Cochrane review, compared the effects of “culturally modified family 6 therapy“(CMFT) to the effects of ”behavioural family therapy“ (BFT serving as 7 the control condition) in patients with schizophrenia. This study included 166 8 participants, with ages ranging from 17 to 55 years. The majority of the 9 patients came from a low socio-economic background. The CMFT was 10 delivered by a psychiatrist and sessions were given monthly for the first 3 11 months and then every 6 weeks in the following months. The CMFT consisted 12 of a “….Socio-cultural approach of family education, drug intervention 13 programme and problem-solving skills. The socio-cultural approaches to 14 family education include explanations of the concept of schizophrenia from a 15 cultural perspective and an attempt to correct negative attitudes toward 16 modern treatment. The family education and drug intervention was delivered 17 as a package. The drug intervention programme included drug counselling, 18 clear instruction about dose, frequency and possible side effects, the role of 19 carers in supervision of medication at home, and close monitoring of 20 compliance by a drug intake check-list presented in every follow-up visit 21 (Haynes, R. B., Yao, X., Degani, A. et al , 2005).” This study was conducted in 22 Malaysia. 23 At six months and one year, patients in the intervention group (CMFT) had 24 significantly higher compliance than those in the control (BFT) group. 25 Strang (1981) (Strang, J. S., Falloon, I.-R. H., and Moss, H. B., 1981), a RCT 26 from the Cochrane review, compared family therapy or individual support 27 sessions in patients with schizophrenia with positive results. Thirty two 28 patients were enrolled in this study. No information on the ages of the patients 29 was given. All patients had scheduled therapy and monthly medication 30 appointments. Patients were allocated to family therapy or individual support 31 sessions. This study was conducted in the UK. Medicines concordance: full guideline DRAFT (July 2008) page 269 of 373 DRAFT FOR CONSULTATION 1 The patients in the family therapy group were significantly more adherent than 2 those in the individual support group. 3 Xiong (1994) (Xiong, W., Phillips, M. R., Hu, X. et al , 1994), a study from 4 the Cochrane review, compared a family based intervention with standard 5 care in patients with schizophrenia with negative results. Sixty three families 6 were enrolled in this study and mean age was 31 years (ranging from 17 to 54 7 years). The family based intervention “…included monthly 45 minute 8 counselling sessions focused on the management of social and occupational 9 problems, medication management, family education, family group meetings, 10 and crisis intervention”. (Haynes, R. B., Yao, X., Degani, A. et al , 2005). This 11 study was conducted in China. There was no difference in terms of adherence 12 between the two groups. 13 Zhang (1994) (Zhang, M., Wang, M., Li, J. et al , 1994), a study from the 14 Cochrane review, compared a family intervention with no addition care above 15 standard care in patients discharged after their first admission to the hospital 16 for schizophrenia with negative results. This study included 83 patients. Mean 17 ages were 23.5 ± (7.6) (s.d.) for the intervention group and 24.1 ± (8.1) (s.d.) 18 for the control group Families and patients in the family intervention group 19 were, “…assigned to one of two counsellors for their ongoing care, were 20 invited to come to a discharge session that focused on education about the 21 management of the patient’s treatment, asked to come to a family group 22 counselling session with other families three months after discharge, and then 23 attend three-monthly group sessions with other families with similar patient 24 problems. Non-attendance triggered a visit from study staff. Each family was 25 contacted at least once during the 18-month follow-up.” (Haynes, R. B., Yao, 26 X., Degani, A. et al , 2005). This study was conducted in China. There was no 27 difference in terms of adherence between the two groups. 28 29 8.10.2.2 Couples Therapy 30 Remien (2005) (Remien, R. H., Stirratt, M. J., Dolezal, C. et al , 2005) 31 compared a couple-based ART adherence intervention with usual care in HIVMedicines concordance: full guideline DRAFT (July 2008) page 270 of 373 DRAFT FOR CONSULTATION 1 serodiscordant couples with positive results. The intervention included 2 structured discussions and instruction, as well as specific problem-solving and 3 couple-communication exercises. A total of 215 couples aged >18 years were 4 enrolled in the study. The mean age of the participants was 42 years. The 5 study sample mainly consisted of lower-income racial/ethnic minorities. Key 6 components included education about the importance of adherence to avoid 7 viral resistance and maintain health, identifying patterns of non-adherence, 8 developing communication and problem-solving strategies to overcome 9 adherence barriers, optimising partner support and building confidence in the 10 couple for achieving and maintaining improved adherence. In addition the 11 intervention sought to help couples to address issues of sex and intimacy. The 12 intervention was administered to each couple by a nurse practitioner through 13 45-60 min sessions held over 5 weeks. The study was conducted in the USA. 14 At 6 months there were significant differences in adherence change between 15 the 2 groups. 16 Significant group differences in adherence change from baseline to week 8 in 17 terms of proportion of prescribed doses taken (p = 0.021) and proportion of 18 doses taken within specified windows (p< 0.001). At 3 months, only the 19 proportion of doses taken within specified time windows was significant (p = 20 0.028). 21 8.10.2.3 22 Ruskin (2004) (Ruskin, P. E., Silver-Aylaian, M., Kling, M. A. et al , 2004) 23 compared patients being seen by a psychiatrist, either in person or by means 24 of telepsychiatry, who had one of the following five diagnoses: major 25 depressive disorder, dysthymic disorder, adjustment disorder with depressed 26 mood, mood disorder due to a general medical condition, or depressive 27 disorder not otherwise specified with negative results. One hundred and thirty 28 one patients were enrolled in the study. Mean age of participants was 49.7 29 (12.8) years. Treatment sessions lasted approximately 20 minutes and 30 consisted of antidepressant medication management, psycho-education, and 31 brief supportive counselling. Treatment consisted of eight sessions with a 32 psychiatrist over a 6-month period. This study was conducted in the USA. Medicines concordance: full guideline DRAFT (July 2008) page 271 of 373 Telepsychiatry DRAFT FOR CONSULTATION 1 There was no difference in the percentage of adherent patients between the 2 two treatment groups. 3 8.10.2.4 4 Kemp’s (1996, 1998) (Kemp, R., Hayward, P., Applewhaite, G. et al , 1996) 5 (Kemp, R, Kirov, G, Everitt, B et al , 1998) RCTs from the Cochrane review, 6 compared “compliance therapy” with supportive counselling sessions (serving 7 as the control group) in patients with psychotic disorders with significant 8 results. Forty-seven patients aged 18 to 65 years were included in the 1996 9 study and 74 patients also aged 18 to 65 years in the 1998 study. Compliance Compliance Therapy 10 therapy consisted of 4 to 6 sessions and was defined as, “…a strategy that 11 borrows from motivational interviewing. During session 1 and session 2, 12 patients reviewed their illness and conceptualized the problem. In the next 2 13 sessions, patients focused on symptoms and the side effects of treatment. In 14 the last 2 sessions, the stigma of drug treatment was addressed (Haynes, R. 15 B., Yao, X., Degani, A. et al , 2005)”. This study was conducted in the UK. 16 At 12 months patients receiving compliance therapy received higher 17 adherence ratings (p < 0.001) than those patients receiving non-specific 18 counselling. 19 O’Donnell (2003) (O'Donnell, Colin, Donohoe, Gary, Sharkey, Louise et al , 20 2003), a RCT from the Cochrane review compared “compliance therapy” with 21 non-specific counselling (as the control group) in patients with schizophrenia, 22 with negative results. The study included 94 patients aged between 18 and 65 23 years. The mean age for both of the groups was 32 years (SD=9). The 24 intervention lasted 5 sessions, each session lasting 30-60 minutes. It is 25 reported that, “…the sessions covered a review of the patient’s illness history, 26 understanding of the illness and his or her ambivalence to treatment, 27 maintenance medication and stigma. Compliance therapy is a cognitive 28 behavioural intervention with techniques adapted from motivational 29 interviewing, other cognitive therapies and psycho-education.” (Haynes, R. B., 30 Yao, X., Degani, A. et al , 2005). This study was conducted in Ireland. 31 There was no difference in terms of adherence between the two groups. Medicines concordance: full guideline DRAFT (July 2008) page 272 of 373 DRAFT FOR CONSULTATION 1 8.10.2.5 Multicomponent intervention 2 Pradier (2003) (Pradier, Christian, Bentz, Laurence, Spire, Bruno et al , 3 2003), a RCT from the Cochrane review, compared a combined educational 4 and counselling intervention with a control condition in patients with HIV with 5 positive results. The study included 244 patients aged >18 years. Median age 6 of the participants was 40 years in the intervention group and 38 years in the 7 control group. The intervention consisted of 3 individual sessions delivered by 8 nurses lasting 45-60 minutes. The intervention was, “…founded on the 9 principles of motivational psychology, client centred therapy and the use of an 10 ”empathic therapeutic to enhance participants’ self efficacy“. The intervention 11 focused on cognitive, emotional, social and behavioural determinants affecting 12 adherence.” (Haynes, R. B., Yao, X., Degani, A. et al , 2005). This study was 13 conducted in France. 14 Self-reported adherence between baseline and six months was significantly 15 improved in the intervention group, versus control. 16 Van Servellen (2005) (van, ServellenG, Nyamathi, A., Carpio, F. et al , 2005) 17 compared an enhanced adherence intervention with standard clinical care in 18 patients (n = 85) taking antiretroviral medications for at least 3 months with 19 negative results. Of note is that to be eligible to take part in this study 20 participants had to be able to speak Spanish. The enhanced adherence 21 intervention consisted of two parts the first being modular instruction which 22 was aimed at increasing patients HIV knowledge and ability to communicate 23 with medical staff and was delivered over 5 sessions by health educators and 24 nurses. These were followed up case management sessions, delivered either 25 face to face or via a telephone by a nurse, which concentrated on addressing 26 patient’s potential or actual risks for non adherence using motivational 27 interviewing techniques. Content involved going over things misunderstood in 28 the modular instruction stage, identifying barriers to adherence and finding 29 strategies to challenge these and helping to find community, treatment and 30 social support/referrals to help address adherence barriers. This study was 31 conducted in the USA. Medicines concordance: full guideline DRAFT (July 2008) page 273 of 373 DRAFT FOR CONSULTATION 1 There where no significant differences between the group at 6 months in: self 2 efficacy of adherence management (control group, -0.06, SD 0.59, 3 intervention group, 0.12, SD 0.95); 2 or more doses missed in last 4 days 4 (control group, 6.79%, intervention group, -5.69%); 2 or more doses missed 5 pasted 24 hours (control group, 18.21%, intervention group, -32%); average 6 doses missed in last 4 days (control group, 0.04, SD 0.13, intervention group, 7 0.02, SD, 0.14); proportion over 95% adherent in last four days (control group, 8 -4.85% ,intervention group, 1.71%); proportion > 90% adherent in last four 9 days (control group, -11.47% intervention group, -0.49%); follow medication 10 special instructions for 4 days (control group, 0.06, SD 0.34, intervention 11 group, -0.07, SD 0.36) and following medication schedule (control group, - 12 0.09, SD 1.60, intervention group, 0.33, SD 1.58). 13 8.10.2.6 14 Antoni (2006) (Antoni, M. H., Carrico, A. W., Duran, R. E. et al , 2006) 15 compared cognitive behavioural stress management (CSBM) in addition to 16 antiretroviral medication adherence training (MAT) with MAT alone in patients 17 with HIV with negative results. CSBM sessions included a didactic 18 component, as well as group discussion, with opportunities provided to apply 19 newly learned techniques. One hundred and thirty patients aged between 18 20 and 65 years were included in the study. The mean age was 41.6 (SD=8.3) 21 years. Homework was assigned to provide opportunities for participants to 22 practice techniques and increase their self-efficacy. The treatment was 23 focused extensively on eliciting participant experiences with adherence and 24 medication side effects. Throughout the 10-week, 135-minute group sessions 25 (90 minute stress management and 45 minute relaxation), facilitators 26 encouraged participants to examine potentially distorted cognitions and how 27 these may influence adherence to HAART (as well as other relevant self-care 28 behaviours). During cognitive restructuring exercises, participants were asked 29 to examine medication-relevant thoughts both in session and through 30 homework exercises. Adherence was also a key target during the skills 31 training sessions. This study appears to have been conducted in the USA (not 32 explicitly stated). Cognitive Behavioural Treatment Medicines concordance: full guideline DRAFT (July 2008) page 274 of 373 DRAFT FOR CONSULTATION 1 The experimental conditions did not differ significantly in participant-reported 2 medication adherence throughout the 15-month investigation period. 3 Bechdolf (2004 and 2005) (Bechdolf, A., Knost, B., Kuntermann, C. et al , 4 2004) (Bechdolf, A., Köhn, D., Knost, B. et al , 2005) compared group CBT 5 with group psycho-education (PE) in patients who had suffered an episode of 6 a schizophrenia or a related disorder with negative results. Eighty-eight 7 patients aged between 18 and 64 years were included in the study. Only a 8 minority of patients were employed. Group CBT focused on assessment and 9 engagement (sharing information about voices and delusions, models of 10 psychosis), improving self-esteem, formulation of key-problems and 11 developing interventions directed at reducing the severity and the occurrence 12 of key problems, relapse prevention/keeping well and enhancing medication 13 compliance. There was a specific focus on the component "improving self- 14 esteem" to foster feelings of hope and engagement with therapy. Group CBT 15 involved 16 sessions in 8 weeks by a psychiatrist or clinical psychologist. This 16 study was conducted in Germany. 17 Compliance was high initially (group CBT mean: 3.9 (SD 0.3) vs. PE group: 18 3.8 (SD 0.5)). At 8 weeks post-treatment, there was no significant difference 19 between groups (3.9 (SD 0.3) vs. 3.7 (SD 0.7)) nor at 6 months (3.5 (SD 0.9) 20 vs. 3.2 (SD 1.0). This remained non-significant when corrected for pre- 21 treatment scores. At 24 month follow-up again no significant differences were 22 seen (3.4 (SD 0.7) vs. 2.9 (SD 1.1). 23 Weber (2004) (Weber, R., Christen, L., Christen, S. et al , 2004), a RCT from 24 the Cochrane review, compared cognitive behavioural therapy in addition to 25 usual care to usual care alone in patients with HIV with negative results. The 26 study included 60 patients, and the median age was 41 years. The 27 intervention was delivered by a psychotherapist. The Cochrane Review 28 informs that, “…protocol defined a minimum of three and a maximum of 25 29 sessions within the 1- year study period. Participant and psychotherapist 30 determined the frequency of appointments and set their own goals for future 31 interventions. The intervention had to be based on concepts of cognitive Medicines concordance: full guideline DRAFT (July 2008) page 275 of 373 DRAFT FOR CONSULTATION 1 behavioral therapy” (Haynes, R. B., Yao, X., Degani, A. et al , 2005). This 2 study was conducted in Switzerland. 3 There was no significant difference in mean adherence between the two 4 groups, but both groups had very high mean adherence rates (92.8% versus 5 88.9%), and a higher proportion of intervention group patients were at or 6 above 95% adherence (70% versus 50%, p = 0.014). 7 There was no difference in terms of adherence between the two groups. 8 Lam (2003) (Lam, D. H., Watkins, E. R., Hayward, P. et al , 2003) compared 9 cognitive therapy and minimal psychiatric care v minimal psychiatric care 10 alone in patients with bi-polar disorder with positive results. One hundred and 11 three patients with ages ranging from 18 to 70 years were enrolled in the 12 study. Mean age was 46.4 ± (12.1) (s.d.) years for the intervention group and 13 41.5± (10.8) (s.d.) for the control group. Traditional cognitive therapy for 14 depression was provided by clinical psychologists with new elements 15 highlighting the need for combined psychological and drug treatment, CBT 16 skills for monitoring mood and preventing relapse and highlighting the 17 importance of sleep and routine. The therapy also addressed illness beliefs. 18 Cognitive therapy involved 12 to 18 individual sessions within the first 6 19 months and 2 booster sessions in the second 6 months. This study was 20 conducted in the UK. 21 At 14 months, medication adherence was 95.5% (SD: 7.7%) in the cognitive 22 therapy group and 69.1% (SD 16.4%) in the control group (p < 0.001). 23 Proportions of people who had at least 80% adherence rates were 97.4% in 24 the cognitive therapy group and 21.7% in the control group (p < 0.001). 25 Gray (2006) (Gray, R., Leese, M., Bindman, J. et al , 2006) compared 26 adherence therapy (AT) with health education (HE) (serving as the control 27 group) in patients with schizophrenia with negative results. Adherence therapy 28 is a brief, individual CBT approach. 6 elements formed the core of the therapy: 29 assessment, medication problem solving, medication timeline, exploring 30 ambivalence, discussing beliefs and concerns about medication and using 31 medication in the future. Three hundred patients were included in the study, Medicines concordance: full guideline DRAFT (July 2008) page 276 of 373 DRAFT FOR CONSULTATION 1 and the mean age (s.d.) was 41.5 (11.5) years. Key therapy skills that the 2 therapists use included exchanging information, developing discrepancies 3 between participant’s thoughts and behaviours about medications and 4 working with resistance to discussing psychiatric medication and treatment. 5 The overall aim of process was to achieve a joint decision about the 6 medication. Participants were offered a maximum of 8 sessions lasting 30-50 7 minutes over a 5 month period and the intervention was delivered by 9 8 therapists (four psychologists, three psychiatrists and 2 mental health nurses). 9 The study was conducted in 4 countries: The Netherlands, Germany, England 10 and Italy. 11 At 12 months, there were no significant differences between the groups using 12 either patient assessment (AT group: 3.20 (SD 1.07), HE group: 3.33 (SD 13 1.02), 95% CI -0.35 to 0.08) or clinical assessment (AT group: 5.22 (SD 1.57) 14 HE group: 5.03 (SD 1.55), 95% CI -0.12 to 0.52) of adherence. 15 Wagner (2006) (Wagner, G. J., Kanouse, D. E., Golinelli, D. et al , 1909) 16 compared a cognitive behavioural treatment with an enhanced condition of the 17 treatment (a 2 week pre-treatment practice trial) and a control group, for the 18 effect on adherence to a new regime of antiretroviral therapy. The study 19 included 230 patients with a mean age of 39 (ranging from 21 to 70 years), 20 80% were male, 49% were Latino(a) and 65% were unemployed. The study 21 was set in the USA. The intervention involved five sessions of cognitive 22 behavioural therapy. Questionnaires were administered and blood was drawn 23 at screening (four weeks before treatment baseline), and periodically up to 48 24 weeks from the start of treatment. There was no difference in adherence 25 between the intervention and the enhanced intervention group. There was 26 initially a significant increase in attaining ‘good’ adherence (90% of prescribed 27 dose) for the intervention groups compared to the control group (82% versus 28 65%, p=0.01). The difference reduced in the following weeks and was not 29 significant. At week 48 the difference was reversed to 57% (intervention 30 group) versus 65% (control group), but this was also non-significant (p=0.52). 31 Wyatt (2004) (Wyatt, G. E., Longshore, D., Chin, D. et al , 2004) compared a 32 cognitive behavioural approach (the Enhanced Sexual Health Intervention) to Medicines concordance: full guideline DRAFT (July 2008) page 277 of 373 DRAFT FOR CONSULTATION 1 usual care for risk reduction and treatment adherence for 147 women who 2 had HIV and a history of childhood sexual abuse. The mean age was 41 (8.2), 3 25-65 years, 51% were African American and 49% were Latina and primarily 4 unemployed. The study was set in the USA. The intervention involved 11 5 weekly sessions for 2.5 hours per week of psycho-educational content relating 6 to child sexual abuse and HIV status. They were followed up at the end of the 7 11 weeks and then again at 3 and 6 months. Although an effect was found for 8 risk reduction there was no increase in adherence to medication in the 9 intervention group (75.6% versus 73.3%, OR=1.13, p=0.41). However a 10 significant effect was found for adherence for those who attended at least 11 eight sessions (91.3%) compared to seven or fewer (49.7%), OR=4.90, 12 p=0.044. The difference in adherence of the high attendees was 91.3% 13 compared to the control group 74.7%, this was statistically significant. 14 15 8.11 Key Clinical Question: Would a contractual agreement between HCP and patient affect adherence? 16 Related Evidence statements Evidence into recommendations references (summary of evidence) Bosch- 1.Evidence from one high For general discussion of limitations of evidence Capblanch quality systematic review of see section 7.3 (2007) RCTS suggests that the (Bosch- use of contracts within a Capblanch, healthcare setting does not X., Abba, appear to increase K., Prictor, adherence to prescribed M. et al , medication. There is no evidence to show that contractual arrangements have any impact in improving adherence and the GDG did not wish to make a recommendation in this area. 2007) 17 18 *e.g. relative value place on different outcomes, trade-offs between benefits and harms, uncertainties in Evidence, likely impact if implements, feasibility, patient perspective. Medicines concordance: full guideline DRAFT (July 2008) page 278 of 373 DRAFT FOR CONSULTATION 1 8.11.1 Methods of the evidence review 2 This paper includes a narrative summary of the included evidence, structured 3 according to the category of the intervention, following the agreed reviewing 4 protocol: 5 Types of studies - Randomized controlled trials (RCTs) of interventions to 6 increase adherence. The excluded studies list from the Cochrane review was 7 checked as we have included those studies with less than 80% follow-up of 8 participants. 9 Types of participants - people prescribed medication for a medical condition. 10 Duration of studies - six months follow up from the time of patient entry for 11 long-term regimens. No time limit specified for short-term conditions. 12 Types of interventions - any interventions intended to change adherence to 13 prescribed medication. As the Cochrane review is presented by condition, we 14 have used the evidence extracted in that review and reconfigured it by 15 intervention. 16 Types of outcome measures - inclusion criteria (as defined in the Cochrane 17 review) were expanded by including studies that used adherence as the only 18 outcome variable as opposed to adherence and treatment outcome variables. 19 The excluded studies list of the Cochrane review was cross-referenced to 20 ensure that no potentially relevant study was missed out. 21 8.11.2 Definitions of contracts 22 Contracts can be generically viewed as reciprocal agreements between two or 23 more parties in which one or more will need to do something. 24 From a behavioural strategy perspective, a contract to increase patient’s 25 adherence can be defined as “a process of specifying a set of rules regarding 26 some behaviour of interest and formalizing a commitment to adhere to them” 27 (Haynes, T. B., Taylor, D. W., and Sackett, D. L., 1979). Medicines concordance: full guideline DRAFT (July 2008) page 279 of 373 DRAFT FOR CONSULTATION 1 Contracts can be written or verbal. Most contracts are between healthcare 2 practitioners and patients, but they may also occur between practitioners and 3 carers, carers and patients or by a patient with him/herself. 4 8.11.3 Evidence review 5 We retrieved one Cochrane Review that aimed to assess whether contracts 6 between healthcare practitioners and patients had an effect on patients' 7 adherence to treatment, prevention and health promotion activities (Bosch- 8 Capblanch, X., Abba, K., Prictor, M. et al , 2007). Although this review 9 included settings other than clinical settings, we decided to include this high 10 quality systematic review of RCTs on the grounds that the results were 11 reported in groups, thus allowing us to make conclusions from those settings 12 relevant to the guideline. It also included other treatment groups that are 13 outside the remit of the guideline, such as substance addiction treatments and 14 interventions for hypertension and overweight without prescribed medication. 15 Other areas that were included were acne, acute bacterial infections, arthritis, 16 asthma, breast self examination, contact lens care, depression, diabetes, 17 phobias, promotion of healthy diet and exercise and tuberculosis. 18 The Cochrane review also assessed the effects of contracts on other 19 outcomes, including patient participation and satisfaction, health practitioner 20 behaviour and views, health status, harms, costs, and ethical issues. 21 Seven trials assessed contracts between HCPs and patients, nine trials 22 assessed contracts between patients and carers, peers or others, and 23 between HCPs and carers in one trial. Four trials assessed contracts between 24 HCPs, patients and carers, two trials assessed self-contract and the other 25 seven trials did not report which type of contract was being used. Twenty one 26 trials included some type of financial incentive 27 Several of the trials were of poor quality and included small numbers of 28 people. Most were conducted in the USA and were conducted in specialised 29 services. Medicines concordance: full guideline DRAFT (July 2008) page 280 of 373 DRAFT FOR CONSULTATION 1 Two trials that examined the effects of contracts in the context of hypertension 2 management reported adherence outcomes. However, only one showed 3 statistically significant results in favour of the group with contracts. 4 In the miscellaneous section, six trials in the contexts of acne, acute bacterial 5 infection, asthma, depression, diabetes and tuberculosis reported adherence 6 outcomes. However, in some cases it was not possible to determine whether 7 adherence was also related to prescribed medication or with an overall 8 treatment regime (depression, and diabetes). From these, five trials did not 9 report any statistical significance in favour of the contracts groups, whilst the 10 acute bacterial infection trial reported significantly better results (based on pill 11 count) in the contract group. However, there was no difference between 12 groups in self-reported adherence, nor in the number of additional 13 prescriptions to finalise the treatment. 14 Based on the results for the trials that included an assessment of patients’ 15 adherence to medication, the use of contracts does not appear to improve 16 adherence. 17 Overall, the conclusions from the Cochrane authors state that there is limited 18 evidence that contracts can have a positive effect in improving adherence. In 19 addition they argue that there is insufficient evidence from large, good quality 20 studies to routinely recommend contracts for improving adherence to 21 treatment or preventive health regimens (Bosch, Capblanch, X and Garner, 22 P., 2006). 23 24 25 26 Medicines concordance: full guideline DRAFT (July 2008) page 281 of 373 DRAFT FOR CONSULTATION 1 2 8.12 Key Clinical Question: Do reminders (and what types of 3 reminders, text messaging etc) increase adherence to 4 prescribed medication? Related Evidence statements references (summary of Evidence into recommendations evidence) Heneghan (2006) 1. A high quality For general discussion of limitations of evidence (Heneghan, C. J., systematic review see section 7.3 Glasziou, P., and suggests that Perera, R., 2006) medication packaging that incorporates a reminder system may improve adherence for patients taking longterm medications. The available evidence suggests that reminders can be of value in improving adherence for some patients in some situations. A wide range of reminders were found from reminders related to packaging to telephone reminders. The GDG did not consider the evidence sufficient to make general recommendations about reminders such Stewart (2005) 2. Three RCTs show as telephone and text reminders but did consider (Stewart, A., that a reminder given that packaging may have a role in reminding Noakes, T., via a telephone call in patients. This overlaps recommendation Eales, C. et al , a multicomponent regarding drug packaging. 2005); intervention can Urien(2004) increase adherence to (Urien, A. M., prescribed medication. Guillen, V. F., Beltran, D. O. et al , 2004); Piette (2000) (Piette, J. D., Weinberger, M., McPhee, S. J. et al , 2000) Medicines concordance: full guideline DRAFT (July 2008) page 282 of 373 DRAFT FOR CONSULTATION Beaucage (2006) 3. One RCT showed (Beaucage, K., no significant results Lachance- on the effect of a Demers, H., Ngo, reminder given via a T. T. et al , 2006) telephone call in a multicomponent intervention in increasing adherence to prescribed medication Hamet 4. There is conflicting (2003)(Hamet, P., evidence on the effect Campbell, N., of reminders via mail in Curnew, G et al , a multicomponent 2003); Peterson intervention in (1984) (Peterson, increasing adherence G. M., to prescribed Fitzmaurice, K. medication. D., Naunton, M. et al , 2004) Vrijens (2006) 5. There is conflicting (Vrijens, B., evidence on the effect Belmans, A., of electronic reminders Matthys, K. et al , in a multicomponent 2006); intervention in Mannheimer increasing adherence (2006) to prescribed (Mannheimer, S. medication. ( B., Morse, E., Matts, J. P. et al , Medicines concordance: full guideline DRAFT (July 2008) page 283 of 373 DRAFT FOR CONSULTATION 2006); Sackett (1975) (Sackett, D. L., Gibson, E. S., and Taylor, D. W., 1975) Guthrie (2001) 6. One RCT showed (Guthrie, R. M., that using a 2001) combination of postal and telephone reminders in a multicomponent intervention showed no effect on adherence to prescribed medication. 1 2 *e.g. relative value place on different outcomes, trade-offs between benefits and harms, uncertainties in Evidence, likely impact if implements, feasibility, patient perspective Medicines concordance: full guideline DRAFT (July 2008) page 284 of 373 DRAFT FOR CONSULTATION 1 8.12.1 Methods of the evidence review 2 This paper includes a narrative summary of the included evidence, structured 3 according to the category of the intervention, following the agreed reviewing 4 protocol: 5 Types of studies - randomized controlled trials (RCTs) of interventions to 6 increase adherence. The excluded studies list from the Cochrane review was 7 checked as we have included those studies with less than 80% follow-up of 8 participants. 9 Types of participants - people prescribed medication for a medical condition. 10 Duration of studies - six months follow up from the time of patient entry for 11 long-term regimens. No time limit specified for short-term conditions. 12 Types of interventions - any interventions intended to change adherence to 13 prescribed medication. As the Cochrane review is presented by condition, we 14 have used the evidence extracted in that review and reconfigured it by 15 intervention. 16 Types of outcome measures – inclusion criteria (as defined in the Cochrane 17 review) were expanded by including studies that used adherence as the only 18 outcome variable as opposed to adherence and treatment outcome variables. 19 The excluded studies list of the Cochrane review was cross-referenced to 20 ensure that no potentially relevant study was missed out. 21 8.12.2 Evidence review 22 One high quality systematic review by Heneghan (2006) (Heneghan, C. J., 23 Glasziou, P., and Perera, R., 2006) aimed to determine the effects of reminder 24 packaging to increase patient adherence with self-administered long-term 25 medications. 26 The systematic review included eight studies containing 1137 participants. All 27 types of setting were included and no age limits were set. Studies where 28 direct observation of therapy occurred through a health professional were 29 excluded. Interventions that were included required a reminder system for the Medicines concordance: full guideline DRAFT (July 2008) page 285 of 373 DRAFT FOR CONSULTATION 1 day of the week or the time that the medication was to be taken, and it had to 2 form part of the packaging. Reminders that were separate to the intervention 3 (e.g. mailed reminders) were excluded. Electronic systems were also 4 excluded. Packaging aids were included irrespective of whether the 5 medication required a prescription or not. 6 The primary outcome of importance was adherence to medication which was 7 measured by pill counts and/or self-reporting. 8 Reminder packaging showed a significant increase in the percentage of pills 9 taken, weighted mean difference 11% (95% confidence interval (CI) 6% to 10 17%). Thus, the authors conclude that reminder packaging may represent a 11 simple method for improving adherence for patients in certain conditions. 12 8.12.2.1 13 Stewart (2005) (Stewart, A., Noakes, T., Eales, C. et al , 2005) compared four 14 once monthly educational sessions, the prescription of a home based walking 15 program and once monthly phone calls with four once monthly educational 16 sessions, the prescription of a home based walking program without once 17 monthly phone calls in patients attending a hypertension clinic with positive 18 results at week 24 but not 36. Sample size was comprised 83 participants. 19 During the phone calls patients (or a family member) were asked about the 20 exercise program and reminded about diet and medication. In total 5 (pairs) of 21 telephone calls (to patient and family member) were made once monthly over 22 24 weeks by a physiotherapist. This study was conducted in South Africa. 23 At week 24 significantly more patients in the group receiving telephone calls 24 (65%) were taking their medications as prescribed compared to the group not 25 receiving telephone calls (44.7%, p = 0.05), however, there was no difference 26 between the groups at week 36 (82.4% vs 86.7%) (Stewart, A., Noakes, T., 27 Eales, C. et al , 2005). 28 Urien (2004) (Urien, A. M., Guillen, V. F., Beltran, D. O. et al , 2004) 29 compared a telephone-delivered intervention plus educational advice with 30 educational advice alone in patients receiving antibiotic treatment. The sample Telephone reminders Medicines concordance: full guideline DRAFT (July 2008) page 286 of 373 DRAFT FOR CONSULTATION 1 was comprised of 128 participants aged ≥18 years. The telephone call was 2 undertaken on the 4th day after the start of treatment, when the first box of 3 antibiotics should have been finished. The patient was advised to continue the 4 treatment according to the dosage and number of days that had been 5 prescribed. The patient was also reminded that although he or she may feel 6 better or even cured, the treatment was to be continued for 10 days. This 7 study was conducted in Spain. 8 Adherence was significantly higher in the intervention group (78.3%) than in 9 the control group (54.1%) (p = 0.005) (Urien, A. M., Guillen, V. F., Beltran, D. 10 O. et al , 2004). 11 Beaucage (2006) (Beaucage, K., Lachance-Demers, H., Ngo, T. T. et al , 12 2006) compared a pharmacist telephone intervention with usual care in 13 patients on antibiotic treatment with negative results. A telephone call was 14 made to patients in the intervention group by a pharmacist 3 days into 15 treatment. The pharmacist asked about the patient’s general condition, on the 16 presence of adverse effects and the participants understanding of dosing. The 17 pharmacist emphasized the importance of adherence and offered motivation 18 to the patient. The patients were offered an opportunity to ask questions and 19 were given the pharmacists contact details in case they wanted to make 20 contacted their pharmacist at a later time. This study was conducted in 21 Canada. 22 23 Mean adherence to treatment was 94% +/- 9% and 94% +/- 12% in the 24 intervention and control groups respectively (p = 0.803). The proportion of 25 patients with less than 80% adherence was similar in the two groups 26 (Intervention group: 8%, control group: 9%). 27 Piette (2000) (Piette, J. D., Weinberger, M., McPhee, S. J. et al , 2000) 28 compared a telephone intervention with usual care in patients with diabetes (n 29 = 280) with positive results. Patients were excluded if they were above 75 30 years of age. The intervention consisted of, “…a series of automated 31 telephone assessments designed to identify patients with health and self-care Medicines concordance: full guideline DRAFT (July 2008) page 287 of 373 DRAFT FOR CONSULTATION 1 problems (TeleminderModel IV automated telephone messaging computer). 2 Calls were made on a biweekly basis, up to 6 attempted calls, and involved a 3 5 to 8-minute assessment. During each assessment, patients used the touch- 4 tone keypad to report information about self-monitored blood glucose 5 readings, self-care, perceived glycaemic control, and symptoms of poor 6 glycaemic control, foot problems, chest pain, and breathing problems, with 7 automated prompts for out-of-range errors. The automated telephone calls 8 were also used to deliver, at the patient’s option, 1 of 30 targeted and tailored 9 self-care education messages at the end of each telephone session. Patients 10 only received a 1-page instruction sheet on the use of the phone. Each week, 11 the automated assessment system generated reports organized according to 12 the urgency of the reported problems, and a diabetes nurse educator used 13 these reports to prioritize contacts for a telephone follow-up. During follow-up 14 calls, the nurse addressed problems reported during the assessments and 15 provided more general self-care information. After several months, 16 intervention group patients were offered additional automated self-care calls 17 that focused on glucose self-monitoring, foot care and medication adherence. 18 In the medication adherence part of these sessions, patients were asked 19 about their adherence to insulin, oral hypoglycaemic medications, 20 antihypertensive medications, and antilipidemic medications. For each type of 21 medication, patients without adherence problems received positive feedback 22 and reinforcement. Patients reporting less than optimal adherence were asked 23 about specific barriers and were given advice from the nurse about 24 overcoming each barrier. The nurse was located outside the clinic and had no 25 access to medical records other than the baseline info collected at enrolment 26 and her own notes. She did not have any face-to-face contact with patients. 27 The nurse addressed problems raised by patients in the automated calls and 28 also gave general self-care education. The nurse also checked on patients 29 who rarely responded to automated calls. A small number of patients initiated 30 calls to the nurse by toll free no. She referred these to the primary care 31 physician as appropriate.” (Haynes, R. B., Yao, X., Degani, A. et al , 2005). 32 This study was conducted in the United States. Medicines concordance: full guideline DRAFT (July 2008) page 288 of 373 DRAFT FOR CONSULTATION 1 Compared with usual care, patients in the intervention group reported fewer 2 problems with medication adherence (P < 0.003). 3 8.12.2.2 4 Hamet (2003) (Hamet, P., Campbell, N., Curnew, G et al , 2003) compared 5 the avapromise intervention (designed to modify behaviour by medication 6 adherence through reinforcement and lifestyle modification) with usual care in 7 patients with high blood pressure. This was a study that comprised a total of 8 4864 participants. The ages of participants ranged from 16 to 89 years. The 9 intervention was made up of two elements that were delivered together. The Mail reminders 10 first element attempted to reinforce medication behaviours by using 11 medication reminder letters, blood pressure diaries and telephone nurse 12 counselling sessions. The second element addressed lifestyle management 13 through educational brochures. Patients assigned to the intervention group 14 were mailed the material at one, two, three, four, six and 12 months. This 15 study was conducted in Canada. 16 A total of 25% of patients discontinued their treatment from the intervention 17 group and 25.5% from the control group (p = 0.94). There was no statistically 18 significant difference in the duration of irbesartan compliance between the 19 treatment groups (Hamet, P., Campbell, N., Curnew, G et al , 2003). 20 In Peterson’s (1984) (Peterson, G. M., Fitzmaurice, K. D., Naunton, M. et 21 al , 2004) (from the Cochrane review) RCT, the intervention group received 22 several adherence-improving strategies: patients were counselled on the 23 goals of anticonvulsant therapy and the importance of good adherence in 24 achieving these goals; a schedule of medication-taking was devised that 25 corresponded with the patient's everyday habits; patients were given a copy of 26 an educational leaflet; each patient was provided with a 'dosette' medication 27 container and counselled on its utility; patients were instructed to use a 28 medication/seizure diary; and patients were reminded by mail of upcoming 29 appointments and of missed prescription refills. The control group received 30 none of these interventions. Patient compliance improved significantly with the 31 intervention group patients. Fifty three adults and teenagers were enrolled in Medicines concordance: full guideline DRAFT (July 2008) page 289 of 373 DRAFT FOR CONSULTATION 1 the study, and the median age was 35 years for the control group and 28 2 years for the intervention group. The study was conducted in Australia. 3 8.12.2.3 4 Guthrie (2001) (Guthrie, R. M., 2001) delivered an intervention involving 5 postal and telephone reminders regarding coronary risk reduction and 6 medication compliance, which were sent during the first 2 months of 7 pravastatin treatment, or usual care. This was a large study that comprised a 8 total of 13,100 participants. Mean age was 58.0 years. Both groups received 9 reminder postcards at 4 and 5 months, in addition to counselling by physicians Mail and telephone reminders 10 about coronary risk reduction. At study discontinuation, patients completed 11 and mailed (to the program-coordinating centre) questionnaires concerning 12 compliance with care, as well as self-reported adoption of other lifestyle 13 modifications. This study was conducted in the USA. Neither early reminders 14 nor baseline patient characteristics were significantly associated with reported 15 pravastatin compliance rates. 16 8.12.2.4 17 Vrijens (2006) (Vrijens, B., Belmans, A., Matthys, K. et al , 2006) compared a 18 supportive intervention program with usual care in patients who had been 19 taking atorvastatin with positive results. Four hundred and twenty nine 20 participants aged >18 years entered the study. As part of the supportive 21 intervention program participants were supplied with a ‘beep-card’ that 22 reminded the patient of the dosing time, and also gave educational reminders. 23 The supportive intervention program also provided patients with a medication 24 review by the patients’ pharmacist of their electronically compiled dosing 25 history (through MEMs). At each follow-up visit the pharmacist delivered an 26 educational message, updated the patients‘ compliance passport and 27 analysed with the patient their electronically compiled dosing history over the 28 previous month/3 months (depending on the gap between follow-up 29 appointments). Baseline adherence was significantly higher in the intervention 30 group compared to the control group (96.43% vs. 94.33%, p = 0.003). At 12 31 months, the intervention group had an increased adherence of 6.5% 32 compared to the control group (95.89% vs. 89.37%, p < 0.001). The analyses Electronic reminders Medicines concordance: full guideline DRAFT (July 2008) page 290 of 373 DRAFT FOR CONSULTATION 1 were stratified by baseline compliance and language region. Over time, the 2 difference between groups increased, with approximately 17% difference in 3 adherence between groups at 300 days. 13% (n = 25) of the intervention 4 group discontinued medication before 300 days, compared to 25% (n = 51) in 5 the control group. Persistence was significantly higher in the intervention 6 group compared to the control group (87% vs. 74%, p = 0.002). This remained 7 significant when adjusted for multiple baseline variables. This study was 8 conducted in Belgium (Vrijens, B., Belmans, A., Matthys, K. et al , 2006). 9 In a study conducted by Sackett (1975) (Sackett, D. L., Gibson, E. S., and 10 Taylor, D. W., 1975) (from the Cochrane Review), subjects in one of the 11 interventions (augmented convenience) groups visited company physicians, 12 rather than their family physicians, for hypertensive and follow-up care during 13 paid working hours. Two hundred and thirty Canadian steelworkers were 14 enrolled. The ages of the participants were not reported in the study. The 15 second intervention, mastery learning, aimed to give the facts about 16 hypertension, its effects upon target organs, health, and life expectancy, the 17 benefits of antihypertensive therapy, the need for adherence with medications 18 and some simple reminders for taking pills (which was provided in a slide-tape 19 format, and reinforced by a secondary-school graduate). No statistically 20 significant results were reported. This study was conducted in Canada. 21 Mannheimer (2006) (Mannheimer, S. B., Morse, E., Matts, J. P. et al , 2006) 22 conducted cluster randomised controlled trial (2x2 factorial design) to assess 23 two interventions to increase adherence in 928 patients who were taking Anti- 24 retroviral Therapy in the USA. One intervention is the Medication Manager 25 (MM) which involved a trained research staff member working with the 26 participants individually to provide tailored adherence support according to a 27 standardised protocol, identifying and addressing information, motivation and 28 skills for antiretroviral adherence (using detailed questionnaires). The second 29 intervention was the electronic medication reminder system, a small portable 30 alarm (ALR) which was programmed to flash and sound when antiretroviral 31 doses were due. Participants were followed up with assessments at 1 and 4 32 months after randomisation and 4 months thereafter. The MM group had Medicines concordance: full guideline DRAFT (July 2008) page 291 of 373 DRAFT FOR CONSULTATION 1 significantly higher reporting of 100% adherence over time compared to non- 2 MM interventions (OR=1.42, p<0.001). There were no significant differences 3 between the ALR group and the non-ALR groups for adherence. 4 8.12.2.5 5 De Geest (2006) (De, Geest, S, Schafer-Keller, P., Denhaerynck, K. et al , 6 2003) compared a nurse-led intervention and usual care with usual care alone 7 in patients who had undergone a kidney transplant with negative results. This 8 was a small study that comprised a total of 18 participants aged > 18 years. 9 The intervention group received one home visit and three telephone Type of reminders used not stated 10 interviews, one at the end of the month for three consecutive months. During 11 the home visit printouts were discussed with patient for problem detection, 12 and adherence goals were made. All patients received self-efficacy 13 interventions. Nurses also implemented additional educational, behavioural 14 (e.g. the use of reminders) and/or social support interventions if they felt this 15 would help the patient. Telephone calls served to discuss adherence in 16 previous month, to check on health status, and discuss and change, if 17 appropriate, adherence interventions. This study was conducted Switzerland. 18 Adherence increased in both groups over the first 3 months (p = 0.04). The 19 overall difference between groups was non-significant at 3 months (p = 0.31) 20 and at 9 months (p = 0.58) with a gradual decline over the total 9 months to a 21 level still higher than initial levels (De, GeestS, Schafer-Keller, P., 22 Denhaerynck, K. et al , 2003). 23 24 Medicines concordance: full guideline DRAFT (July 2008) page 292 of 373 DRAFT FOR CONSULTATION 1 2 8.13 Key Clinical Question : Does being involved in self- 3 monitoring (e.g. of own blood pressure) increase 4 adherence to prescribed medication? Related Evidence statements references (summary of Evidence into recommendations evidence) Haynes 1976 1. The majority of For general discussion of limitations of evidence se (Haynes, R. B., evidence suggests that section 7.3 Sackett, D. L., involving patients in the and Gibson, E. self-monitoring of their S., 1976); Sadik medication adherence 2005 (Sadik, A., (e.g. through recording Yousif, M., and adherence in diary McElnay, J. C., logs) appears to 2005); Tsuyuki increase adherence as 2004 (Tsuyuki, part of a multi R. T., Fradette, component M., Johnson, J. intervention. The evidence of the value of self-monitoring in increasing adherence was conflicting. The GDG considered that this was perhaps not surprising given that qualitative evidence had indicated that patients may use such measures to inform their ow decisions and evaluations of treatments rather tha to ensure they follow a previous decision. Self monitoring is also used in conditions where the patient can alter treatment according to results and this group will have different characteristics and A. et al , 2004); needs than a general patient group. Sadik 2005 2. There is conflicting (Sadik, A., evidence in regard to Yousif, M., and whether having McElnay, J. C., patients record 2005); Bailey information (e.g. in a 1990 (Bailey, W. diary log) relevant to C., Richards-Jr, their condition (e.g. J. M., Brooks, C. symptoms) can M. et al , 1990); increase adherence as Medicines concordance: full guideline DRAFT (July 2008) page 293 of 373 DRAFT FOR CONSULTATION Cote 1997 (Coté, part of a multi J., Cartier, A., component Robichaud, P. et intervention. al , 1997); Cote 2001(Côté, J., Bowie, D. M., Robichaud, P. et al , 2001); Friedman 1996 (Friedman, R. H., Kazis, L. E., Jette, A. et al , 1996); Haynes 1976 (Haynes, R. B., Sackett, D. L., and Gibson, E. S., 1976); Johnson 1978 (Johnson, A. L., Taylor, D. W., and Sackett, D. L., 1978); Morice 2001(Morice, A. H. and Wrench, C., 2001); Peterson (1984) (Peterson, G. M., Fitzmaurice, K. D., Naunton, M. et al , 2004) Haynes 1976 3. There is conflicting (Haynes, R. B., evidence with regard to Medicines concordance: full guideline DRAFT (July 2008) page 294 of 373 DRAFT FOR CONSULTATION Sackett, D. L., whether providing and Gibson, E. participants with S., 1976); Cote information on how to 1997 (Coté, J., adjust their treatment Cartier, A., based on their own Robichaud, P. et self-monitoring effects al , 1997); adherence. Morice 2001 (Morice, A. H. and Wrench, C., 2001); Bailey 1990 (Bailey, W. C., Richards-Jr, J. M., Brooks, C. M. et al , 1990); Bailey (1999) (Bailey, W. C., Richards-Jr, J. M., Brooks, C. M. et al , 1990); ; Cote, 2001 (Cote, J., Bowie, D. M., Robichaud, P. et al , 2001). Morice 2001 4. There is conflicting (Morice, A. H. evidence that involving and Wrench, C., patients more in their 2001); Cote care through the self- 1997 (Coté, J., monitoring of Cartier, A., respiratory function Robichaud, P. et through measurement Medicines concordance: full guideline DRAFT (July 2008) page 295 of 373 DRAFT FOR CONSULTATION al , 1997); Bailey of PEF increases 1990 (Bailey, W. adherence as part of a C., Richards-Jr, multi component J. M., Brooks, C. intervention. M. et al , 1990); Cote 2001(Cote, J., Bowie, D. M., Robichaud, P. et al , 2001); Levy (2000) (Levy, M. L., Robb, M, Allen, J et al , 2000). Friedman 1996 5. There is conflicting (Friedman, R. H., evidence with regard to Kazis, L. E., whether involving Jette, A. et al , patients in the self- 1996); Haynes monitoring of their 1979 (Haynes, blood pressure R. B., Sackett, D. improves adherence as L., and Gibson, part of a multi E. S., 1976); component Johnson 1978 intervention. (Johnson, A. L., Taylor, D. W., and Sackett, D. L., 1978); MarquezContreras (2006) (MarquezContreras, E., Medicines concordance: full guideline DRAFT (July 2008) page 296 of 373 DRAFT FOR CONSULTATION Martell-Claros, N., Gil-Guillen, V. et al , 2006); Rudd (2004 (Rudd, P., Miller, N. H., Kaufman, J. et al , 2004) 1 2 *e.g. relative value place on different outcomes, trade-offs between benefits and harms, uncertainties in Evidence, likely impact if implements, feasibility, patient perspective 3 4 8.13.1 Methods of the evidence review 5 This paper includes a narrative summary of the included evidence, structured 6 according to the category of the intervention, following the agreed reviewing 7 protocol: 8 Types of studies - randomized controlled trials (RCTs) of interventions to 9 increase adherence. The excluded studies list from the Cochrane review was 10 checked as we have included those studies with less than 80% follow-up of 11 participants. 12 Types of participants - people prescribed medication for a medical condition. 13 Duration of studies - six months follow up from the time of patient entry for 14 long-term regimens. No time limit specified for short-term conditions. 15 Types of interventions - any interventions intended to change adherence to 16 prescribed medication. As the Cochrane review is presented by condition, we 17 have used the evidence extracted in that review and reconfigured it by 18 intervention. 19 Types of outcome measures – inclusion criteria (as defined in the Cochrane 20 review) were expanded by including studies that used adherence as the only 21 outcome variable as opposed to adherence and treatment outcome variables. Medicines concordance: full guideline DRAFT (July 2008) page 297 of 373 DRAFT FOR CONSULTATION 1 The excluded studies list of the Cochrane review was cross-referenced to 2 ensure that no potentially relevant study was missed out. 3 8.13.2 Evidence review 4 Although a handful of studies were found which addressed self monitoring we 5 did not find any study where the study interventions sole purpose was to 6 address self monitoring (although arguably a few appear in the Cochrane 7 Review studies). Within the studies which did address self monitoring there 8 was variability in how they did this and to what extent self monitoring was a 9 focus of the intervention. Narratives of each study, which give full details of 10 the entire intervention used in each individual study, are summarised below. 11 Tsuyuki (2004) (Tsuyuki, R. T., Fradette, M., Johnson, J. A. et al , 2004) 12 compared a patient support program with usual care in patients with heart 13 failure with negative results. Seven hundred and sixty six patients aged above 14 18 years were enrolled in the study. The mean age of the patients was 74 15 years. Patients in the support group received educational material consisting 16 of information about heart failure, non-drug treatment, medication information 17 (with special emphasis on proven benefits of therapies) and self-monitoring, 18 all written at a grade 8 reading level. Patients also received adherence aids 19 including a medication organizer, medication administration schedule, and 20 daily weight log. Community follow-up in the patient support program 21 consisted of telephone contact by the local research coordinator at 2 and 4 22 weeks and then monthly there after for 6 months after discharge The 23 telephone contact was to reinforce education and adherence relating to heart 24 failure and other self-care activities, focusing on the 5 essential components: 25 salt and fluid restriction, daily weighting, exercise alternating with rest periods, 26 proper medication use and knowing when to call their physician. This study 27 was conducted in Canada. ACE inhibitor adherence over 6 months after 28 discharge was 86.2 ± 29% in the usual care group vs. 83.5 ± 29% in the 29 patient support group (p = NS). 30 Sadik (2005) (Sadik, A., Yousif, M., and McElnay, J. C., 2005) compared a 31 pharmacist delivered intervention with usual care in patients with heart failure Medicines concordance: full guideline DRAFT (July 2008) page 298 of 373 DRAFT FOR CONSULTATION 1 with positive results. A total of 221 patients were enrolled in the study. Mean 2 age of participants was 58 years. For the intervention group patients the 3 research pharmacist discussed with their physicians if rationalization of drug 4 therapy or simplification of dosage regimens were considered appropriate. 5 Intervention patients were also educated (in a structured fashion) on HF, their 6 prescribed medication and the management of HF symptoms by the research 7 pharmacist. A printed booklet developed for this type of education programme 8 was used and each patient was given a copy to take home. The booklet 9 contained information on HF, its symptoms, the aims of treatment, the types of 10 medication used and their possible side-effects, diet and lifestyle changes, 11 advice to stick to one brand of digoxin (it having a narrow therapeutic index) 12 and information on the action to take if doses of medication were missed. 13 Intervention group patients were also instructed on a self-monitoring 14 programme (signs and symptoms of HF; compliance with prescribed 15 medication) in which they were asked to become involved; a monitoring diary 16 card (covering 1 month) was used. Patients were asked to complete their 17 monitoring diary cards at home and to show them to their physicians when 18 attending an appointment. The patients were asked to return their completed 19 diary cards to the research pharmacist for review when they visited the 20 hospital to receive medication refills. Reinforcement of the educational 21 message was carried out by the pharmacist as deemed necessary. This study 22 was conducted in the United Arab Emirates. 23 The number of intervention group patients vs. control patients who exhibited 24 self-reported compliance with the prescribed medicines (85 vs. 35) and 25 lifestyle adjustment (75 vs. 29) was higher than in control group patients at 12 26 months (p < 0.05). 27 Bailey (1990) (Bailey, W. C., Richards-Jr, J. M., Brooks, C. M. et al , 1990), a 28 study from the Cochrane review, compared a multi-faceted intervention with 29 standard care in patients with recurrent episodes of wheezing with positive 30 results. A total of 267 patients aged above 18 years were enrolled in the 31 study. Patients in the intervention group were, “…provided with the 32 standardized asthma pamphlets and were provided with a skill-oriented selfMedicines concordance: full guideline DRAFT (July 2008) page 299 of 373 DRAFT FOR CONSULTATION 1 help workbook, a one-to-one counselling session, and were subject to several 2 adherence-enhancing strategies, such as attending an asthma support group 3 and receiving telephone calls from a health educator. Physicians emphasized 4 these skills at regular clinic visits.” (Haynes, R. B., Yao, X., Degani, A. et al , 5 2005). The Cochrane review also states that the intervention included 6 involving patients more in their care through self-monitoring of their respiratory 7 function. This study was conducted in the USA. The intervention group were 8 significantly more adherent than the control group. 9 Bailey (1999) (Bailey, W. C., Richards-Jr, J. M., Brooks, C. M. et al , 1990), a 10 study from the Cochrane review, conducted a randomised controlled trial of 11 asthma self-management. 236 patients stratified by moderate or severe 12 asthma in Alabama, USA, were randomised to the University of Alabama 13 Birmingham (UAB) asthma self-management group, the UAB core-elements 14 group and usual care group. The core components involved a skill-oriented 15 self-help asthma workbook, which the patients were counselled in for one 16 hour. The participants also received 2 telephone calls and a letter at 1, 2 and 17 4 weeks to discuss problems and peak flow readings (see Bailey 1990). The 18 core elements program involved a shorter version of the workbook of which 19 counselling was given briefly (15-20 minutes). They were trained to use 20 inhalers and peak flow meters. Participants were followed up by telephone a 21 week later and a letter 2 weeks later. The usual care group received usual 22 education from their GP and informational leaflets. There were no significant 23 differences between groups in adherence. 24 Cote (1997) (Coté, J., Cartier, A., Robichaud, P. et al , 1997), a study from the 25 Cochrane review, compared two intervention groups and one control group in 26 patients with moderate to severe asthma with negative results. A total of 188 27 patients aged above 16 years were enrolled in the study. The intervention was 28 an “…asthma education program with an action plan based on peak-flow 29 monitoring (Group P) or an action plan based on asthma symptoms (Group 30 S). The Control group (Group C) received instructions from their 31 pulmonologists regarding medication use and influence of allergenic and non- 32 allergenic triggers. They were taught how to use their inhaler properly by the Medicines concordance: full guideline DRAFT (July 2008) page 300 of 373 DRAFT FOR CONSULTATION 1 educator. A verbal action plan could be given by the physician. Groups P and 2 S received the same education as the Controls plus individual counselling with 3 the specialised educator during a 1-hour session. All participants received a 4 book titled ”Understand and Control Your Asthma“ at no extra charge. Group 5 P received a self-management plan based on peak expiratory flow (PEF). 6 They were asked to continue measuring PEF twice a day and to keep a diary 7 of the results. Each time, subjects only recorded the best of three 8 measurements. Every attempt was made to ensure that patients knew how to 9 interpret the measurement and how to respond to a change in PEF. At each 10 follow-up visit, the patient’s diary card was reviewed, and if the action plan 11 had not been implemented when required, further explanations were given 12 regarding when treatment should be modified. Group S received a self- 13 management plan based on asthma symptom monitoring. These patients 14 were asked to keep a daily diary of asthma symptom scores, using a scale of 15 0 (no symptoms) to 3 (night time asthma symptoms, severe daily symptoms 16 preventing usual activities), and adjust their medications according to the 17 severity of respiratory symptoms using the guidelines of the action plan.” 18 (Haynes, R. B., Yao, X., Degani, A. et al , 2005). This study was conducted in 19 Canada. Neither intervention had a significant effect on participants’ 20 adherence. 21 Cote (2001) (Cote, J., Bowie, D. M., Robichaud, P. et al , 2001), a study from 22 the Cochrane review, compared two different educational interventions with 23 usual care for adult patients consulting with an acute asthma exacerbation 24 with negative results. A total of 126 patients aged above 18 years entered the 25 study. Patients in the Group Limited Education (LE) group were given a self- 26 action plan that was explained by the on-call physician. The action plan used 27 ”traffic lights“ (green, yellow, red) to describe specific states of asthma control 28 based on Peak Expiratory Flow and symptoms and actions that the patient 29 should take for each state. Patients in a ”Structured Educational group (SE)“, 30 in addition to what patients in Group LE received, participated in a structured 31 asthma educational program based on the PRECEDE model of health 32 education. This model took into consideration three different issues that were 33 important when dealing with health-related behaviours: predisposing factors Medicines concordance: full guideline DRAFT (July 2008) page 301 of 373 DRAFT FOR CONSULTATION 1 (belief, attitude, knowledge), enabling factors (community resource, family 2 support), and reinforcement…Reinforcement was provided at the 6-month 3 follow-up visit…the teaching was provided individually or in small groups 4 according to patient preference…Both intervention groups also received usual 5 care.” (Haynes, R. B., Yao, X., Degani, A. et al , 2005). This study was 6 conducted in Canada. 7 Neither intervention had a significant effect on participants’ adherence. 8 Levy (2000) (Levy, M. L., Robb, M, Allen, J et al , 2000) compared a nurse 9 delivered intervention with usual care in patients with asthma (n = 211) with 10 positive results. The intervention group were, “…invited to attend a 1h 11 consultation with one of the nurses beginning 2 weeks after entry to the study, 12 followed by two or more lasting half an hour, at 6-weekly intervals. The 13 second and third could be substituted by a telephone call. Patients were 14 phoned, by the nurse before each appointment in order to improve attendance 15 rates. Patient’s asthma control and management were assessed followed by 16 education on recognition and self-treatment of episodes of asthma. The 17 patients were taught to step-up medication when they recognized uncontrolled 18 asthma using PEF or symptoms. The advice was in accordance with national 19 guideline. Prescriptions were obtained from one of the doctors in the clinic or 20 by providing the patient with a letter to their general practitioner. Patients 21 presenting with severe asthma (severe symptoms of PEF below 60% of their 22 best/normal)were referred immediately to the consultant.” (Haynes, R. B., 23 Yao, X., Degani, A. et al , 2005). This study was conducted in the UK 24 Self-reported compliance was significantly higher in the intervention group for 25 use of inhaled topical steroids and rescue medication for severe asthmatic 26 attacks, but there was no significant difference between the groups for use of 27 these medications for mild attacks. 28 Friedman (1996) (Friedman, R. H., Kazis, L. E., Jette, A. et al , 1996), a study 29 from the Cochrane review, compared a telephone-linked computer system 30 (TLC) intervention for monitoring and counselling patients with usual care in 31 patients with hypertension with positive results. Two hundred and sixty seven Medicines concordance: full guideline DRAFT (July 2008) page 302 of 373 DRAFT FOR CONSULTATION 1 patients aged ≥ 60 years were recruited for the study. The mean age was 76 2 years for the TLC group and 77 years for the usual care group. TLC is, “…an 3 interactive computer-based telecommunications system that converses with 4 patients in their homes, using computer-controlled speech, between office 5 visits to their physicians. The intervention patients would call the TLC on a 6 weekly basis. Before calling, subjects would record their own blood pressure 7 using an automated sphygmomanometer with a digital readout. During the 8 conversation, subjects would answer a standard series of questions and the 9 TLC would provide education and motivational counselling to improve 10 medication adherence. The TLC then transmitted the reported information to 11 the subject’s physician”. (Haynes, R. B., Yao, X., Degani, A. et al , 2005). This 12 study appears to have been conducted in the USA. 13 The unadjusted results did not demonstrate significant improvement in 14 compliance or clinical outcome in patients using TLC as compared to those 15 patients receiving usual care. However, when the data were adjusted for age, 16 sex, and baseline adherence, the patients using TLC demonstrated a greater 17 improvement in medication adherence than those receiving usual care (p < 18 0.05). Sub-group analysis showed, in people who were non-adherent at 19 baseline, patients using TLC had greater improvement in medication 20 compliance (p < 0.05) than those receiving usual care. In people who were 21 adherent at baseline, TLC showed no significant difference in adherence 22 between the two groups over the course of the trial. 23 Haynes (1976) (Haynes, R. B., Sackett, D. L., and Gibson, E. S., 1976), a 24 study from the Cochrane review, compared a multi-component intervention 25 with usual care in patients with high blood pressure with positive results. Thirty 26 nine patients were enrolled in the study. It is not clear from the study the ages 27 of the patients. Patients in the intervention group were, “…all taught the 28 correct method to measure their own blood pressures, were asked to chart 29 their home blood pressures and pill taking, and taught how to tailor pill taking 30 to their daily habits and rituals. They also visited fortnightly at the worksite a 31 high-school graduate with no formal health professional training who 32 reinforced the experimental manoeuvres and rewarded improvements in Medicines concordance: full guideline DRAFT (July 2008) page 303 of 373 DRAFT FOR CONSULTATION 1 adherence and blood pressure. Rewards included allowing participants to 2 earn credit, for improvements in adherence and blood pressure, which could 3 be applied towards the eventual purchase of the blood pressure apparatus 4 they had been loaned for the trial”. (Haynes, R. B., Yao, X., Degani, A. et al , 5 2005). This study was conducted in Canada. There was a significant increase 6 in adherence associated with the intervention. 7 Johnson (1978) (Johnson, A. L., Taylor, D. W., and Sackett, D. L., 1978), a 8 study from the Cochrane review, compared four groups (1) self-recording and 9 monthly home visits, (2) self recording only, (3) monthly home visits, and a 10 control group consisted of (4) neither self-recording nor home visits with 11 negative results. One hundred and forty patients aged between 35 and 65 12 years were included in the study. Patients receiving antihypertensive 13 medications were studied. Participants, “…in groups (1) and (2) received a 14 blood pressure kit and instruction in self-recording. Patients in the self- 15 recording groups were to keep charts of their daily blood pressure readings 16 and were instructed to bring these charts to their physician at each 17 appointment. Subjects in groups (1) and (3) had their blood pressure 18 measured in their homes every four weeks, and the results were reported to 19 both the patient and the physician”. (Haynes, R. B., Yao, X., Degani, A. et al , 20 2005). This study was conducted in Canada. There was no effect on 21 adherence from either intervention. 22 Marquez-Contreras (2006) (Marquez-Contreras, E., Martell-Claros, N., Gil- 23 Guillen, V. et al , 2006), included in the Cochrane Updated Review, 24 conducted a randomised controlled trial of a programme of home blood 25 pressure management (HBPM) in patients with mild-to-moderate arterial 26 hypertension. This study was conducted in 40 primary care centres in Spain. 27 250 patients were included with data for 226. Mean age of participants was 59 28 years, and around 50% females/males. The no. of diseases was significantly 29 higher in the intervention group 2.6 (SD=1.6) vs 2.2 (SD=1.2), p=0.023). 30 Patients in the control group received usual GP care and the intervention 31 group received the intervention from GPs plus an OMRON automatic monitor 32 for HBPM. The programme was measuring their blood pressure 3 days a Medicines concordance: full guideline DRAFT (July 2008) page 304 of 373 DRAFT FOR CONSULTATION 1 week (Tuesdays, Thursdays and Saturdays), twice before breakfast and twice 2 before supper and record the results on a card. 74% of the control group and 3 92% of the intervention group were compliant (measured by MEMS) (95% CI 4 63.9-84.1 and 86.7-97.3, p=0.0001); the mean percentage compliances were 5 87.6% for the control group and 93.5% for the intervention group (95% CI 6 81.2-94.0 and 88.7-98.3, p=0.0001); the percentage of days the drug was 7 taken correctly were 83.6% and 89.4; the percentages of participants taking 8 medication at the correct time was 79.89 vs. 88.06. 9 Rudd (2004) (Rudd, P., Miller, N. H., Kaufman, J. et al , 2004) (included in 10 the Cochrane Updated Review) conducted a RCT of a system for patients to 11 monitor their own blood pressure. Patients of two medical clinics in California 12 were randomised to receive routine care (n=76) or an automated blood 13 pressure device at home with management by a nurse care manager (n=74). 14 The mean age for the intervention was 59 and 60 for the control group. 50% 15 of the intervention and 56% of the control group were female. Patients 16 recorded their blood pressure twice a day at the same time using the semi- 17 automated portable device. At the end of the week the device printed a report 18 of up to 14 measurements and every two weeks patients were to mail the 19 values on the printout to the nurse care manager who used the data to guide 20 drug therapy. Any new blood pressure drug initiation was requested from the 21 doctor. The primary outcome was change of BP at 6 months. Secondary 22 analyses were made for frequency of drug changes and adherence. 23 Adherence was measured by a drug event monitor (a microchip in the pill 24 bottle lid of the most frequently used medication) and the patients were 25 required to return to the clinic at 3 and 6 months so this data could be 26 downloaded, although this was not used as feedback to patients, physicians 27 or nurse care managers. The mean daily adherence rate was 80% (SD= 23%) 28 for the intervention group and 69% (SD=31%) for the control group, p=0.03). 29 Morice (2001) (Morice, A. H. and Wrench, C., 2001), a study from the 30 Cochrane review, compared an asthma-nurse led intervention with routine 31 care in patients with asthma with negative results. A group of 80 patients (53 32 women) aged between 16 and 72 years were included in the study. Mean age Medicines concordance: full guideline DRAFT (July 2008) page 305 of 373 DRAFT FOR CONSULTATION 1 was 36.1 years. Compared with the control group, patients in the educational 2 intervention group had a minimum of two separate sessions, lasting on 3 average 30 minutes each. These were carried out on an individual basis. The 4 first session involved discussion on the basic mechanisms of asthma, 5 including common triggers and an explanation of the changes which occur to 6 the airways resulting in the symptoms experienced by the patient. This was 7 supported by illustrations in the ‘Regular Therapy with Asthma’ booklet which 8 was given to each intervention group patient. Lifestyle influences, such as 9 occupation and leisure activities were discussed where appropriate to the 10 individual. The need for ‘preventer’ and ‘reliever’ medication was also 11 emphasized during this session. Patients were encouraged to actively 12 participate in the session and relatives were included at the patients’ request. 13 The second session took place on the following day. Previously given 14 information was briefly summarized with input from the patient as a means of 15 checking understanding. An agreed individualized self-management plan was 16 determined, with written instructions using the ‘Sheffield Asthma Card’. This 17 also contained a telephone contact number. Each patient was given a peak 18 low meter to take home and instructions on monitoring, with documentation of 19 predicted peak low measurement and parameters for altering treatment, as 20 well as clear written guidelines on when to seek emergency care. Home 21 intervention was based upon a combination of symptoms and peak low 22 recordings and all guidance offered throughout the educational programme 23 was based on the BTS guidelines for the management of asthma in adults” 24 (Haynes, R. B., Yao, X., Degani, A. et al , 2005). This study was conducted in 25 the UK. There were no significant improvements in compliance at six months. 26 Medicines concordance: full guideline DRAFT (July 2008) page 306 of 373 DRAFT FOR CONSULTATION 1 2 8.14 Key Clinical Question : Does the use of dosette boxes increase adherence to prescribed medication? Related Evidence statements references (summary of Evidence into recommendations evidence) 1. No evidence was The GDG were aware of regular use in the found on the effects of community of compliance aids such as dosette dosette boxes in boxes in the community. When general searches increasing adherence did not locate evidence regarding their use the GD to prescribed asking for specific searches. The GDG recognized medication the potential use of these devices for patients who may have wish to organize their treatment regimen but did not consider it appropriate to make a gener recommendation about their use. 3 4 8.14.1 Methods of the evidence review 5 This paper includes a narrative summary of the included evidence, structured 6 according to the category of the intervention, following the agreed reviewing 7 protocol: 8 Types of studies – no restrictions on study design. 9 Types of participants- people prescribed medication for a medical condition. 10 Duration of studies - no time limit specified. 11 Types of interventions - any interventions intended to assess the correlation 12 between the use of dosette boxes and the impact on adherence to prescribed 13 medication. 14 8.14.2 Evidence review 15 No studies that were relevant to this clinical question were retrieved. 16 Medicines concordance: full guideline DRAFT (July 2008) page 307 of 373 DRAFT FOR CONSULTATION 1 9 Reviewing medicines 2 9.1 Recommendations *listed at the start of the guideline according to chapter 3 4 9.2 5 Review of medicines and medicine taking is seen as an important aspect of 6 health care. Professionals involved in prescribing and dispensing of medicines 7 are currently reimbursed for reviewing medicines. General practitioners in the 8 UK are remunerated for medication review via the Quality and Outcomes 9 Framework (QOF). Community pharmacists are reimbursed for carrying out 10 reviews which are called Medicines Use Reviews (MURs). The Dispensing 11 Review of Use of Medicines (DRUM) is part of the Dispensing Services 12 Quality Scheme for GP surgeries. 13 The terminology in this area is not standardised and is subject to change. The 14 Medicines Partnership Programme 15 defined medication review as ‘a 15 structured, critical examination of a patient’s medicines with the objective of 16 reaching an agreement with the patient about treatment, optimising the impact 17 of medicines, minimising the number of medication-related problems and 18 reducing waste’. It is implicit in this definition that the patients is involved. In 19 ‘Room for Review’ in 2002 they suggested four levels of medication review – 20 level 0 which is an ad-hoc opportunistic review; level 1 a prescription review 21 which is a technical review of a patients list of medicines; level 2 is a 22 treatment review which is a review of medicines with the patients full notes 23 and level 3 which is a clinical medication review which is a face-to face review 24 with patients of medicine and condition. A review with the patient’s notes but 25 not necessarily with the patient (as in level 2 as described above) fulfils the 26 criteria for QoF. An MUR is described as a one-one conversation between 27 people and pharmacists that are designed to identify any problems a person is 28 experiencing with their medicines (Pharmacy in England White paper 2008). 29 Community pharmacists carrying out these reviews will not generally have 15 Introduction http://www.npc.co.uk/med_partnership/assets/room_for_review.pdf Medicines concordance: full guideline DRAFT (July 2008) page 308 of 373 DRAFT FOR CONSULTATION 1 access to clinical information about patients. The recent Pharmacy in England 2 White Paper (2008)16 reports that many people report satisfaction with this 3 service but longer term impacts can not be assessed. The White Paper 4 reports that government plans for MUR services to be prioritised to meet 5 health needs and ensuring funding rewards health outcomes. 6 The National Prescribing Centre has recently revisited the topic in A Guide to 7 Medication Review (2008). The guide aims to advise those providing and 8 commissioning medication reviews. This characterises 3 types of medication 9 review with an emphasis on the purpose of the review: Type 1 Prescription 10 review; Type 2 Concordance and compliance review and Type 3 Clinical 11 medication review. The three types of medication review replace the earlier 12 levels of medication review. This reclassification appears to make clearer the 13 role of the review and the place of the patient and clinical information in 14 different types of review., 15 16 The GDG were interested in whether there was any evidence that medication 17 review improved either shared decision-making or adherence. In this context 18 medication review has to involve a face to face meeting with professionals 19 and patient. The professional involved was not pre-defined. The evidence 20 search used ‘medication review’ as a generic term. 21 16 http://www.official-documents.gov.uk/document/cm73/7341/7341.pdf Medicines concordance: full guideline DRAFT (July 2008) page 309 of 373 DRAFT FOR CONSULTATION 1 2 3 9.3 Key Clinical Question: does medication review increase shared decision–making or adherence? Related references Evidence Evidence into recommendations statements (summary of evidence) All retrieved 1. There is The GDG considered that review of prescribed evidence. conflicting evidence medicines is most commonly undertaken in with regards to clinical settings as part of management of whether medication patients and their medical problems. In this review increases setting it is seen as integral to continuing care adherence. and not separate from it. The GDG considered that all levels of medication review as described Lowe (2000) (Lowe, 2. Five RCTs C. J., Raynor, D. K., conducted in the UK Purvis, J. et al , shows that 2000); Sturgess medication review (2003) (Sturgess, I. increased K., McElnay, J. C., adherence to Hughes, C. M. et al , prescribed 2003); Nazareth medication. in ‘Room for Review’ take place in this setting and have a role. Revisiting a decision to prescribe medicines and exploring patients medicine-taking behaviour was considered by the GDG to be part of the dynamic process that long term medicine prescribing required. The research evidence primarily addresses (2001) (Nazareth, I., medication reviews that take place separate Burton, A., Shulman, from the delivery of clinical care, often by S. et al , 2001); practitioners who do not have access to clinical Bernsten (2008) history and notes. These have been a recent (Bernsten, C, development and have primarily involved Bjorkman, I, pharmacists. Most of the evidence on reviews Caramona, M et al , by pharmacists comes from studies that 2001); Begley (1997) targeted older adults on multiple medications. (Begley, S., Many studies include quite complex Medicines concordance: full guideline DRAFT (July 2008) page 310 of 373 DRAFT FOR CONSULTATION Livingstone, C., pharmaceutical care programmes where the Hodges, N. et al , interventions consists of a number of 1997). components including education and follow up which the GDG considered more intensive than Lipton (1994) (Lipton, 3. There is H. L. and Bird, J. A., conflicting evidence 1994); Hanlon (1996) from six RCTs (Hanlon, J. T., conducted outside Medication review can have benefits for the Weinberger, M., the UK that patient but evidence was conflicting whether Samsa, G. P. et al , medication review this led to improvements in adherence to 1996); Chisholm increased prescribed medication. (2001) (Chisholm, M. adherence to A., Mulloy, L. L., prescribed Jagadeesan, M. et al medication. is currently provided in any type of medication review provided in the UK. The GDG were particularly concerned that reviews of medication carried out remote from the clinical settings needed to feed back to , 2001); Taylor (2003) clinicians who were involved in prescribing and (Taylor, C. T., Byrd, other aspects of care. Increasing the number of D. C., and Krueger, medication reviews and the personnel involved K., 2003); Grymonpre in carrying them out might not be effective if (2001) (Grymonpre, communication and follow up is not achieved. R. E., Williamson, D. A., and Montgomery, The GDG were clear that the lack of research P. R., 2001); on reviews conducted as part of clinical care Sookanekun (2004) should not indicate that these were not of value. (Sookaneknun, P., Review of medications will continue to be part Richards, R. M., of delivery of health care. As responses to Sanguansermsri, J. et medication can change over time, both in terms al , 2004) of patient behaviour (adherence) and clinical Grymonpre 4. Medication review (2001)(Grymonpre, R. was carried out by E., Williamson, D. A., pharmacists in all of and Montgomery, P. the RCTs, except for R., 2001) one RCT where a trained volunteer outcomes, a process of medication review is likely to be necessary and be part of on-going processes of decision-making and medicinetaking. An informal review of medication should continue as part of good clinical practice but it is not possible to be recommend precise timings Medicines concordance: full guideline DRAFT (July 2008) page 311 of 373 DRAFT FOR CONSULTATION undertook the for formal medication reviews outside the review which was clinical setting. then reviewed by a pharmacist consultant. Lowe (2000)(Lowe, 5. Most of the RCTs C. J., Raynor, D. K., included only Purvis, J. et al , participants over 65 2000); Hanlon years old. (1996)(Hanlon, J. T., Weinberger, M., Samsa, G. P. et al , 1996); Grymonpre (2001)(Grymonpre, R. E., Williamson, D. A., and Montgomery, P. R., 2001); Nazareth (2001)(Nazareth, I., Burton, A., Shulman, S. et al , 2001); Taylor (2003)(Taylor, C. T., Byrd, D. C., and Krueger, K., 2003); Bernsten (2008) (Bernsten, C, Bjorkman, I, Caramona, M et al , 2001); Begley (1997) (Begley, S., Livingstone, C., Hodges, N. et al , 1997); Sturgess Medicines concordance: full guideline DRAFT (July 2008) page 312 of 373 DRAFT FOR CONSULTATION (2003) (Sturgess, I. K., McElnay, J. C., Hughes, C. M. et al , 2003) Medicines concordance: full guideline DRAFT (July 2008) page 313 of 373 DRAFT FOR CONSULTATION Methods of the evidence review 1 9.3.1 2 The titles and abstracts of studies retrieved by an electronic search for 3 medication review were scanned for relevance to the question of whether 4 medication review increases adherence to medications. Any potentially 5 relevant publications were obtained in full text. These were then reviewed to 6 identify the most appropriate evidence to help answer the question and to 7 ensure that the recommendations are based on the best available evidence. 8 This process required four main tasks: selection of relevant studies; 9 assessment of study quality; synthesis of the results; and grading of the 10 evidence. 11 This paper includes a narrative summary of the included evidence, following 12 the agreed reviewing protocol: 13 Types of studies - randomized controlled trials (RCTs) of medication review 14 interventions to increase adherence. 15 Types of participants - people prescribed medication for a medical condition. 16 Medication review performed by any healthcare professional or trained 17 personnel. 18 Setting- Carried out in the community. 19 Duration of studies - No time limit specified for this evidence review. 20 Types of interventions - any medication review ( as implying face to face 21 meeting between the patient and the health care professional doing the 22 review) interventions intended to change adherence to prescribed medication. 23 The content and delivery of interventions are not standardized in the literature. 24 The term ‘pharmaceutical care programme’ is used and this applies to 25 pharmacist led programmes which assess medication use, develop an 26 intervention and provide long term follow up to patients including liaison with 27 the prescriber. Some of these interventions provide intensive support.. Some 28 subjective assessment of studies was required as content is often not well 29 defined. Many of studies on interventions to increase adherence used Medicines concordance: full guideline DRAFT (July 2008) page 314 of 373 DRAFT FOR CONSULTATION 1 pharmacists to carry out the intervention but we have included here only 2 studies that were carried out in the community and were providing general 3 review rather than disease specific support. 4 Types of outcome measures - any prescribed medication adherence 5 outcomes which changed as a result of the medication review. Outcomes 6 relevant to patient involvement were reported as part of the evidence review. 7 9.3.2 8 Of the RCTs found relating to medication review, many had to be excluded. as 9 they did not have adherence outcomes. Instead they focused on hospital re- 10 admissions, care home admissions, death and cost effectiveness. One high 11 quality RCT conducted by Zermansky (2002) (Zermansky, A. G., Petty, D. R., 12 Raynor, D. K. et al , 2002) for HTA could not be included as there were no 13 specific adherence outcomes Zermansky (2002) studied whether a trained 14 pharmacist could conduct effective clinical medication reviews of elderly 15 patients who were on repeat prescriptions from their GP. The participants 16 were 65 years or over on repeat medication, who were not resident in a 17 nursing or residential home and were not terminally ill. The study lasted 12 18 months and the intervention involved the pharmacist assessing the patient, 19 their illnesses and their medication regimen and making recommendations. 20 The primary outcome measure was the number of repeat medication changes 21 per patient, which was 2.2 in the intervention group and 1.9 in the control 22 group (Difference of 0.31, 95% CI 0.06 to 0.57), p=0.02). The secondary 23 outcome was the effect on cost of medication. There was a rise in repeat 24 medication items for both groups, but this was significantly less for the 25 intervention group (intervention mean 0.2, SD 1.55; control mean 0.4, SD 26 1.53, difference -0.2, 95% CI, -0.4 to-0.1). The cost saving for the intervention 27 group compared to the control group was £4.75 per 28-day month, a total of 28 £61.75 per patient per year. 29 A systematic review (Holland 2007) focusing specifically on pharmacist-led 30 medication review was found. However the primary outcome of interest was 31 reduction of hospital admissions and deaths in older people and adherence Evidence review Medicines concordance: full guideline DRAFT (July 2008) page 315 of 373 DRAFT FOR CONSULTATION 1 was a secondary outcome. The studies included all forms of medication 2 review for checking and optimising the patients’ drug regimens apart from 3 those with only knowledge and/or adherence outcomes. They reported that 14 4 of the trials included adherence, with 7 reporting a significant effect and 7 5 reporting a non-significant positive effect. 6 The term ‘pharmaceutical care programme’ is also used in the literature and 7 this generally applies to pharmacist led programmes which assess medication 8 use, develop an intervention and provide long term follow up to patients. 9 Although more intensive that any programmes currently delivered in the UK 10 we included these studies as it was important to assess whether such 11 structured and intensive support was either clinically or cost effective. 12 Some subjective assessment of the studies was necessary as the content of 13 the reviews and pharmaceutical programmes is not always clearly defined. 14 9.3.2.1 15 Sturgess (2003) (Sturgess, I. K., McElnay, J. C., Hughes, C. M. et al , 2003) 16 measured a structured pharmaceutical care programme provided to elderly 17 patients by community pharmacists 191 elderly patients with a mean age of 18 73.1 ± 5.0 for the intervention group and 74.2 ± 6.3 for the control group . This 19 RCT was conducted in Northern Ireland. In the intervention pharmacists 20 assessed patients to identify drug-related problems. A number of information 21 sources were used by intervention pharmacists during this assessment 22 procedure including: the patient (via informal questioning), the patient’s GP, 23 study questionnaires and computerised medication records. During the 24 assessment, pharmacists were asked to document any identified drug-related 25 problems and to form with the patient an intervention and monitoring plan e.g. 26 education, implementation of adherence improving strategies. Pharmacists 27 visited patients at home to assess storage of medicines where problems were 28 identified. 29 Self reported compliance: between-group analysis at each assessment point 30 indicated that a significantly higher proportion of intervention patients were 31 compliant with their medicine at 12 (intervention group: 40.4%, control group: RCTs conducted in the UK Medicines concordance: full guideline DRAFT (July 2008) page 316 of 373 DRAFT FOR CONSULTATION 1 24.4%) and 18 (intervention group: 47.3%, control group: 14.7%) months 2 compared to control patients (p < 0.05) (6 months: intervention group: 34.5%, 3 control group: 29.4%). Analysis of change in compliance during the study 4 (change in compliance status compared to that reported at baseline) showed 5 that a significantly higher proportion of intervention patients changed from 6 non-compliant to compliant compared to control patients (intervention 13.4% 7 vs. control 9.1%) and a significantly higher proportion of control patients 8 changed from compliant to non-compliant compared to intervention patients at 9 18 months (control 36.4% vs. intervention 4.5%). 10 Lowe (2000) (Lowe, C. J., Raynor, D. K., Purvis, J. et al , 2000) determined 11 whether a medicine review and education programme influenced elderly 12 patients’ compliance and knowledge compared to a control group in a RCT. 13 161 participants, mean age 77.5 (SD 65-96) for the intervention group and 75 14 (SD 65-88) for the control group, mainly female (67%), living with spouse or 15 relative 55% (intervention group) and 57% (control group) and prescribed an 16 average of 4 medicines (ranging from 1 to 8). The RCT was conducted in a 17 GP practice in Leeds, UK. An investigator visited patients and filled in a 18 structured questionnaire regarding their medicines, which medicine had been 19 used and patients’ understanding and ability to take medications. The 20 investigator then reported the findings to doctors where there was a need to 21 reduce dosage and discontinue medication, then liaised with the pharmacist 22 for modifications to medicine containers. At the second visit after a month they 23 delivered 1 months supply of medication and removed any other prescribed 24 medications. They discussed the regimen, the purpose of the medications 25 and the correct way to take them, with the use of a reminder chart if needed. 26 At 3 weeks follow-up participants were given a further months supply and 27 assessed on their knowledge and compliance, by counting the medications 28 left from the last visit. The mean compliance score was 91.3% for the 29 intervention group (95% CI, 89%-94%) and 79.5% for the control group (75%- 30 84%), which was significantly different (p<0.0001). Medicines concordance: full guideline DRAFT (July 2008) page 317 of 373 DRAFT FOR CONSULTATION 1 N.B It should be noted that that this study was under 6 months duration but 2 the patients were followed up twice at 3 week to monthly intervals and the 3 study was of particular relevance. 4 Nazareth (2001) (Nazareth, I., Burton, A., Shulman, S. et al , 2001) 5 compared patients who had been discharged from hospital with a discharge 6 plan with those who had a standard discharge letter. This RCT included 362 7 patients from four hospitals in central London. The participants had a mean 8 age was 84 years in both groups (SD 5.2 and 5.4 respectively), mainly female 9 (62% and 66%), white (97%) with a mean of three chronic medical conditions 10 and prescribed a mean of 6 drugs (SD=2). The discharge plan included 11 assessing the prescribed medication, rationalising the drug treatment and 12 assessing patients’ medication management, knowledge and support. The 13 participants were then followed up 7 to14 days later at home by community 14 pharmacists who compared medicine taking with prescribed medications and 15 their understanding and adherence to the medication regimen. They 16 intervened when necessary and provided medication counselling, disposal of 17 excess medications and liaison with GPs. There was no significant difference 18 in adherence to medicines for either group at 3 months or at 6 months (49% 19 intervention group versus 50% for the control group, with a mean difference of 20 0.02 (95% CI -0.106 to 0.126). 21 Begley (1997) (Begley, S., Livingstone, C., Hodges, N. et al , 1997) 22 assessed the influence of domiciliary pharmacy visits on medication 23 management in sample of elderly people recently discharged from hospital to 24 their own homes. Patients were aged 75 years or older. The study included 25 one intervention group receiving home visits and counselling, in which 26 structured patient interviews were conducted during the domiciliary visits and 27 consisted of six sections: patient information; drug knowledge; Patient 28 dexterity; abbreviated mental test; medication management; and compliance 29 with medication regimen. Patients were seen during 12 months. There were 30 two control groups: one which was the control and received visits only (called 31 V group), and other which was the control group that received traditional Medicines concordance: full guideline DRAFT (July 2008) page 318 of 373 DRAFT FOR CONSULTATION 1 pharmaceutical services with no visits except for the beginning and the end of 2 the study (NV group). 3 At each visit there were significant differences between the groups in terms of 4 distribution of patients at the various levels of compliance 5 (p<0.001).Compliance was higher at 3 months and 12 months for the 6 intervention group compared to the other control groups (p<0.001), despite 7 the low compliance value for the intervention group at the 12 month visit. 8 Patients in the intervention group who increased their compliance rates 9 between visits also increased their drug knowledge scores (p<0.005). Mean 10 scores for drug knowledge did not differ significantly between the groups at 11 any of the visits, although the mean score for the intervention group increased 12 significantly between the initial and the two weeks visits (p=0.001). There 13 were no changes for patient dexterity scores between groups at any point of 14 the study. Contacts with GP and health workers was lower for the intervention 15 group than for the control (V) in each of the four time periods (p<0.01). 16 Bernsten (2008) (Bernsten, C, Bjorkman, I, Caramona, M et al , 2001) 17 conducted a multicentre RCT in seven European Countries including the UK 18 that evaluated a pharmaceutical care programme provided to elderly patients 19 (aged 65 or older) taking 4 or more medications by community pharmacists. A 20 total of 1290 intervention patients and 1164 control patients were recruited. 21 The programme interventions included: 1) educating the patient about their 22 drug regimen and their condition; 2) implementing compliance-improving 23 interventions such as drug reminder charts; 3) rationalising and simplifying 24 drug regimens in collaboration with the patients GP. This was a continuous 25 process throughout the 18 months of the study. 26 Generally, the programme had some positive effects on humanistic health 27 outcomes such as satisfaction with treatment, and sign and symptom control, 28 and on economic outcomes, but had less impact than anticipated on drug 29 therapy, drug knowledge and compliance with medication. An analysis of 30 changes in compliance during the study indicated that at 18 months a 31 significantly higher proportion of the intervention patients changed from being 32 noncompliant to compliant compared with the control groups (p=0.028). Medicines concordance: full guideline DRAFT (July 2008) page 319 of 373 DRAFT FOR CONSULTATION 1 Intervention patients rated the services provided higher that the control at 6 2 and 18 months (p<0.05). There was a small statistically significant increase in 3 satisfaction in the intervention group over time (baseline vs. 12 months 4 p=0.039). 5 9.3.2.2 6 Lipton (1994) (Lipton, H. L. and Bird, J. A., 1994) assessed the impact of 7 clinical pharmacists' consultations on drug regimens, compliance, and health 8 service use of 706 geriatric hospitalized patients discharged on 3 or more 9 medications. The RCT was conducted in the USA. Mean age was 74.6 in the 10 experimental group and 74.4 in the control group. Pharmacists consulted with 11 experimental patients at discharge and 3 months thereafter, and with 12 physicians as needed. Controls received usual care. At 6-8 weeks after 13 enrolment, experimental patients were more knowledgeable about regimens 14 than controls. At 12-14 weeks, they were on fewer medications and less 15 complex regimens, and had better compliance scores p<0.001). There was no 16 effect on service use or charges, perhaps due to inadequate sample size and 17 lack of targeted drug groups analysis. 18 Hanlon (1996) (Hanlon, J. T., Weinberger, M., Samsa, G. P. et al , 1996) 19 was an RCT which compared the effects on elderly outpatients who had an 20 additional pharmacist intervention with those who received usual care from 21 their physician. Most of the patients were male (98% in the intervention and 22 100% in the control group), white (79% and 75% respectively), married 23 (65.7% intervention, 85.4% control), with baseline compliance rates of 73% 24 and 74% respectively. The mean age of participants was 70 years old. The 25 RCT was conducted in Durham, North Carolina. Before attending the 26 physician the pharmacist reviewed their medical records and medication lists 27 to ascertain their current medication use, drug-related problems and to 28 evaluate their needs by applying the Medication Appropriateness Index. Their 29 findings were then reported to the physician. The Pharmacist educated the 30 patient on drug-related problems and encouraged compliance through 31 strategies such as medication reminders and written patient materials. They 32 reviewed safe medicine use and the importance of discussing medications RCTs conducted outside the UK Medicines concordance: full guideline DRAFT (July 2008) page 320 of 373 DRAFT FOR CONSULTATION 1 with physicians. There were no significant differences between the groups at 2 the end of the follow-up period with regard to medication compliance (77.4% 3 of intervention group and 76.1% of control group complied, p=0.88). 4 Chisholm (2001) (Chisholm, M. A., Mulloy, L. L., Jagadeesan, M. et al , 5 2001) studied the compliance rates of patients who received a clinical 6 pharmacist intervention in addition to usual care compared to control patients 7 after a renal transplant. The RCT included 24 participants, 75% male with a 8 mean of 49 years (+/-10 years) and 58.3% Caucasian, 37.5% African- 9 American and 1 Hispanic. The RCT was conducted in Augusta, Georgia, 10 USA. The Pharmacist obtained medication histories and reviewed 11 medications monthly. They made recommendations to the nephrologists and 12 counselled the patients on their medication, including instructing how to use 13 the medication. Patients were encouraged to call them with any questions. 14 The patients were assessed on their understanding of their medication and 15 advised on how to enhance compliance. 16 At 12 months the compliance rate was statistically significant for the 17 intervention group 96.1% (+/- 4.7%) compared to the control group 81.6% (+/- 18 11.5%), p<0.001. For 6 of the 12 months there were differences in compliance 19 rates (64 to 100% for control group and 89 to 100% for the intervention group) 20 with the intervention group always at a higher rate (p<0.05). The duration that 21 patients complied for also differed with the intervention group remaining 75% 22 compliant each month compared to only 33.3% of the control group (p<0.05). 23 Taylor (2003) (Taylor, C. T., Byrd, D. C., and Krueger, K., 2003) conducted 24 an RCT of patients attending a community-based physician and compared 25 those who additionally received a Pharmacist intervention to a control group. 26 The majority of participants were women (63.6% in the intervention group vs 27 72.2% in the control group, p=0.445), most were white (60.6% vs 61.1%, 28 p=.966) with a mean age of 64.4 and 66.7 years respectively (p=0.467), the 29 majority were married 75.8% vs 72.2 (p=0.935) with 12 years mean 30 education. They were taking on average six medications each. The RCT was 31 set in medicine clinics in Alabama, USA. Medicines concordance: full guideline DRAFT (July 2008) page 321 of 373 DRAFT FOR CONSULTATION 1 The pharmacists evaluated the patients’ medication, reviewed medical 2 records and examined medication history to determine compliance and 3 complications with medication. Therapeutic recommendations were made to 4 the physicians and they made follow-up visits to patients and gave 5 individualised education and were available to answer questions. Patients' 6 responses to drugs were monitored and their medication regimens 7 consolidated, dosage frequency was reduced and medication reminders and 8 techniques for using certain devices were taught. 9 The number within the intervention group to have compliance scores of 80- 10 100% increased by 15% but there was no change for the control group (time 11 period not stated). By 12 months this difference was not significant, 100% of 12 patients in the intervention group versus 88.9% (SD 6.3) of the control group 13 had compliance scores of 80-100%, p=0.115). At baseline this was 84.9% 14 (SD=6.7) and 88.9% respectively (SD 5.8, p=0.728). 15 The most frequently sited reasons for not complying with medication were 16 forgetting to take the medications (n=10), having too many to take (n=9), 17 finding it hard to read or understand the directions (n=4) and too much trouble 18 to take (n=4). 19 Grymonpre (2001) (Grymonpre, R. E., Williamson, D. A., and 20 Montgomery, P. R., 2001) compared the impact on geriatric patients who 21 received pharmaceutical care compared to those patients who did not in a 22 RCT. Most of the patients were female (75% intervention vs 83% control, 23 p=0.254), aged 77 (SD 8.0 to 9.0), Caucasian (100%) and lived alone 61% vs 24 77% respectively, p=0.018. The RCT was conducted in a community-based 25 health clinic in Manitoba, Winnipeg, Canada. Volunteers and staff were 26 trained to conduct a comprehensive medication review which was utilised by 27 the pharmacist to identify and document potential and actual drug-related 28 issues and to address these with the patient and they physician. Their use of 29 prescribed and non-prescribed medicines, social drugs, home remedies, 30 regimen, adherence and communication with GPs, problems or side effects 31 with drugs were all assessed. The recommendations were given in a letter to Medicines concordance: full guideline DRAFT (July 2008) page 322 of 373 DRAFT FOR CONSULTATION 1 physicians and the patients were followed up by the pharmacist when required 2 to monitor therapeutic endpoints and sort out any problems that had arisen. 3 The mean number of mediations adhered to at follow-up was 87 (+/-46) for 4 the intervention and 85 (+/-41) for the control group, p=0.895, showing no 5 significant difference in adherence. 6 Sookaneknun (2004) (Sookaneknun, P., Richards, R. M., 7 Sanguansermsri, J. et al , 2004) compared hypertensive patients assigned 8 to a pharmacist-involved group with those who had no pharmacist 9 involvement, with the objective of stabilising blood pressure. The participants 10 of the RCT included 235 patients, mean age 63 years old and mainly female 11 (64% in intervention and 71% in control group). The RCT was conducted in 12 Thailand. The intervention group’s blood pressure was measured every month 13 by the pharmacist and they assessed the patients understanding of 14 medications, adherence and reviewed adverse effects from the medications. 15 Medication counselling was given and drug-related problems were identified, 16 resolved and prevented. The recommendations for change of medication 17 regimen were given to the physicians. The adherence at pre-test was not 18 significantly different but at post-test the treatment group had significantly 19 increased adherence compared to the control group, Pre-test adherence of 20 80% or more was found in 51% of the treatment group and 56% of the control 21 group. At post-test this had increased to 63% for the treatment group and had 22 remained constant (55%) for the control group (p=0.014). 23 Quality of studies 24 It should be noted that the quality of many of the RCTs was low. This was 25 mainly due to the possibility of bias occurring within the methodology. 26 27 28 Medicines concordance: full guideline DRAFT (July 2008) page 323 of 373 DRAFT FOR CONSULTATION 1 2 10 Health Economics and Medicines Concordance 3 4 10.1 Introduction 5 Health economics is about improving the health of the population through the 6 efficient use of resources, so it necessarily applies at all levels, including 7 individual clinical decisions. Clinicians already take resources and value for 8 money into account when making clinical decisions, and the incorporation of 9 good-quality health-economic evidence into clinical guidelines can help make 10 this more consistent. The guideline development group (GDG) is required to 11 make decisions based on the best available evidence of both clinical and cost 12 effectiveness. 13 This chapter will outline the methodology, approach and findings of health 14 economic evidence that were used to inform the GDG in their decision 15 making. 16 10.1.1 How terminology of compliance, concordance and 17 adherence is used in HE literature 18 The terminology relevant to medicine taking i.e. the term compliance, 19 adherence and concordance have a different meaning when using a health 20 economic perspective. In Chapter 3, the relevant concepts have been 21 discussed. In a recent economic review paper (Elliott, Rachel A., Barber, Nick, 22 and Horne, Rob, 2005), the authors state that “compliance and adherence 23 imply patient behaviour being congruent with healthcare providers’ 24 recommendation”. We found that most economic papers use adherence and 25 compliance interchangeably, despite the evident conceptual difference in 26 meaning. In accordance with this guideline, this chapter will use “shared 27 decision making” (SDM) to describe “concordance as a patient centred 28 process where health care professional (HCP) make a therapeutic alliance 29 with a patient, one result of which may be increased compliance” and Medicines concordance: full guideline DRAFT (July 2008) page 324 of 373 DRAFT FOR CONSULTATION 1 adherence for describing “medicine taking behaviour congruent with 2 prescriber’s recommendation”. 3 4 As discussed in chapter 3.6, the structure chosen for the guideline was to look 5 separately at decision –making about medicines and medicines taking. This 6 understanding suggests two separate processes - firstly the consultation with 7 the health care provider (HCP) where shared decision making is possible and 8 secondly the medicine taking after the consultation. This includes additional 9 contacts with dispensers and HCPs other than the prescriber as represented 10 in the care pathway. This division will be reflected in the structure of this 11 chapter. 12 13 14 10.2 Health Economics methods 15 Economic evaluation provides a formal comparison of benefits and harms as 16 well as the costs of alternative health programmes. It helps to identify, 17 measure, value and compare costs and consequences of alternative 18 treatment options. These outcomes are usually synthesised in cost 19 effectiveness (CEA) or cost utility analysis (CUA), which reflect the principle of 20 opportunity costs. For example, if a particular treatment strategy were found to 21 yield little health gain relative to the resources used, then it could be 22 advantageous to re-deploy resources to other activities that yield greater 23 health gain. 24 25 To assess the cost-effectiveness of interventions to increase adherence, we 26 conducted a comprehensive systematic review of the economic literature 27 relating to medicines and nonadherence. 28 29 In accordance with the NICE social value judgement the primary criteria 30 applied for an intervention to be considered cost effective were either: 31 Medicines concordance: full guideline DRAFT (July 2008) page 325 of 373 DRAFT FOR CONSULTATION 1 a) The intervention dominated other relevant strategies (that is it is both less 2 costly in terms of resource use and more clinically effective compared with the 3 other relevant alternative strategies); or 4 5 b) The intervention cost less than £20,000 per quality-adjusted life-year 6 (QALY) gained compared with the next best strategy (or usual care) 7 8 10.2.1 Health Economic evidence review methodology 9 10 The following information sources were searched: 11 • Medline (Ovid) (1966-June 2006) 12 • Embase (1980-June 2006) 13 • NHS Economic Evaluations Database (NHS EED) 14 • PsycINFO 15 • Cumulative Index to Nursing and Allied Health Literature (CINAHL) 16 17 The electronic search strategies were developed in Medline and adapted for 18 use with the other information databases. The clinical search strategy was 19 supplemented with economic search terms. Titles and abstracts retrieved 20 were subjected to an inclusion/exclusion criterion and relevant papers were 21 ordered. No criteria for study design were imposed a priori. In this way the 22 searches were not constrained to randomised controlled trials (RCTs) 23 containing formal economic evaluations. Papers included were: 24 • Full/partial economic evaluations 25 • Considered patients over 16 years of age 26 • Written in English, and reported health economic information that could 27 be generalised to UK. 28 29 The full papers were critically appraised by a health economist using a 30 standard validated checklist. A general descriptive overview of the studies, 31 their quality, and conclusions was presented and summarised in the form of a 32 narrative review. Medicines concordance: full guideline DRAFT (July 2008) page 326 of 373 DRAFT FOR CONSULTATION 1 Each study was categorized as one of the following: cost-effectiveness 2 analysis or cost utility analysis (i.e. cost-effectiveness analysis with 3 effectiveness measured in terms of QALYs or life year gained). Some studies 4 were categorised as ‘cost consequences analysis’ or ‘cost minimisation 5 analysis’. These studies did not provide an overall measure of health gain or 6 attempt to synthesise costs and benefits. Such studies were considered as 7 partial economic evaluations. 8 9 10.2.2 Cost-effectiveness modelling methods 10 De novo modelling was considered for aspects of medicine taking. However, 11 due to heterogeneity in the population covered by this guideline this was not 12 possible. This is discussed in more detail in below.. 13 14 15 16 10.3 Shared decision making and medicine taking from a health economic perspective 17 10.3.1 Patient involvement and shared decision making 18 From an economic view, one aim of facilitating shared decision making (SDM) 19 on measurable outcomes is adherence to the joint decision taken. As 20 discussed previously in Chapters 3 and 4 of this guidance, the decision would 21 be taken jointly between HCP and patient and/or carer and based on the 22 appropriate information shared between them. The decision may result in the 23 agreement to prescribe and take a medication, or equally, to not prescribe and 24 take a different medication or no medication, depending on the individual 25 patient case. A programme that facilitates shared decision making between a 26 HCP and a patient can be seen as an intervention to increase adherence to 27 joint decisions including prescribed medication, and thereby health of the 28 population. 29 30 An alternative view focuses on the notion of sharing the decision itself. The 31 process of shared decision making can increase patient wellbeing by Medicines concordance: full guideline DRAFT (July 2008) page 327 of 373 DRAFT FOR CONSULTATION 1 improving patient satisfaction with the consultation, as well as the wellbeing of 2 possibly doctors and carers. This change in satisfaction and related wellbeing 3 can be summarised by the concept used in health economics of process utility 4 (Brouwer, Werner B. F., Exel, N. Job, Berg, Bernard van den et al , 2005). 5 Generally, being informed and involved in important decisions concerning our 6 lives, for example in an occupational setting, often increases well being and 7 satisfaction. As a result, terms such as enrichment and involvement were 8 introduced into management studies (Bambra, Clare, Egan, Matt, Thomas, 9 Sian et al , 2007). In health care, similar effects exist for patients. Thus, the 10 process and meaning of shared decision making can have an effect on well 11 being. However, it should be noted that, although many people may perceive 12 involvement in their care decisions as beneficial, not everybody will value 13 shared decision making in this way. As a result, a change in process utility 14 may both be positive or negative. Health economic research has so far mostly 15 concentrated on health-related consequences. More work is needed to 16 understand the magnitude of effects and value of process utility better. 17 18 There is a body of literature which describes and analyses peoples’ 19 preferences for various combinations of attributes. Discrete choice 20 experiments, also referred to as conjoint analyses, evaluate what people 21 actually value in a product or service. In recent years health economic 22 research has started to adapt this methodology to health care. Some papers 23 exist that investigate the relative importance of consultation features. This 24 evidence has not been systematically reviewed and appraised as part of he 25 health economic evidence reviews, however, the following papers were 26 identified as examples of conjoint analyses (Albus-Christian, Schmeisser- 27 Norbert Salzberger-Bernd Faetkenheuer, 2005); (Aristides, M., Weston, A. R., 28 FitzGerald, P. et al , 2004); (Beusterien, K. M., Dziekan, K., Flood, E. et al , 29 2005); (Frileux-Stéphanie, Sastre-María-Teresa-Muñoz Mullet-Etienne Sorum, 30 2004); (Goldring, A. B., Taylor, S. E., Kemeny, M. E. et al , 2002); (Honish, A., 31 Westerfield, W., Ashby, A. et al , 2006); (Kellett, N., West, F., and Finlay, A. 32 Y., 2006); (King, Madeleine T., Hall, Jane, Lancsar, Emily et al , 2007); 33 (Lamiraud, Karine and Moatti, Jean Paul, 2006); (Lancsar, E. J., Hall, J. P., Medicines concordance: full guideline DRAFT (July 2008) page 328 of 373 DRAFT FOR CONSULTATION 1 King, M. et al , 2007); (Mahadevia, Parthiv J., Shah, Shailen, Leibman, 2 Christopher et al , 2004); (Mark, T. L. and Swait, J., 2004); (Reginster, J. Y., 3 Rabenda, V., Neuprez, A. et al , 2006); (Richter, J., Eisemann, M., and 4 Zgonnikova, E., 2001); (Ryan, Mandy and Farrar, Shelley, 2000); (Stone, 5 Valerie E., Jordan, Jamie, Tolson, Jerry et al , 2004); (Vick, S. and Scott, A., 6 1998). 7 8 Despite limitations of the evidence, it seems likely that the shared decision 9 making process would improve cost-effectiveness by enabling patients to 10 make a prediction of their individual valuation of harms and benefits and 11 subsequently opt in or out of treatment. Theoretically it is likely to be of 12 economic benefit to enable patients to decline a suggested prescription as it 13 prevents people from accepting and filling prescriptions they would otherwise 14 had not intended to take. 15 16 This analysis does not include any anticipated impact on the resources used 17 to inform patients as well as on the overall length of the consultation. A 18 systematic review requested by the development group sought to answer the 19 question whether interventions to increase patient involvement increase the 20 length of the consultation between HCP and patient and other health care 21 resource use. The included studies resulted in a variety of results. Please 22 refer to Chapter 4 for more information on the clinical narrative. Overall, it was 23 found that simple interventions do not always increase overall length of 24 consultation. The GDG felt it important to look beyond the initial consultation 25 process when evaluating the benefits of health programmes to improve 26 patient involvement. Extra time invested in the initial consultation where 27 medicines are suggested and/or prescribed was deemed likely to have 28 benefits later on. Specifically, time of follow up consultation could be reduced 29 since concerns have been addressed and the perceived need to seek 30 information from other professionals could be reduced. The group 31 emphasized that an improvement of skill and knowledge on both sides can 32 enable the HCP to be more effective and efficient in the consultation process, 33 which may have a positive effect on the cost effective use of resources. Medicines concordance: full guideline DRAFT (July 2008) page 329 of 373 DRAFT FOR CONSULTATION 1 2 10.3.2 Medicine-taking 3 Medicine taking is a complex behaviour, specific to individuals. People are 4 understood to differ in their valuation of benefits and risks, thus making 5 choices not only based on information but also on their own preferences. 6 Apart from these more conscious, or intentional, choices, non adherence to 7 medication may be unintentional, when, for example, patients forget to take 8 their tablets or have physical difficulties to do so. 9 10 Interventions to raise adherence need to address underlying factors or 11 barriers that influence patients in their decisions about medicine taking in 12 order to improve their clinical outcome. It should be noted that the relationship 13 between adherence and clinical benefit is not necessarily positively linear, i.e. 14 it can not be presumed that the more adherent the patient the better the 15 outcome, as, for example, medications can also induce side effects and 16 adverse events. From a cost effectiveness perspective such interventions to 17 increase adherence would have to be effective in raising adherence and 18 improve survival and/or quality of life for patients enough to warrant any 19 additional costs. The effect of adherence to medicines on health is also 20 dependent on factors such as the natural history of the individual disease as 21 well as specific drug pharmacology and pharmacokinetics in particular. For 22 example, missing a small proportion of drugs that lower cholesterol levels may 23 have only marginal impact on cardiovascular risk. However, missing a similar 24 proportion of immunosuppressant drugs in the short term after renal 25 transplantation may cause renal rejection or even death. This ability of a drug 26 to sustain its pharmacological action after a dose has been missed has been 27 termed ‘drug forgiveness’ in the literature and has been defined as “the 28 postdose duration of action of the drug minus the prescribed interval between 29 doses” (Urquhart, J., 1996) (Girvin, B. G. and Johnston, G. D., 2004). The 30 variability of these factors imply significant heterogeneity in the population and 31 interventions covered in this guideline. 32 Medicines concordance: full guideline DRAFT (July 2008) page 330 of 373 DRAFT FOR CONSULTATION 1 Introducing an adherence enhancing intervention aimed at raising adherence 2 rates with a medication regimen will entail effects on costs and benefits. The 3 following sections will look in more detail at these theoretical effects. 4 5 Costs consequences 6 Costs include the costs of providing the intervention to increase adherence 7 comprising service costs incurred by the health service; knock on costs of 8 prescriptions that had formerly not been collected but as a result of the 9 intervention will be; associated opportunity costs to filled prescriptions that are 10 not taken and whose benefit is thus foregone. Also, if patients were to make 11 an informed decision jointly with the HCP, and were open to decide against 12 medication during consultation this could prevent a suggested prescription 13 from being filled. As a result, this could result in a possible monetary saving. 14 15 Health consequences 16 The effectiveness of receiving an intervention to increase adherence in form 17 of health benefits can be grouped in two main categories: treatment effects 18 from a particular medication and possible process utility gains or losses from 19 receiving the intervention. 20 21 Differences in medication treatment effect will depend on the effectiveness of 22 the adherence enhancing intervention, as well as the dose-response related 23 efficacy of the medication (cf. drug ‘forgiveness’, as mentioned in section 24 10.3.2). Assuming the intervention is effective and raises adherence rates, it 25 still remains uncertain what the incremental treatment effect of the medication 26 will be. The intervention may result in a potential QALY loss due to side 27 effects of the medication (where before non compliance would have prevented 28 them). 29 30 As discussed in above, there may be benefits attributable to an effect of 31 health outcome, but there may also be changes in process utility. Such 32 changes are dependent on individual preferences respective to attributes of Medicines concordance: full guideline DRAFT (July 2008) page 331 of 373 DRAFT FOR CONSULTATION 1 the provided intervention. The mentioned literature on conjoint analysis or 2 discrete choice experiments may shed light on such preferences. Thus, 3 receiving the intervention may result in a process utility gain or loss. Changes 4 in utility may also incur during the process of taking medicines, through 5 practical (swallowing a tablet) or psychological issues (cognitive dissonance). 6 These changes may be positive or negative depending on the individual 7 valuation of processes and agreement with diagnosis and choice of therapy. 8 9 10 11 10.4 Cost effectiveness literature review of Interventions to increase adherence 12 10.4.1 Methods of the review 13 The aim of this literature search was to provide an overview of the health 14 economic evidence base of the cost effectiveness of interventions to increase 15 adherence. For details on the methods please refer to section 2.6. The 16 chosen study methods, reported outcomes and results from the analyses 17 were presented. Titles and abstracts of records were retrieved by the 18 searches, suggested by the GDG or submitted by stakeholders. They were 19 then scanned for relevance to the key questions. Any potentially relevant 20 publications were obtained in full text. Papers were then reviewed to identify 21 the most appropriate evidence to help answer the key questions and to 22 ensure that the recommendations are based on the best available evidence. 23 This process required three main tasks: selection of relevant studies; 24 assessment of study quality; synthesis of the results. From this, three 25 systematic reviews were identified and selected by the GDG for subsequent 26 reviewing. Evidence reviews were undertaken using systematic, transparent 27 approaches following the Guidelines Manual 2007 (www.nice.org.uk). 28 29 Types of studies: Systematic reviews of cost effectiveness studies or 30 comparative economic analyses based on modelling or randomised controlled 31 trials (RCTs) of interventions to increase adherence. Medicines concordance: full guideline DRAFT (July 2008) page 332 of 373 DRAFT FOR CONSULTATION 1 Types of participants: people prescribed medication for a medical condition 2 from healthcare professionals in any health service setting. 3 Duration of studies: No time limit was applied. 4 Types of interventions: any interventions intended to change adherence to 5 prescribed medication. 6 Types of outcome measures: adherence/compliance levels as well as clinical 7 outcome. 8 9 10.4.2 Interventions to increase adherence: cost effectiveness 10 systematic reviews narrative 11 Three systematic reviews were found and included. First, the evidence from 12 these three systematic reviews will be presented in section 10.4.3 and an 13 update review of the latest systematic review (Elliott, Rachel A., Barber, Nick, 14 and Horne, Rob, 2005) will be presented as case studies in Section 10.5. 15 16 10.4.3 Summary of findings and methodological issues 17 highlighted by systematic reviews 18 A UK systematic review by Elliott (2005) (Elliott, Rachel A., Barber, Nick, and 19 Horne, Rob, 2005) found 42 papers, 30 from the US, two from the UK and the 20 remainder from a range of countries. This review does not report the cost 21 effectiveness findings of the individual studies, but concentrated on critiquing 22 the variability and prevailing inadequacy of the methods used. The authors 23 used a variety of minimum quality criteria comprising economic evaluation 24 quality criteria, standard hierarchies of evidence, and adherence-specific 25 design issues. 26 27 The second included systematic review by Hughes (2001) (Hughes, D. A., 28 Bagust, A., Haycox, A. et al , 2001) was from the UK. They only included 29 pharmacoeconomic evaluations that applied sensitivity analysis to non- 30 adherence rates in order to evaluate the impact of non-adherence on the cost- 31 effectiveness of different drug therapies. 22 evaluations were included for 32 reviewing of which 13 were from the US, 5 were Canadian and 2 UK papers. Medicines concordance: full guideline DRAFT (July 2008) page 333 of 373 DRAFT FOR CONSULTATION 1 Hughes et al. reviewed only two studies that were clearly described as 2 specifically investigating interventions that aim at increasing adherence, which 3 had mixed results. 4 5 A third, Dutch systematic review by Cleemput (2002) (Cleemput, I, Kesteloot, 6 K, and DeGeest, S, 2002) included 18 studies on the economics of 7 therapeutic non-adherence, which were assessed according to their definition 8 and measurement of medication non-adherence, study design, and 9 identification and valuation of costs and outcomes. The majority of articles 10 dealt exclusively with medication non-adherence. Eight studies examined the 11 economics of interventions to increase adherence. Of these, three were 12 excluded from this review on methodological grounds. 13 14 15 Expectedly, there was high heterogeneity. All three systematic reviews found 16 that the reviewed analyses were conducted in different clinical areas, 17 including chronic, acute and infectious, and studied a wide variety of 18 interventions. For example, Elliott (2005) reviewed programmes to improve 19 convenience of care, information, counselling, reminders, self-monitoring, 20 reinforcement, family therapy, and other forms of additional supervision or 21 attention. Elliott (2005) and Cleemput (2002) described considerable variation 22 between studies in terms of the form of delivery, such as telephone or postal 23 communication. Most interventions contained the provision by a specified 24 health professional, an educational component, and often used more than one 25 component. Cleemput (2002), Hughes (2001) and Elliott (2005) also 26 described heterogeneity in terms of study design. Of the 42 papers reviewed 27 by Elliott (2005), 21 were based on RCTs, of which 4 used a modelling 28 methodology. According to Hughes et al., a decision analytic model was 29 employed in most instances and some of the evaluations modelling chronic 30 illnesses adopted a Markov-type decision analysis approach. It was not 31 specified if models were based on one trial exclusively or on multiple sources. 32 The papers reviewed by Cleemput (2002) included different types of economic 33 evaluations, ranging from cost effectiveness over cost benefit to cost outcome Medicines concordance: full guideline DRAFT (July 2008) page 334 of 373 DRAFT FOR CONSULTATION 1 descriptive analyses. Time horizons varied considerably between studies and 2 ranged from 2 weeks to lifetime, with 3 studies spanning time up to 1 year and 3 10 papers of over 10 years. Cleemput (2002) found time horizons to be less 4 than 18 months in all studies, and said that long-term benefits of compliance- 5 enhancing interventions could thus rarely be shown. 6 7 All three systematic reviews found inadequate definitions given for non- 8 adherence. Some studies used previously unvalidated adherence measures. 9 Often no attempt at defining the measure of adherence was made. 10 Alternatively, adherence was defined arbitrarily at some cut off point, or simply 11 assumed to be increased based on opinion. 12 13 All three reviews found effectiveness measures to also be substantially 14 heterogeneous. They were mostly disease specific, and ranged from QALYs 15 to ‘fractures avoided’. Some studies did not report an outcome at all. 16 The link between adherence and health benefit is important, however, was 17 described by both reviews as particularly problematic. Some analyses 18 provided no indication of the differences in health benefits which would be 19 observed when patients were non-adherent. Hughes found that, although 20 some studies were found to link non-adherence to changes in risk 21 probabilities or outcomes, only few referenced an evidence based source. Of 22 those studies which disclosed sources for values and assumptions, only very 23 few used sources other than opinion. Not surprisingly, none of the five studies 24 reviewed by Cleemput (2002) seemed to have linked adherence to clinically 25 relevant outcomes, such as health outcomes represented in QALYs. 26 27 The majority of studies conducted insufficient sensitivity analysis, particularly 28 of adherence rates. For example, Elliott (2005) found that of the 42 studies 29 reviewed only nine conducted sensitivity analysis to quantify the measure of 30 uncertainty which is an important methodological omission. Moreover, no 31 study assessed how the use of a particular adherence measure may have 32 affected ICERs or how sensitive the results were to changes in adherence 33 rates. This limits the generalisability of any results reported. Medicines concordance: full guideline DRAFT (July 2008) page 335 of 373 DRAFT FOR CONSULTATION 1 2 The reviews also noted other methodological problems of the studies 3 reported. Cleemput (2002) and Elliott (2005) found costing methods and 4 appropriate statistical analysis for cost data to be often invalid. Further 5 problems included stating the perspective of evaluation, appropriate 6 discounting and quantification of uncertainty, as well as incremental analysis. 7 8 The Elliott review concluded that a meaningful comparison across studies was 9 not possible due to heterogeneity. Hughes (2001) concluded that it was not 10 possible for the review to compare the magnitude of the impact of non- 11 adherence among different drug-disease combinations. Of the two studies 12 that specifically investigated interventions to increase adherence, reviewed by 13 Hughes (2001), one found that the intervention with a higher associated 14 adherence rate was cost effective compared to that with lower adherence, 15 whereas the other found the intervention with lower adherence to be cost 16 effective. Cleemput (2002) could not conclude on a certain approach to have 17 a clear advantage compared with another. Irrespective of the methodological 18 shortcomings, some of the reviewed studies on Interventions to increase 19 adherence show an improvement in terms of cost savings or reduced 20 noncompliance. They also found that one study on an intervention to increase 21 adherence for antimalarials and one for antihypertensives showed some 22 increase in the efficacy of treatment. However, the cost-effectiveness of an 23 intervention to increase adherence will depend upon the costs and health 24 effects associated with usual care and the intervention’s own costs and health 25 effects. 26 27 The systematic reviews did, however, raise some other important issues. 28 Hughes (2001) found that the nature of non-adherence, the severity and 29 pathophysiology of the disease, and the extent to which a drug ‘forgives’ to 30 non-adherence all contribute in determining the extent of the clinical and 31 economic consequences of non-adherence. Drug forgiveness describes the 32 ability of a drug to sustain its pharmacological action after a dose has been 33 missed. Medicines concordance: full guideline DRAFT (July 2008) page 336 of 373 DRAFT FOR CONSULTATION 1 In terms of the review findings, studies showed that non-adherence generally 2 results in a reduction in efficacy. Such a relationship was not as clear in cost 3 terms. While eight of the reviewed studies show that costs increase as 4 adherence decreases, six found the opposite trend. This discrepancy did not 5 appear to be related to the nature of the disease, the measure of non- 6 adherence or the assumptions relating to the health benefits experienced by 7 non-compliers. 8 The authors emphasise the need to systematically include non-adherence in 9 pharmacoeconomic evaluations. The vast majority of such studies were based 10 upon trials designed to demonstrate efficacy, and not effectiveness (i.e. 11 protocol effectiveness versus effectiveness in clinical practice). While the 12 randomised clinical trial remains the ‘gold standard’ for comparing alternative 13 treatments, the high internal validity required to demonstrate efficacy comes at 14 the expense of external validity, that is, generalisability of results to the ‘real 15 world’ in medical practice. 16 17 The review by Cleemput (2002) described the important nonlinearity in the 18 relationship between quality of life and non-adherence. Non-adherence may 19 improve patients’ quality of life, for instance when they deliberately adapt their 20 medication schedule to their own lifestyle, or it may decrease their quality of 21 life due to increased morbidity, adverse events and/or side-effects. From a 22 pharmacological perspective, under-dosing or extended time intervals 23 between two medication intakes may increase morbidity (and subsequently 24 costs), whereas over-dosing or shorter intervals between two medication 25 intakes may increase unpleasant side-effects or toxicity of the medication. 26 From an economic view, buying medication without actually taking it implies a 27 pure economic production cost. 28 The authors highlight the importance for clinicians, policymakers in healthcare 29 as well as patients to take non-adherence on the cost effectiveness of health 30 interventions into account. Methodologies to do so adequately need to be 31 improved and used in a standardised way. 32 Medicines concordance: full guideline DRAFT (July 2008) page 337 of 373 DRAFT FOR CONSULTATION 1 10.4.4 Conclusion 2 It is not possible to conclude which intervention to increase adherence is most 3 cost effective on the basis of the reviewed evidence. The disease specific 4 evaluation of interventions to increase adherence is important to establish 5 what works best for whom and in what circumstance. This is paramount for 6 the cost effectiveness profile of different interventions to increase adherence. 7 The reviews emphasised the importance of standardising the methods to take 8 non-adherence into account when assessing the effectiveness and cost 9 effectiveness of medicines. Moreover, the necessity to measure adherence 10 and to establish a link to clinical outcome appropriately were highlighted. The 11 need to distinguish persistence from compliance/adherence outcomes was 12 described by Hughes and they define medication persistence as “the length of 13 time from initiation to discontinuation of therapy and is measured in units of 14 time”(Hughes, Dyfrig, Cowell, W, Koncz, T et al , 2007). 15 The evidence from the three reviews confirmed that for the analysis of the 16 cost–effectiveness of Interventions to increase adherence it is paramount to 17 look at both costs of the intervention and outcomes, not only in terms of 18 adherence, but also in terms of the subjective value of the clinical outcome for 19 the patient. If the intervention offers a better health outcome than usual care 20 by effectively increasing adherence, or if its cost is more than offset by the 21 cost-savings from reduced non-adherence, the intervention is worthwhile. It is 22 unlikely to be possible to identify a single intervention to increase adherence 23 that fits all patients with their differing health situations, beliefs and 24 preferences in a generic format. 25 26 10.5 Update of the systematic review by Elliott 27 10.5.1 Methods of the review 28 The aim of this literature search is to update the systematic review by Elliott 29 (2005) and colleagues (Elliott, Barber, & Horne 2005) with relevant papers 30 published from 2004 onwards. The titles and abstracts of records retrieved by 31 the searches, suggested by the GDG or submitted by stakeholders were 32 scanned for relevance to the key questions. Any potentially relevant Medicines concordance: full guideline DRAFT (July 2008) page 338 of 373 DRAFT FOR CONSULTATION 1 publications were obtained in full text. This yielded eight papers, which were 2 then reviewed to identify the most appropriate evidence to help answer the 3 key questions and to ensure that the recommendations are based on the best 4 available evidence. This process required three main tasks: selection of 5 relevant studies; assessment of study quality; synthesis of the results. Three 6 of the eight references were subsequently excluded as they either failed to 7 describe and impute adherence specifically as outcome measure or did not 8 meet the methodological requirements. Five papers were reviewed. Evidence 9 reviews were undertaken using systematic, transparent approaches following 10 the Guidelines Manual 2007 (www.nice.org.uk). 11 Types of studies: Systematic reviews of cost effectiveness studies or 12 comparative economic analyses based on modelling or randomized controlled 13 trials (RCTs) of interventions to increase adherence. 14 Types of participants: people aged 16 and over prescribed medication for a 15 medical condition from healthcare professionals in any health service setting. 16 Duration of studies: No time limit was applied. 17 Types of interventions: any interventions intended to change adherence to 18 prescribed medication. 19 Types of outcome measures: adherence/compliance levels as well as clinical 20 outcome. 21 22 10.5.2 Update review narrative 23 Five papers were found and included. One Dutch paper by Bosmans 24 (Bosmans 2007) assessed the cost effectiveness of a pharmacy based 25 coaching programme to improve adherence to antidepressants. Another 26 Dutch paper by Cleemput (2004) (Cleemput et al. 2004) compared renal 27 transplantation with haemodialysis for patients with renal failure. A paper by 28 Brunenberg (2007) (Brunenberg-Danielle et al. 2007) examined the cost 29 effectiveness of an adherence improving programme comprising monitoring 30 system and adherence training for patients with hypertension receiving 31 statins. One US paper by Edwards (2005) {8837} assessed the cost 32 effectiveness of long acting risperidone compared to other oral agents in 33 patients with schizophrenia in a decision model. Finally Munakata (2006) Medicines concordance: full guideline DRAFT (July 2008) page 339 of 373 DRAFT FOR CONSULTATION 1 (Munakata et al. 2006) analysed in a theoretical scenario how much an 2 adherence enhancing intervention could cost for it to be cost effective when 3 provided for HIV positive patients who receive HAART. 4 5 The first Dutch based cost effectiveness analysis by Bosmans (2007) 6 (Bosmans, J. E., Brook, O. H., Van-Hout, H.-P. J. et al , 2007) was based on a 7 RCT of 151 patients with a prescription for non-tricyclic antidepressants from 8 their GP for depressive complaints. They were randomised to either receiving 9 an intervention consisting of three personal coaching contacts with a 10 pharmacist and an educational video to take home, or usual care including 11 standard oral and written information. 12 Adherence was measured using an electronic pill container (eDEM) and was 13 primary outcome, with the Hopkins depression 13 item subscale (SCL) used 14 as secondary outcome for depressive symptoms. Mean adherence did not 15 differ significantly between the intervention group (88%) and the control group 16 (86%) at six months (mean difference 2.1%, 95% CI -5.6, 9.8). In respect to 17 SCL subscale, there was no statistically significant difference between the 18 groups either despite a slight improvement in the pharmacist intervention 19 group (-0.15, 95% CI -0.54, 0.23). 20 The ICER for coaching and education by pharmacists compared with usual 21 care was €149 per 1% improvement in adherence and €2550 per point 22 improvement in the SCL depression mean item score. Uncertainty was 23 considerable, reflected by insignificance of mean differences. Pairs of costs 24 and effects were distributed in all four quadrants of the cost effectiveness 25 plane. The Cost effectiveness acceptability curve (CEAC) for adherence 26 showed extreme uncertainty, guiding decision makers to have little belief that 27 coaching and education by pharmacists is cost effective as a means of 28 increasing adherence to antidepressants compared with usual care. 29 30 The cost utility analysis by Cleemput (2004) (Cleemput, Irina, Kesteloot, 31 Katrien, Vanrenterghem, Yves et al , 2004) compared haemodialysis with 32 renal transplantation and was based on a decision analytic model. The model 33 drew on data from a prospective study of 126 adults with chronic renal failure Medicines concordance: full guideline DRAFT (July 2008) page 340 of 373 DRAFT FOR CONSULTATION 1 and varying adherence levels. Of these, 23 received renal transplant. 2 Adherence was measured using an electronic event monitoring (EEM) device. 3 Five study subjects were defined nonadherent with medication which gives a 4 rate of nonadherence of 21% in this population. 5 6 Lifetime costs after transplantation in the adherent patient group are higher 7 than lifetime costs in the non adherent group, mainly because adherent 8 patients live longer after transplantation. Compared with dialysis, renal 9 transplantation offers better outcome in both adherent and nonadherent 10 patients. Transplant dominated haemodialysis on all adherence levels and 11 was therefore found to be more cost effective. When full adherence is 12 assumed, transplant generates a cost saving relative of dialysis and 5.19 13 additional QALYs. In a heterogeneous group of adherent and nonadherent 14 patients, the saving was greater but fewer QALYs were generated (5.06). This 15 was mainly due to a reduced life expectancy. Among transplant patients, 16 adherence with immunosuppressants after transplantation is associated with a 17 QALY gain, albeit at a higher cost which was mainly due to a longer overall 18 life span. Mean costs per QALY in adherent patients relative to nonadherent 19 patients after transplantation amounted to €35 021 (95%CI 26 959 - 46 620). 20 In 2004, the authors concluded that this leaves scope for an effective 21 adherence enhancing intervention assuming a willingness to pay threshold 22 similar to that currently used by NICE. This study illustrates the effect 23 nonadherence can have on the findings of an economic evaluation. Assuming 24 full or good adherence, which seems common in RCTs, has the tendency to 25 overestimate cost effectiveness by producing more effect and fewer costs in a 26 scenario like this study. 27 28 The cost utility study by Brunenberg (2007) (Brunenberg-Danielle, E. M., 29 Wetzels-Gwenn, E. C., Nelemans, Patricia J. et al , 2007) compared a 30 medication events monitoring system MEMS) plus adherence training with 31 usual care alone over a 5 month follow up. The MEMS is a drug container and 32 cap equipped with a microchip that registers the date and time of each 33 opening. This study was based on a randomised controlled trial. There were Medicines concordance: full guideline DRAFT (July 2008) page 341 of 373 DRAFT FOR CONSULTATION 1 164 hypertensive patients in the MEMS arm and 89 in usual care group and 2 they had a systolic blood pressure (BP) >160mm Hg and/or diastolic BP 3 >95mm Hg despite being antihypertensive drug eligible. 4 Adherence was defined as intake minimum 85% of days as prescribed. From 5 the healthcare perspective, electronic monitoring led to a cost saving of €100 6 and an additional 3.1% of patients achieved normal blood pressure than in the 7 usual care arm and was therefore dominating. Sensitivity analysis revealed 8 considerable uncertainty. Using the QALY as an outcome measure, this 9 probability was between 45% and 51%, albeit from a societal perspective. 10 Overall, effect sizes were small and not statistically significant. From both, 11 healthcare and societal perspectives, the CEA bootstrap replicates on the CE 12 plane covered the origin. This means that the analysed pairs of cost and 13 effectiveness estimates result in both cost effective and not cost effective 14 decision rules, thereby disabling the decision maker to reach a certain 15 conclusion. 16 For statins as a long term treatment of hypertension, the length of follow up of 17 five months appears insufficiently short to anticipate the long term effect of the 18 intervention on adherence, and the effect of adherence on outcome. 19 Theoretically, if in the UK the costs for electronic monitoring do not exceed 20 those of a potential drug cost saving, even a moderate increase in adherence 21 would be cost effective. This would have to be proven in a different trial set 22 up. 23 24 Edwards (1999) (Edwards, G. and Anderson, I., 1999) conducted a cost 25 effective analysis based on a decision analytic model that compared long 26 acting risperidone with a range of other antipsychotic agents, including oral 27 risperidone and depot haloperidol. The population was drawn from patients 28 with schizophrenia in community dwellings who have previously suffered 29 relapse requiring hospitalisation. 30 Compliance was assumed to be improved by the long acting injectable 31 risperidone formula. It was estimated that a 20% point difference in 32 compliance would predict a 3.1 point improvement in the PANSS (Positive 33 and Negative Syndrome Scale for Schizophrenia). Such improvement in turn Medicines concordance: full guideline DRAFT (July 2008) page 342 of 373 DRAFT FOR CONSULTATION 1 stabilised patients so that a further 6.1 point in PANSS was achieved by 2 further improved medicine taking behaviour, and aversion of relapse. The 3 model predicts that patients receiving long acting risperidone will have the 4 best clinical outcomes in terms of the frequency and duration of relapses over 5 the one year duration. For example, on long acting risperidone 26% of 6 patients will experience relapse requiring hospitalisation and 24% relapse not 7 requiring hospitalisation. On haloperidol nearly two thirds of patients are 8 predicted to have relapses requiring hospitalisation and over 60% not 9 requiring hospitalisation. This analysis predicts dominance of long acting 10 risperidone over the comparators, with providing a health outcome 11 improvement in terms of days of relapse averted whilst costing less over the 12 time horizon of one year. Univariate sensitivity analysis was reported to have 13 been robust. However, at the upper bound of the 95%CI for relapse rates 14 requiring hospitalisation there was an incremental cost for long acting 15 risperidone with an ICER of US$821per days of hospitalisation averted 16 compared to oral risperidone. The model seems also sensitive to the cost of 17 hospitalisation as well as frequency rates of relapse. 18 Overall, the analysis seems of interest; however, there are issues with its 19 robustness. Adherence and clinical outcomes were estimated and not 20 quantified in a single measure, such as the QALY. The outcome of cost per 21 days of hospitalisation averted poses a challenge on the interpretation of the 22 findings in the context of this guideline. Values used in the sensitivity analysis 23 seem relatively conservative. The short time horizon could be an issue and 24 has not been thoroughly discussed. Quantifying treatment effect and quality of 25 life losses in one measurement such as the QALY could considerably help 26 interpret the findings from the analysis. 27 28 The cost utility analysis conducted by Munakata (2006) (Munakata, J., 29 Benner, J. S., Becker, S. et al , 2006) was based on a decision analytic 30 model. The aim was to quantify the clinical and economic effects of 31 nonadherence and estimate the cost effectiveness of improving adherence in 32 treatment naïve patients. For this, HAART treatment with an assumed good 33 adherence was compared with HAART on ‘typical’ adherence. The authors Medicines concordance: full guideline DRAFT (July 2008) page 343 of 373 DRAFT FOR CONSULTATION 1 drew on data from randomised controlled trials and observational data for the 2 comparators, respectively. The model population was HIV positive, with a 3 mean age of 33. The assumed portion of drugs consumed of 0.98 (0.95-1.0) 4 was defined as adherent and 0.55 (0-0.95) as nonadherent. The proportion of 5 adherent patients in the typical comparator arm was imputed as 0.52 (0.3- 6 0.88). Lifetime discounted costs in the typical and ideal scenarios were $308 7 000 and $341 000, respectively. This gives an incremental cost of $33 000. 8 People in the ideal scenario generated 10.2 QALYs per patient compared to 9 9.0 QALYs per patient in the typical scenario. This gives an incremental effect 10 of 1.2 QALYs. The incremental cost effectiveness ratio (ICER) resulted in $29 11 400 per QALY. This means that there is scope for an intervention to increase 12 adherence. The authors calculated a willingness to pay (WTP) ceiling value 13 for an intervention to increase adherence. They conclude that $1 600 could be 14 spent per patient to increase adherence to ideal levels, giving 15-33% 15 reductions in treatment failure. Univariate sensitivity analysis was conducted 16 for all parameters, as well as multivariate SA for selected values. The analysis 17 was described as robust in sensitivity analysis. In severe diseases where 18 adherence and related comorbidities are a big issue, adherence improving 19 interventions may be cost effective. Given that there are interventions that are 20 effective in increasing adherence, this analysis found that $1 600 per patient 21 could be spent. 22 23 Please refer to the discussion section 10.6 for a summary of the evidence. Medicines concordance: full guideline DRAFT (July 2008) page 344 of 373 DRAFT FOR CONSULTATION 1 2 10.6 3 It has been shown that the relationship between cost effectiveness and 4 adherence is not a simple a priori relationship. An increase in adherence does 5 not necessarily lead to an improvement in the cost effectiveness profile, even 6 for highly cost-effective interventions. 7 In order to make any predictions on the cost effectiveness of an intervention to 8 increase adherence (or indeed patient involvement as discussed) it will be 9 necessary to determine outcomes from direct evidence, comprising both 10 Discussion harms and benefits. 11 12 10.6.1 Theoretical discussion of review findings, caveats and 13 opportunities for future research 14 The presented findings from the studies selected for this review update 15 underpins the discussion from the previous evidence narrative of the three 16 systematic reviews (Cleemput, I, Kesteloot, K, and DeGeest, S, 2002) (Elliott, 17 Rachel A., Barber, Nick, and Horne, Rob, 2005) (Hughes, D. A., Bagust, A., 18 Haycox, A. et al , 2001). This states that, where nonadherence is prevalent 19 and leads to reductions in quality of life and even survival, there is scope for 20 an intervention that effectively raises adherence. However, it could not be 21 concluded which intervention could increase adherence appropriately to be 22 cost effective. This update review gives examples such as HAART and renal 23 transplant patients where there may be scope for such interventions, provided 24 they are effective in increasing adherence. 25 26 The reviews have revealed areas where there are particular methodological 27 difficulties/caveats in this area of research. Firstly, it became evident that 28 there is an important distinction between efficacy and effectiveness. Clinical 29 trials may demonstrate higher adherence rates than would be found in clinical 30 practice. This may be due to bias or length of follow up, particularly for 31 medicines where medication is indicated for the long term (adherence to 32 statins, for example, may be significantly lower after four years rather than Medicines concordance: full guideline DRAFT (July 2008) page 345 of 373 DRAFT FOR CONSULTATION 1 three months after initiation, as often measured in clinical trials (Avorn, J., 2 Monette, J., Lacour, A. et al , 1998). Limitations to follow up result from a 3 number of factors, including funding. There have been suggestions for 4 research to overcome the problem of follow up in form of persistence 5 measures (Elliott, R. A., Shinogle, J. A., Peele, P. et al , 2008) (Hughes, 6 Dyfrig, Cowell, W, Koncz, T et al , 2007). This will be discussed in more detail 7 below. 8 9 Secondly, the review of three systematic reviews presented in the previous 10 section found ‘forgiveness’ in connection with severity and pathophysiology of 11 the disease an important factor in determining the cost effectiveness of an 12 intervention to increase adherence. As described in section 10.3.2, a drug is 13 ‘forgiving’ if it is still efficacious despite missed doses. In such situations the 14 scope for an intervention is reduced. For example, statins are considered to 15 be relatively forgiving. From the update review, the study of monitoring 16 devices for the use of statins in hypertensive patients confirmed a limited use 17 of adherence enhancing measures in terms of QALY gain and cost 18 effectiveness (Brunenberg-Danielle, E. M., Wetzels-Gwenn, E. C., Nelemans, 19 Patricia J. et al , 2007). Measuring and reporting adherence as well as health 20 related outcomes is paramount for research to be interpreted meaningfully. 21 22 Thirdly, process utility was not measured in any of the studies reviewed, but 23 there is likely to be an effect on the quality of life by the process of taking the 24 medication. Please refer to Section 10.3 for some further detail on process 25 utility. 26 27 Despite all problems and caveats, the area of medicine taking is an interesting 28 and important one to investigate. As some studies in the review have shown, 29 there is scope to provide cost effective interventions to increase adherence. 30 Firstly, the capacity to benefit is evident to be of importance here. Where not 31 taking medication has likely severe effects on quality of life perhaps via effect 32 of the illness (e.g. in antidepressants) or comorbidities (e.g. in HIV AIDS 33 defining illnesses) and even survival (e.g. HIV, Renal transplants), effective Medicines concordance: full guideline DRAFT (July 2008) page 346 of 373 DRAFT FOR CONSULTATION 1 interventions to enhance adherence are more likely to be cost effective. The 2 papers by Cleemput et al. (Cleemput, Irina, Kesteloot, Katrien, 3 Vanrenterghem, Yves et al , 2004) and Munakata et al. (Munakata, J., 4 Benner, J. S., Becker, S. et al , 2006) provide examples for this. 5 6 Secondly, an effective intervention to increase adherence is more likely to be 7 cost effective where even only few missed doses of an ‘unforgiving’ have 8 more immediate consequences. Munakata et al. (Munakata, J., Benner, J. S., 9 Becker, S. et al , 2006) suggested that for people taking HAART medication, 10 there is scope to provide effective programmes to help patients take their 11 medication. Other examples could be antiepileptic or immunosuppressant 12 drugs. 13 Thirdly, externalities may widen the role of interventions to increase 14 adherence to a societal level. In economics, externalities (or spill over effects) 15 occur when firms or people impose costs or benefits on others outside the 16 market place (Samuelson, Paul A. and William D Nordhaus, 2004). An 17 example of this occurs in communicable diseases such as for example 18 tuberculosis. The social value curve for high adherence to active TB treatment 19 may deviate from the individual value curve for some patients as they may 20 experience a situation under treatment they would not personally have 21 chosen. An example of negative externality is antibiotic resistance which 22 occurs with poor adherence to TB treatment, and we see the devastating 23 effect it can have in some Baltic states and Latin American countries (Pablos- 24 Mendez, Ariel, Raviglione, Mario C., Laszlo, Adalbert et al , 1998). 25 26 27 10.6.2 Recommendations for minimum reporting requirements 28 The review of interventions to increase adherence has shown limitations of 29 the reporting style from studies. Often, reporting was inconsistent, incomplete 30 and/or not validated. In accordance with the NICE technical manual for 31 guideline development, we recommend the following data to be reported for 32 health economic analyses: 33 Medicines concordance: full guideline DRAFT (July 2008) page 347 of 373 DRAFT FOR CONSULTATION 1 2 3 1. Adequate time horizons and length of follow up need to be considered and reported. 2. Perspective of the evaluation needs to be reported. Where perspective not 4 healthcare, all relevant healthcare and social costs need to be presented 5 separately 6 3. Instead of presenting total costs per patient, microlevel costing and 7 providing mean resource use and unit costs is preferred in order to be able 8 to adjust for the prevailing setting. 9 4. Incremental analysis is used in health economics to reveal the value of 10 adopting a decision to provide a programme over and above an 11 alternative. 12 5. To report health related outcome, including all related side effects and co- 13 morbidities. The QALY is the preferred choice as it combines all health 14 effects into one measure, thereby allowing comparisons across 15 programmes. 16 6. Essentially, a definition of adherence needs to be stated, justified and used 17 consistently. It would be useful to develop standardised definitions for 18 specific formulations. For long term follow up, a binary measure such as 19 persistence may be most useful, particularly in more forgiving treatment 20 options (Stason, W. B., 1999) (Hughes, Dyfrig, Cowell, W, Koncz, T et al , 21 2007). Beware of censoring when using persistence measures, this 22 phenomenon can often be found with other methods of survival analysis 23 (Lamiraud, Karine and Moatti, Jean Paul, 2006) 24 7. A modelling approach may be of value in order to reflect the complexity of 25 concordance and adherence behaviour. Where decision trees oversimplify, 26 cohort based Markov modelling and patient-level discrete event simulation 27 may be of more valuable methodologies (Hughes, Dyfrig, Cowell, W, 28 Koncz, T et al , 2007) Medicines concordance: full guideline DRAFT (July 2008) page 348 of 373 DRAFT FOR CONSULTATION 1 8. Probabilistic sensitivity analysis should be carried out in order to estimate 2 uncertainty around incremental estimates and robustness of assumptions. 3 (Claxton, K., Sculpher, M., McCabe, C. et al , 2005). 4 9. Heterogeneity of intervention needs to be addressed appropriately by 5 providing an in-depth description of all components of the intervention. 6 External validity and replicability issues need to be adequately considered. 7 10. Heterogeneity of the patient population needs to be addressed by 8 providing an in-depth description of the study population. This is of great 9 importance for considering validity and replicability issues, especially since 10 medicine taking is an individual behaviour and qualitative research 11 approaches could be advocated. 12 Recommendations for future research 13 10.6.3 14 Further from the minimum reporting requirements set out in the previous 15 section, there are methodological options that the research community could 16 utilise to address the specific complexities described in this document more 17 adequately. 18 19 Minimum reporting recommendations would give scope for doing 20 economic modelling in the specific area. A decision model could 21 directly analyse the effect of changing intervention adherence levels on 22 outcome and cost effectiveness. A price ceiling for the intervention 23 could be established, and the uncertainty and robustness of single 24 parameters, such as rate of relapse, age etc evaluated by means of 25 formal sensitivity analysis. 26 27 Interventions to increase adherence involve a complex set of 28 components that may be difficult to incorporate into standard cost 29 effectiveness models. More research is needed to identify and assess 30 their components and how they can lead to improvements in quality of 31 life and/or a reduction in costs. It has been suggested that in the field of Medicines concordance: full guideline DRAFT (July 2008) page 349 of 373 DRAFT FOR CONSULTATION 1 adherence research endogeneity and interdependence play an 2 important role that require careful consideration for sound statistical 3 and econometric analysis (Lamiraud, Karine and Moatti, Jean Paul, 4 2006). They also say that using a microeconomic perspective could 5 help to better understand the complex trade offs that patients have to 6 face when deciding about heir level of adherence with medical 7 prescriptions. Further from describing the potential implications from 8 variable endogeneity and complexity to understand the dimensions of 9 behaviour change in both HCP and patient, pioneering research to 10 assess GP behaviour change has drawn upon psychometric tests to 11 assess GP change, as well as testing if actual behaviour has changed 12 (Bonetti, Debbie, Eccles, Martin, Johnston, Marie et al , 2005) (Hrisos, 13 S., Eccles, M., Johnston, M. et al , 2008a) (Hrisos, S., Eccles, M., 14 Johnston, M. et al , 2008b). Incorporated into a model, such data would 15 allow retrospective analysis in order to understand why an intervention 16 worked or did not work. 17 18 The binary measure of treatment persistence could enable long term 19 follow up after clinical trials have stopped, and thereby help determine 20 long term cost effectiveness of an intervention to enhance adherence. 21 This would be of particular interest in treatment situations where 22 treatments affect outcomes over long periods of time and/or where 23 medicines have to be taken for long periods of time and adherence 24 rates tend to drop over time in practice. Please see also section 10.3 25 where the concepts of efficiency and effectiveness were contrasted in 26 this context. 27 28 The evaluation of process utility may be important for evaluating the 29 benefits of shared decision making as well as adherence behaviour. 30 Therefore, further research in this area could play an important role as 31 insight into the individual gain other than health related from a health 32 care programme may change the outcome in terms of cost 33 effectiveness. This has not yet been considered by NICE as robust Medicines concordance: full guideline DRAFT (July 2008) page 350 of 373 DRAFT FOR CONSULTATION 1 methodology is absent, however, as a field of research changes in 2 process utility could be of particular importance for evaluating the 3 benefits and costs of shared decision making. 4 5 6 7 8 9 10 11 12 Medicines concordance: full guideline DRAFT (July 2008) page 351 of 373 DRAFT FOR CONSULTATION Reference List (1) Adam BD, Maticka TE, Cohen JJ. Adherence practices among people living with HIV. AIDS Care: Psychological Socio Medical Aspects of AIDS/HIV. 2003; 15 (2):263-274. (2) Adler DA, Bungay KM, Wilson IB, Pei Y et al. The impact of a pharmacist intervention on 6-month outcomes in depressed primary care patients. Gen Hosp Psychiatry. 2004; 26 (3):199-209. (3) Agård A, Hermerén G, Herlitz J. When is a patient with heart failure adequately informed? A study of patients' knowledge of and attitudes toward medical information. Heart & lung : the journal of critical care. 2004; 33 (4):219-226. (4) Aikens JE, Nease-Donald-E-Jr, Klinkman MS. Explaining patients' beliefs about the necessity and harmfulness of antidepressants. Annals of family medicine. 2008; 6 (1):23-29. (5) Albus-Christian S-NS-BF. Preferences regarding medical and psychosocial support in HIV-infected patients. Patient Educ Couns. 2005; 56 (1):16-20. (6) Alfonso V, Bermbach N, Geller J, Montaner J-SG. Individual variability in barriers affecting people's decision to take HAART: A qualitative study identifying barriers to being on HAART. Aids Patient Care STDS. 2006; 20 (12):848-857. (7) Antoni MH, Carrico AW, Duran RE, Spitzer S et al. Randomized clinical trial of cognitive behavioral stress management on human immunodeficiency virus viral load in gay men treated with highly active antiretroviral therapy. Psychosom Med. 2006; 68 (1):143-151. (8) Aristides M, Weston AR, FitzGerald P, Le RC et al. Patient preference and willingness-to-pay for Humalog Mix25 relative to Humulin 30/70: A multicountry application of a discrete choice experiment. Value Health. 2004; 7 (4):442-454. (9) Aronson B. Medication management behaviors of adherent short-term antibiotic users. Clinical Excellence for Nurse Practitioners. 2005; 9 (1):23-30. (10) Astrom K, Carlsson J, Bates I, Webb DG et al. Desire for information about drugs. A multi-method study in general medical inpatients. Pharm World Sci. 2000; 22 (4):159-164. Medicines concordance: full guideline DRAFT (July 2008) page 352 of 373 DRAFT FOR CONSULTATION (11) Atella V, Schafheutle E, Noyce P, Hassell K. Affordability of medicines and patients' cost-reducing behaviour: empirical evidence based on SUR estimates from Italy and the UK. Appl Health Econ Health Policy. 2005; 4 (1):23-35. (12) Atherton-Naji A, Hamilton R, Riddle W, Naji S. Improving adherence to antidepressant drug treatment in primary care: A feasibility study for a randomized controlled trial of education intervention. Primary Care Psychiatry. 2001; 7 (2):61-67. (13) Avorn J, Monette J, Lacour A, Bohn RL et al. Persistence of use of lipidlowering medications: a cross-national study. JAMA. 1998; 279 (18):14581462. (14) Bailey WC, Richards-Jr JM, Brooks CM, Soong S et al. A randomized trial to improve self-management practices of adults with asthma. Arch Intern Med. 1990; 150 (8):1664-1668. (15) Bain D. Doctor-patient communication in general practice consultations. Medical Education 10, 125-131. 1976. (16) Baird MG, Bentley-Taylor MM, Carruthers SG, Dawson KG et al. A study of efficacy, tolerance and compliance of once-daily versus twice-daily metoprolol (Betaloc) in hypertension. Betaloc Compliance Canadian Cooperative Study Group. Clin Invest Med. 1984; 7 (2):95-102. (17) Bambra C, Egan M, Thomas S, Petticrew M et al. The psychosocial and health effects of workplace reorganisation. 2. A systematic review of task restructuring interventions. J Epidemiol Community Health. 2007; 61 (12):1028-1037. (18) Bane C, Hughes CM, Cupples ME, McElnay JC. The journey to concordance for patients with hypertension: a qualitative study in primary care. Pharmacy world & science : PWS. 2007; 29 (5):534-540. (19) Bangalore S, Kamalakkannan G, Parkar S, Messerli FH. Fixed-dose combinations improve medication compliance: a meta- analysis. Am J Med. 2007; 120 :713-719. (20) Beaucage K, Lachance-Demers H, Ngo TT, Vachon C et al. Telephone follow-up of patients receiving antibiotic prescriptions from community pharmacies. American Journal of Health-System Pharmacy. 2006; 63 (6):557563. (21) Bechdolf A, Knost B, Kuntermann C, Schiller S et al. A randomized comparison of group cognitive-behavioural therapy and group psychoeducation in patients with schizophrenia. Acta Psychiatr Scand. 2004; 110 (1):21-28. (22) Bechdolf A, Köhn D, Knost B, Pukrop R et al. A randomized comparison of group cognitive-behavioural therapy and group psychoeducation in acute patients with schizophrenia: Outcome at 24 months. Acta Psychiatr Scand. 2005; 112 (3):173-179. Medicines concordance: full guideline DRAFT (July 2008) page 353 of 373 DRAFT FOR CONSULTATION (23) Becker LA, Glanz K, Sobel E. A randomized trial of special packaging of antihypertensive medications. J Fam Pract. 1986; 22 (4):357-361. (24) Begley S, Livingstone C, Hodges N, Williamson V. Impact of domiciliary pharmacy visits on medication management in an elderly population. International Journal of Pharmacy Practice. 1997; 5 (3):111-121. (25) Bender B, Milgrom H, Rand C. Nonadherence in asthmatic patients: is there a solution to the problem? Annals of allergy asthma & immunology : official publication of the American College of Allergy Asthma & Immunology. 1997; 79 (3):177-186. (26) Bernsten C, Bjorkman I, Caramona M, Crealey G et al. Improving the wellbeing of elderly patients via community pharmacy-based provision of pharmaceutical care. Drugs Aging. 2001; 18 (1):63-77. (27) Beusterien KM, Dziekan K, Flood E, Harding G et al. Understanding patient preferences for HIV medications using adaptive conjoint analysis: feasibility assessment. Value Health. 2005; 8 (4):453-461. (28) Bikowski RM, Ripsin CM, Lorraine VL. Physician-patient congruence regarding medication regimens. J Am Geriatr Soc. 2001; 49 (10):1353-1357. (29) Bollini P, Tibaldi G, Testa C, Munizza C. Understanding treatment adherence in affective disorders: a qualitative study. J Psychiatr Ment Health Nurs. 2004; 11 (6):668-674. (30) Bonetti D, Eccles M, Johnston M, Steen N et al. Guiding the design and selection of interventions to influence the implementation of evidence-based practice: an experimental simulation of a complex intervention trial. Social Science & Medicine. 2005; 60 (9):2135-2147. (31) Bosch C, X, Garner P. Contracts between patients and healthcare practitioners for improving patients' adherence to treatment, prevention and health promotion activities. 2006. (32) Bosch-Capblanch, X., Abba, K., Prictor, M., and Garner, P. Contracts between patients and healthcare practitioners for improving patients' adherence to treatment, prevention and health promotion activities. 2007 (33) Bosmans JE, Brook OH, Van-Hout H-PJ, De-Bruijne MC et al. Cost effectiveness of a pharmacy-based coaching programme to improve adherence to antidepressants. Pharmacoeconomics. 2007; 25 (1):25-37. (34) Braman AC. Patient personality predicts preference for relationships with doctors. Personality and Individual Differences. 2004; 37 (4):815-826. (35) Brouwer WBF, Exel NJ, Berg Bvd, Bos GAM et al. Process utility from providing informal care: the benefit of caring. Health Policy. 2005; 74 (1):8599. Medicines concordance: full guideline DRAFT (July 2008) page 354 of 373 DRAFT FOR CONSULTATION (36) Brown BG, Bardsley J, Poulin D, Hillger LA et al. Moderate dose, three-drug therapy with niacin, lovastatin, and colestipol to reduce low-density lipoprotein cholesterol <100 mg/dl in patients with hyperlipidemia and coronary artery disease. Am J Cardiol. 1997; 80 (2):111-115. (37) Brown C, Battista DR, Bruehlman R, Sereika SS et al. Beliefs about antidepressant medications in primary care patients: relationship to selfreported adherence. Med Care. 2005; 43 (12):1203-1207. (38) Brunenberg-Danielle EM, Wetzels-Gwenn EC, Nelemans PJ, Dirksen CD et al. Cost effectiveness of an adherence-improving programme in hypertensive patients. Pharmacoeconomics. 2007; 25 (3):239-251. (39) Campero L, Herrera C, Kendall T, Caballero M. Bridging the gap between antiretroviral access and adherence in Mexico. Qual Health Res. 2007; 17 :599-611. (40) Canto-De-Cetina TE, Canto P, Ordoñez LM. Effect of counseling to improve compliance in Mexican women receiving depot-medroxyprogesterone acetate. Contraception. 2001; 63 (3):143-146. (41) Caress A. Patient roles in decision-making. Nurs Times. 1997; 93 (31):45-48. (42) Carrick R, Mitchell A, Powell RA, Lloyd K. The quest for well-being: a qualitative study of the experience of taking antipsychotic medication. Psychology and psychotherapy. 2004; 77 (1):19-33. (43) Chaplin R, Kent A. Informing patients about tardive dyskinesia. Controlled trial of patient education. Br J Psychiatry. 1998; 172 (Jan):78-81. (44) Charles C, Gafni A, Whelan T. Decision-making in the physician-patient encounter: revisiting the shared treatment decision-making model. Soc Sci Med. 1999; 49 (5):651-661. (45) Charles C, Gafni A, Whelan T. Shared decision-making in the medical encounter: what does it mean? (or it takes at least two to tango). Social Science & Medicine. 1997; 44 (5):681-692. (46) Chisholm MA, Mulloy LL, Jagadeesan M, DiPiro JT. Impact of clinical pharmacy services on renal transplant patients' compliance with immunosuppressive medications. Clin Transplant. 2001; 15 (5):330-336. (47) Claxton AJ, Cramer J, Pierce C. A systematic review of the associations between dose regimens and medication compliance. Clin Ther. 2001; 23 (8):1296-1310. (48) Claxton A, Cramer J, Pierce C. A systematic Review of the Associations Between Dose Regimens and Medication Compliance. Clinical Therapeutics 23[8], 1296. 2001. Medicines concordance: full guideline DRAFT (July 2008) page 355 of 373 DRAFT FOR CONSULTATION (49) Claxton K, Sculpher M, McCabe C, Briggs A et al. Probabilistic sensitivity analysis for NICE technology assessment: not an optional extra. Health Econ. 2005; 14 (4):339-347. (50) Cleemput I, Kesteloot K, DeGeest S. A review of the literature on the economics of noncompliance. Room for methodological improvement. Health Policy. 2002; 59 (1):65-94. (51) Cleemput I, Kesteloot K, Vanrenterghem Y, De GS. The economic implications of non-adherence after renal transplantation. Pharmacoeconomics. 2004; 22 (18):1217-1234. (52) Cohen D, Longo MF, Hood K, Edwards A et al. Resource effects of training general practitioners in risk communication skills and shared decision making competences. Journal of Evaluation in Clinical Practice. 2004; 10 (3):439-445. (53) Collier AC, Ribaudo H, Mukherjee AL, Feinberg J et al. A randomized study of serial telephone call support to increase adherence and thereby improve virologic outcome in persons initiating antiretroviral therapy. J Infect Dis. 2002; 192 (8):1398-1406. (54) Connor J, Rafter N, Rodgers A. Do fixed-dose combination pills or unit-of-use packaging improve adherence? A systematic review. [Review] [26 refs]. Bulletin of the World Health Organization. 2004; 82 (12):935-939. (55) Cooper V, Buick D, Horne R, Lambert N et al. Perceptions of HAART among gay men who declined a treatment offer: preliminary results from an interview-based study. AIDS Care. 2002; 14 (3):319-328. (56) Cote J, Bowie DM, Robichaud P, Parent JG et al. Evaluation of two different educational interventions for adult patients consulting with an acute asthma exacerbation. Am J Respir Crit Care Med. 2001; 163 (6):1415-1419. (57) Côté J, Bowie DM, Robichaud P, Parent JG et al. Evaluation of two different educational interventions for adult patients consulting with an acute asthma exacerbation. Am J Respir Crit Care Med. 2001; 163 (6):1415-1419. (58) Coté J, Cartier A, Robichaud P, Boutin H et al. Influence on asthma morbidity of asthma education programs based on self-management plans following treatment optimization. Am J Respir Crit Care Med. 1997; 155 (5):1509-1514. (59) Cox, F, Stevenson, N, Brtten, N, and Dundar, Y. A systematic review of communication between patients and health care professionals about medicine-taking and prescribing. Medicines partnership. 2004 (60) Cox K, Britten N, Hooper R, White P. Patients' involvement in decisions about medicines: GPs' perceptions of their preferences. Br J Gen Pract. 2007; 57 (543):777-784. Medicines concordance: full guideline DRAFT (July 2008) page 356 of 373 DRAFT FOR CONSULTATION (61) Cramer JA. A systematic review of adherence with medications for diabetes. Diabetes Care. 2004; 27 (5):1218-1224. (62) De G, Schafer-Keller P, Denhaerynck K, Thannberger N et al. Supporting medication adherence in renal transplantation (SMART): a pilot RCT to improve adherence to immunosuppressive regimens. Clinical Transplantation. 2003; #2006 May-Jun. (3). (63) Deegan PE. The importance of personal medicine: A qualitative study of resilience in people with psychiatric disabilities. Scandinavian Journal of Public Health. 2005; 33 (SUPPL. 66):29-35. (64) Degner L. Decision making during serious illness: What role do patients really want to play?. Sloan J, editor. Journal of Clinical Epidemiology 45, 941-950. 2008. (65) Doherty C, Doherty W. Patients' preferences for involvement in clinical decision-making within secondary care and the factors that influence their preferences. J Nurs Manag. 2005; 13 (2):119-127. (66) Dowell J, Jones A, Snadden D. Exploring medication use to seek concordance with 'non-adherent' patients: a qualitative study. The British journal of general practice : the journal of the Royal College of General Practitioners. 2002; 52 (474):24-32. (67) Duggan C, Bates I. Development and evaluation of a survey tool to explore patients' perceptions of their prescribed drugs and their need for drug information. International Journal of Pharmacy Practice. 2000; 8 (1):42-52. (68) Duggan PS, Geller G, Cooper LA, Beach MC. The moral nature of patientcenteredness: is it "just the right thing to do"?. [Review] [36 refs]. Patient Education & Counseling. 2006; 62 (2):271-276. (69) Dunbar J, Dunning EJ, Dwyer K. Compliance measurement with arthritis regimen. Arthritis care and research : the official journal of the Arthritis Health Professions Association. 1989;(3):S8-16. (70) Edwards A, Elwyn G. Inside the black box of shared decision making: distinguishing between the process of involvement and who makes the decision. Health Expect. 2006; 9 (4):307-320. (71) Edwards A, Elwyn G, Hood K, Atwell C et al. Patient-based outcome results from a cluster randomized trial of shared decision making skill development and use of risk communication aids in general practice. Fam Pract. 2004; 21 (4):347-354. (72) Edwards G, Anderson I. Systematic review and guide to selection of selective serotonin reuptake inhibitors. Drugs. 1999; 57 (4):507-533. (73) Elliott RA, Shinogle JA, Peele P, Bhosle M et al. Understanding Medication Compliance and Persistence from an Economics Perspective. Value Health. 2008. Medicines concordance: full guideline DRAFT (July 2008) page 357 of 373 DRAFT FOR CONSULTATION (74) Elliott RA, Barber N, Horne R. Cost-effectiveness of adherence-enhancing interventions: a quality assessment of the evidence. The Annals of pharmacotherapy. 2005; 39 (3):508-515. (75) Elliott RA, Ross DD, Adams AS, Safran DG et al. Strategies for coping in a complex world: adherence behavior among older adults with chronic illness. J Gen Intern Med. 2007; 22 (6):805-810. (76) Elwyn G, Edwards A, Dwyn R, Grol R. Towards a feasible model for shared decision-making: focus group study wuth general practice registrars. British Medical Journal 319, 753-756. 1999. (77) Elwyn G, Edwards A, hutchings A, Rapport F et al. The OPTION scale:measuring the extent that clinicians involve patients in decision-making tasks. Health Expectations 8[1], 34-42. 2005. (78) Elwyn G, Edwards A, Mowle S, Wensing M et al. Measuring the involvement of patients in shared decision-making: a systematic review of instruments. [Review] [88 refs]. Patient Education & Counseling. 2001; 43 (1):5-22. (79) Elwyn G, Edwards A, Wensing M, Hood K et al. Shared decision making: developing the OPTION scale for measuring patient involvement. Qual Saf Health Care. 2003; 12 (2):93-99. (80) Ende J, Kazis L, Ash A, Moskowitz MA. Measuring patients' desire for autonomy: decision making and information-seeking preferences among medical patients. Journal of general internal medicine : official journal of the Society for Research and Education in Primary Care Internal Medicine. 1989; 4 (1):23-30. (81) Erwin J, Peters B. Treatment issues for HIV+ Africans in London. Soc Sci Med. 1999; 49 (11):1519-1528. (82) F Légaré, S Ratté D Stacey J Kryworuchko K Gravel L Turcot ID Graham. Interventions for improving the adoption of shared decision making by healthcare professionals. Cochrane Collaboration. 2008 (83) Farmer KC. Methods for measuring and monitoring medication regimen adherence in clinical trials and clinical practice. Clin Ther. 1999; 21 (6):-90. (84) Finley PR, Rens HR, Pont JT, Gess SL et al. Impact of a collaborative care model on depression in a primary care setting: a randomized controlled trial. Pharmacotherapy. 2003; 23 (9):1175-1185. (85) Fraenkel L, Rabidou N, Wittink D, Fried T. Improving informed decisionmaking for patients with knee pain. The Journal of rheumatology. 2007; 34 (9):1894-1898. (86) Friedman RH, Kazis LE, Jette A, Smith MB et al. A telecommunications system for monitoring and counseling patients with hypertension. Impact on medication adherence and blood pressure control. Am J Hypertens. 1996; 9 (4):285-292. Medicines concordance: full guideline DRAFT (July 2008) page 358 of 373 DRAFT FOR CONSULTATION (87) Frileux-Stéphanie S-M-T-MM-ES. The impact of the preventive medical message on intention to change behavior. Patient Educ Couns. 2004; 52 (1):79-88. (88) Gagné-Camille GG. Improving self-report measures of non-adherence to HIV medications. Psychology & Health. 2005; 20 (6):803-816. (89) Garber MC, Nau DP, Erickson SR, Aikens JE et al. The concordance of selfreport with other measures of medication adherence: a summary of the literature. Med Care. 2004; 42 (7):649-652. (90) Garcia-Popa-Lisseanu MG, Greisinger A, Richardson M, Malley KJ et al. Determinants of treatment adherence in ethnically diverse, economically disadvantaged patients with rheumatic disease. J Rheumatol. 2005; 32 (5):913919. (91) George J, Kong-David CM, Stewart K. Adherence to disease management programs in patients with COPD. International journal of chronic obstructive pulmonary disease. 2007; 2 (3):253-262. (92) Girvin B, McDermott BJ, Johnston GD. A comparison of enalapril 20 mg once daily versus 10 mg twice daily in terms of blood pressure lowering and patient compliance. J Hypertens. 1999; 17 (11):1627-1631. (93) Girvin BG, Johnston GD. Comparison of the effects of a 7-day period of noncompliance on blood pressure control using three different antihypertensive agents. J Hypertens. 2004; 22 (7):1409-1414. (94) Goldring AB, Taylor SE, Kemeny ME, Anton PA. Impact of health beliefs, quality of life, and the physician-patient relationship on the treatment intentions of inflammatory bowel disease patients. Health Psychology. 2002; 21 (3):219-228. (95) Gravel K, Legare F, Graham ID. barriers and facilitators to implementing shared decision-making in clinical practice: a systematic review of health care professionals perceptions. implementation science. 2006; 1 (16):1-12. (96) Gray R, Leese M, Bindman J, Becker T et al. Adherence therapy for people with schizophrenia: European multicentre randomised controlled trial. Br J Psychiatry. 2006; 189 (DEC.):508-514. (97) Grymonpre RE, Williamson DA, Montgomery PR. Impact of a pharmaceutical care model for non-institutionalised elderly: Results of a randomised, controlled trial. International Journal of Pharmacy Practice. 2001; 9 (4):235241. (98) Guthrie RM. The effects of postal and telephone reminders on compliance with pravastatin therapy in a national registry: results of the first myocardial infarction risk reduction program. Clin Ther. 2001; 23 (6):970-980. (99) Hamann J, Langer B, Winkler V, Busch R et al. Shared decision making for in-patients with schizophrenia. Acta Psychiatr Scand. 2006; 114 (4):265-273. Medicines concordance: full guideline DRAFT (July 2008) page 359 of 373 DRAFT FOR CONSULTATION (100) Hamann J, Cohen R, Leucht S, Busch R et al. Shared decision making and long-term outcome in schizophrenia treatment. The Journal of clinical psychiatry. 2007; 68 (7):992-997. (101) Hamet P, Campbell N, Curnew G, Eastwood C et al. Avapromise: a randomized clinical trial for increasing adherence through behavioural modification in essential hypertension. Exp Clin Cardiol. 2003; 7 :165-172. (102) Hamilton W, Russell D, Stabb C, Seamark D et al. The effect of patient selfcompletion agenda forms on prescribing and adherence in general practice: a randomized controlled trial. Fam Pract. 2007; 24 (1):77-83. (103) Hanlon JT, Weinberger M, Samsa GP, Schmader KE et al. A randomized, controlled trial of a clinical pharmacist intervention to improve inappropriate prescribing in elderly outpatients with polypharmacy. The American journal of medicine. 1996; 100 (4):428-437. (104) Harrington-Jane NL. Improving patients' communication with doctors: A systematic review of intervention studies. Patient Educ Couns. 2004; 52 (1):716. (105) Hawkshead J, Krousel-Wood MA. Techniques for measuring medication adherence in hypertensive patients in outpatient settings: Advantages and limitations. Disease Management and Health Outcomes. 2007; 15 (2):109-118. (106) Hayes RP, Bowman L, Monahan PO, Marrero DG et al. Understanding diabetes medications from the perspective of patients with type 2 diabetes: prerequisite to medication concordance. The Diabetes educator. 2006; 32 (3):404-414. (107) Haynes RB, Ackloo E, Sahota N, McDonald HP et al. Interventions for enhancing medication adherence. Cochrane Database of Systematic Reviews. 2008;(CD000011). (108) Haynes RB, Sackett DL, Gibson ES. Improvement of medication compliance in uncontrolled hypertension. Lancet. 1976; 1 (7972):1265-1268. (109) Haynes RB, Yao X, Degani A, Kripalani S et al. Interventions to enhance medication adherence. Cochrane Database Syst Rev. 2005;(4):CD000011. (110) Haynes TB, Taylor DW, Sackett DL. Behavioural Strategies for improving compliance. In: Haynes TB, Taylor DW, Sackett DL. Compliance in Health Care. 174-190. 1979. John Hopkins University Press. (111) Hecht FM. Measuring HIV treatment adherence in clinical practice. AIDS clinical care. 1998; 10 (8):57-59. (112) Heneghan CJ, Glasziou P, Perera R. Reminder packaging for improving adherence to self-administered long-term medications. [Review] [86 refs]. Cochrane Database of Systematic Reviews (1):CD005025. 2006;(1). Medicines concordance: full guideline DRAFT (July 2008) page 360 of 373 DRAFT FOR CONSULTATION (113) Henry A, Batey RG. Enhancing compliance not a prerequisite for effective eradication of Helicobacter pylori: the HelP Study. Am J Gastroenterol. 1999; 94 (3):811-815. (114) Hill SA, Laugharne R. Decision making and information seeking preferences among psychiatric patients. J Ment Health. 2006; 15 (1):75-84. (115) Hirth RA, Greer SL, Albert JM, Young EW et al. Out-of-pocket spending and medication adherence among dialysis patients in twelve countries. Health Aff. 2008; 27 (1):89-102. (116) Honish A, Westerfield W, Ashby A, Momin S et al. Health-related quality of life and treatment compliance with diabetes care. DIS MANAGE. 2006; 9 (4):195-200. (117) Horne R, Weinman J. Concordance, A Partnership in Medicine Taking. 2004. The Pharmaceutical Press. (118) Horne R, Weinman J, Barber N, Elliott R et al. Concordance, adherence and compliance in medicine taking. Report for the National Co-ordinatingCentre for NHS Service Delivery andOrganisation R & D (NCCSDO . 2005. (119) Horne R. Beliefs about medicines questionnaire: the development and evaluation of a new method for assessing the cognitive representation of mediciation. Psychology and Health. 2008; 37 (2):113-124. (120) Horne R, Cooper V, Gellaitry G, Date HL et al. Patients' perceptions of highly active antiretroviral therapy in relation to treatment uptake and adherence: The utility of the necessity-concerns framework. J Acquir Immune Defic Syndr. 2007; 45 (3):334-341. (121) Howland JS, Baker MG, Poe T. Does patient education cause side effects? A controlled trial. J Fam Pract. 1990; 31 (1):62-64. (122) Hrisos S, Eccles M, Johnston M, Francis J et al. An intervention modelling experiment to change GPs' intentions to implement evidence-based practice: using theory-based interventions to promote GP management of upper respiratory tract infection without prescribing antibiotics #2. BMC Health Serv Res. 2008a; 8 :10. (123) Hrisos S, Eccles M, Johnston M, Francis J et al. Developing the content of two behavioural interventions: using theory-based interventions to promote GP management of upper respiratory tract infection without prescribing antibiotics #1. BMC Health Serv Res. 2008b; 8 :11. (124) Hughes DA, Bagust A, Haycox A, Walley T. The impact of non-compliance on the cost-effectiveness of pharmaceuticals: a review of the literature. [Review] [52 refs]. Health Economics. 2001; 10 (7):601-615. (125) Hughes D, Cowell W, Koncz T, Cramer J. Methods for integrating medication compliance and persistence in pharmacoeconomic evaluations. Value Health. 2007. Medicines concordance: full guideline DRAFT (July 2008) page 361 of 373 DRAFT FOR CONSULTATION (126) Iskedjian M, Einarson TR, MacKeigan LD, Shear N et al. Relationship between daily dose frequency and adherence to antihypertensive pharmacotherapy: Evidence from a meta-analysis. Clin Ther. 2002; 24 (2):302-316. (127) Jenkins L, Britten N, Stevenson F, Barber N et al. Developing and using quantitative instruments for measuring doctor- patient communication about drugs. Patient Educ Couns. 2003; 50 (3):273-278. (128) Johnson AL, Taylor DW, Sackett DL. Self-recording of blood pressure in the management of hypertension. Can Med Assoc J. 1978; 119 (9):1034-1039. (129) Johnson SB. Methodological issues in diabetes research. Measuring adherence. Diabetes Care. 1992; 15 (11):1658-1667. (130) Katon W, Russo J, Von KM, Lin E et al. Long-term effects of a collaborative care intervention in persistently depressed primary care patients. J Gen Intern Med. 2002; 17 (10):741-748. (131) Kellett N, West F, Finlay AY. Conjoint analysis: a novel, rigorous tool for determining patient preferences for topical antibiotic treatment for acne. A randomised controlled trial. The British journal of dermatology. 2006; 154 (3):524-532. (132) Kemp R, Kirov G, Everitt B, Hayward P et al. Randomized controlled trial of compliance therapy. 18 month follow up. Br J Psychiatry. 1998; 172 :413-419. (133) Kemp R, Hayward P, Applewhaite G, Everitt B et al. Compliance therapy in psychotic patients: Randomised controlled trial. Br Med J. 1996; 312 (7027):345-349. (134) Kemp-Steven F-MH-AW-HC-C. Psychological factors and use of antiepileptic drugs: Pilot work using an objective measure of adherence. Psychology. 2007; 12 (1):107-113. (135) Khanderia U, Townsend KA, Erickson SR, Vlasnik J et al. Medication adherence following coronary artery bypass graft surgery: Assessment of beliefs and attitudes. Ann Pharmacother. 2008; 42 (2):192-199. (136) Kikkert MJ, Schene AH, Koeter-Maarten WJ, Robson D et al. Medication adherence in schizophrenia: exploring patients', carers' and professionals' views. Schizophr Bull. 2006; 32 (4):786-794. (137) King MT, Hall J, Lancsar E, Fiebig D et al. Patient preferences for managing asthma: results from a discrete choice experiment. Health Economics. 2007; 16 :703-717. (138) Kinnersley P, Edwards A, Hood K, Cadbury N et al. Interventions before consultations for helping patients address their information needs. Cochrane database of systematic reviews (Online). 2007;(3):CD004565, 156CD004565,493X. Medicines concordance: full guideline DRAFT (July 2008) page 362 of 373 DRAFT FOR CONSULTATION (139) Kumar K, Gordon C, Toescu V, Buckley CD et al. Beliefs about medicines in patients with rheumatoid arthritis and systemic lupus erythematosus: a comparison between patients of South Asian and White British origin. Rheumatology (Oxford England). 2008; 47 (5):690-697. (140) La FJ. Methods to Measure Patient Compliance with Medication Regimens. JPain-Pall-Care-Pharmacother. 2004; 18 (3):81-87. (141) Lam DH, Watkins ER, Hayward P, Bright J et al. A randomized controlled study of cognitive therapy for relapse prevention for bipolar affective disorder: outcome of the first year. Arch Gen Psychiatry. 2003; 60 (2):145-152. (142) Lamiraud K, Moatti JP. Economic implications of nonadherence to highly active antiretroviral treatment in HIV patients. Expert opinion on pharmacotherapy. 2006; 7 (2):135-143. (143) Lancsar EJ, Hall JP, King M, Kenny P et al. Using discrete choice experiments to investigate subject preferences for preventive asthma medication. Respirology (Carlton, Vic ). 2007; 12 (1):127-136. (144) Langewitz W, Nübling M, Weber H. Hospital patients' preferences for involvement in decision-making: A questionnaire survey of 1040 patients from a Swiss university hospital. Swiss Medical Weekly. 2006; 136 (3-4):59-64. (145) Lawton J, Ahmad N, Hallowell N, Hanna L et al. Perceptions and experiences of taking oral hypoglycaemic agents among people of Pakistani and Indian origin: Qualitative study. Br Med J. 2005; 330 (7502):1247-1249. (146) Lee JK, Grace KA, Taylor AJ. Effect of a pharmacy care program on medication adherence and persistence, blood pressure, and low-density lipoprotein cholesterol: a randomized controlled trial.[see comment]. JAMA. 2006; 296 (21):2563-2571. (147) Levy ML, Robb M, Allen J, Doherty C et al. A randomized controlled evaluation of specialist nurse education following accident and emergency department attendance for acute asthma. Respir Med. 2000; 94 :900-908. (148) Lewin, S. A., Skea, Z. C., Entwistle, V., Zwarenstein, M. et al. Interventions for providers to promote a patient-centred approach in clinical consultations. 2001 (149) Lewis MP, Colbert A, Erlen J, Meyers M. A qualitative study of persons who are 100% adherent to antiretroviral therapy. AIDS Care - Psychological and Socio-Medical Aspects of AIDS/HIV. 2006; 18 (2):140-148. (150) Lipton HL, Bird JA. The impact of clinical pharmacists' consultations on geriatric patients' compliance and medical care use: a randomized controlled trial. The Gerontologist. 1994; 34 (3):307-315. (151) Little P, Dorward M, Warner G, Moore M et al. Randomised controlled trial of effect of leaflets to empower patients in consultations in primary care. Br Med J. 2004; 328 (7437):441-444. Medicines concordance: full guideline DRAFT (July 2008) page 363 of 373 DRAFT FOR CONSULTATION (152) Loh A, Simon D, Wills CE, Kriston L et al. The effects of a shared decisionmaking intervention in primary care of depression: a cluster-randomized controlled trial. Patient Educ Couns. 2007; 67 (3):324-332. (153) Lowe CJ, Raynor DK, Purvis J, Farrin A et al. Effects of a medicine review and education programme for older people in general practice. British journal of clinical pharmacology. 2000; 50 (2):172-175. (154) Lukoschek P. African Americans' beliefs and attitudes regarding hypertension and its treatment: a qualitative study. J Health Care Poor Underserved. 2003; 14 (4):566-587. (155) Mahadevia PJ, Shah S, Leibman C, Kleinman L et al. Patient preferences for sensory attributes of intranasal corticosteroids and willingness to adhere to prescribed therapy for allergic rhinitis: a conjoint analysis. Annals of allergy asthma & immunology : official publication of the American College of Allergy Asthma & Immunology. 2004; 93 (4):345-350. (156) Makoul G, Clayman ML. An integrative model of shared decision making in medical encounters. Patient Educ Couns. 2006; 60 (3):301-312. (157) Mannheimer SB, Morse E, Matts JP, Andrews L et al. Sustained benefit from a long-term antiretroviral adherence intervention: Results of a large randomized clinical trial. J Acquir Immune Defic Syndr. 2006; 43 (SUPPL. 1):S41-S47. (158) Mark TL, Swait J. Using stated preference and revealed preference modeling to evaluate prescribing decisions. Health Economics. 2004; 13 (6):563-573. (159) Marquez-Contreras E, Martell-Claros N, Gil-Guillen V, de lV et al. Efficacy of a home blood pressure monitoring programme on therapeutic compliance in hypertension: the EAPACUM-HTA study. Journal of Hypertension. 2006; 24 (1):169-175. (160) McCann-Simon WJ. Encouraging patient participation in general practice consultations: Effect on consultation length and content, patient satisfaction and health. Psychology & Health. 1996; 11 (6):857-869. (161) McKinstry, B., Ashcroft, R. E., Car, J., Freeman, G. K. et al. Interventions for improving patients' trust in doctors and groups of doctors. 2006 (162) McLean M, Armstrong D. Eliciting patients' concerns: a randomised controlled trial of different approaches by the doctor. British Journal of General Practice. 2004; 54 (506):663-666. (163) Mead N, Power P. Patient-centredness: a conceptual framework and review of the emprirical literature. Social Science & Medicine 51[7], 1087-1110. 2000. (164) Medical Research Council. A Framework for development and evaluation of RCTs for Complex Interventions to Improve Health. 2000. Medicines concordance: full guideline DRAFT (July 2008) page 364 of 373 DRAFT FOR CONSULTATION (165) Menckeberg TT, Bouvy ML, Bracke M, Kaptein AA et al. Beliefs about medicines predict refill adherence to inhaled corticosteroids. J Psychosom Res. 2008; 64 (1):47-54. (166) Middleton JF, McKinley RK, Gillies CL. Effect of patient completed agenda forms and doctors' education about the agenda on the outcome of consultations: randomised controlled trial. Br Med J. 2006; 332 (7552):12381242. (167) Miklowitz DJ, George EL, Richards JA, Simoneau TL et al. A randomized study of family-focused psychoeducation and pharmacotherapy in the outpatient management of bipolar disorder. Arch Gen Psychiatry. 2003; 60 (9):904-912. (168) Miller LG, Hays RD. Measuring adherence to antiretroviral medications in clinical trials. HIV Clin Trials. 2000; 1 (1):36-46. (169) Mills EJ. Adherence to Antiretroviral Therapy in Sub-Saharan Africa and North America: A Meta-Analysis. JAMA. 2006; 296 (6):679-690. (170) Mills EJ, Nachega JB, Bangsberg DR, Singh S et al. Adherence to HAART: a systematic review of developed and developing nation patient-reported barriers and facilitators. PLoS medicine. 2006; 3 (11):2039-2064. (171) Molina JM, Podsadecki TJ, Johnson MA, Wilkin A et al. A lopinavir/ritonavir-based once-daily regimen results in better compliance and is non-inferior to a twice-daily regimen through 96 weeks. AIDS Res Hum Retroviruses. 2007; 23 (12):1505-1514. (172) Montgomery AA, Fahey T, Peters TJ. A factorial randomised controlled trial of decision analysis and an information video plus leaflet for newly diagnosed hypertensive patients. The British journal of general practice : the journal of the Royal College of General Practitioners. 2003; 53 (491):446-453. (173) Morgan-Myfanwy F-M-J. Barriers to Uptake and Adherence with Malaria Prophylaxis by the African Community in London, England: Focus Group Study. Ethnicity & Health. 2005; 10 (4):355-372. (174) Morice AH, Wrench C. The role of the asthma nurse in treatment compliance and self- management following hospital admission. Respir Med. 2001; 95 (11):851-856. (175) Munakata J, Benner JS, Becker S, Dezii CM et al. Clinical and economic outcomes of nonadherence to highly active antiretroviral therapy in patients with human immunodeficiency virus. Med Care. 2006; 44 (10):893-899. (176) Munro SA, Lewin SA, Smith HJ, Engel ME et al. Patient adherence to tuberculosis treatment: a systematic review of qualitative research. PLoS medicine. 2007; 4 (7):1230-1245. Medicines concordance: full guideline DRAFT (July 2008) page 365 of 373 DRAFT FOR CONSULTATION (177) Murray E, Charles C, Gafni A. Shared decision-making in primary care: tailoring the Charles et al. model to fit the context of general practice. [Review] [32 refs]. Patient Education & Counseling. 2006; 62 (2):205-211. (178) Mutchler JE, Bacigalupe G, Coppin A, Gottlieb A. Language barriers surrounding medication use among older Latinos. Journal of cross cultural gerontology. 2007; 22 (1):101-114. (179) Nazareth I, Burton A, Shulman S, Smith P et al. A pharmacy discharge plan for hospitalized elderly patients--a randomized controlled trial. Age Ageing. 2001; 30 (1):33-40. (180) Neame R, Hammond A, Deighton C. Need for information and for involvement in decision making among patients with rheumatoid arthritis: a questionnaire survey. Arthritis Rheum. 2005; 53 :249-255. (181) O'Connor, A. M., Stacey, D., Entwistle, V., Llewellyn-Thomas, H. et al. Decision aids for people facing health treatment or screening decisions. 2003 (182) O'Donnell C, Donohoe G, Sharkey L, Owens N et al. Compliance therapy: a randomised controlled trial in schizophrenia. Br Med J. 2003; 327 (7419):834837. (183) Oakley S, Walley T. A pilot study assessing the effectiveness of a decision aid on patient adherence with oral bisphosphonate medication. Pharmaceutical Journal. 2006; 276 (7399):536-538. (184) Ogedegbe G, Harrison M, Robbins L, Mancuso CA et al. Barriers and facilitators of medication adherence in hypertensive African Americans: a qualitative study. Ethnicity & disease. 2004; 14 (1):3-12. (185) Orton L, Barnish G. Unit-dose packaged drugs for treating malaria. [Review] [40 refs]. Cochrane Database of Systematic Reviews. 2005;(CD004614). (186) Pablos-Mendez A, Raviglione MC, Laszlo A, Binkin N et al. Global Surveillance for Antituberculosis-Drug Resistance, 1994-1997. The New England journal of medicine. 1998; 338 (23):1641-1649. (187) Parienti JJ, Massari V, Reliquet V, Chaillot F et al. Effect of twice-daily nevirapine on adherence in HIV-1-infected patients: a randomized controlled study. AIDS. 2007; 21 (16):2217-2222. (188) Paterson DL, Potoski B, Capitano B. Measurement of adherence to antiretroviral medications. J Acquir Immune Defic Syndr. 2002; 31 (Suppl 3):S103-S106. (189) Peterson GM, Fitzmaurice KD, Naunton M, Vial JH et al. Impact of pharmacist-conducted home visits on the outcomes of lipid-lowering drug therapy. J Clin Pharm Ther. 2004; 29 (1):23-30. Medicines concordance: full guideline DRAFT (July 2008) page 366 of 373 DRAFT FOR CONSULTATION (190) Peveler R, George C, Kinmonth AL, Campbell M et al. Effect of antidepressant drug counselling and information leaflets on adherence to drug treatment in primary care: randomised controlled trial. Br Med J. 1999; 319 (7210):612-615. (191) Piette JD, Weinberger M, McPhee SJ, Mah CA et al. Do automated calls with nurse follow-up improve self-care and glycemic control among vulnerable patients with diabetes? Am J Med. 2000; 108 (1):20-27. (192) Portsmouth SD, Osorio J, McCormick K, Gazzard BG et al. Better maintained adherence on switching from twice-daily to once-daily therapy for HIV: a 24week randomized trial of treatment simplification using stavudine prolongedrelease capsules. HIV Medicine. 2005; 6 (3):185-190. (193) Pound P, Britten N, Morgan M, Yardley L et al. Resisting medicines: a synthesis of qualitative studies of medicine taking. Soc Sci Med. 2005; 61 (1):133-155. (194) Pradier C, Bentz L, Spire B, Tourette TC et al. Efficacy of an educational and counseling intervention on adherence to highly active antiretroviral therapy: French prospective controlled study. HIV Clin Trials. 2003; 4 (2):121-131. (195) Pyne JM, McSweeney J, Kane HS, Harvey S et al. Agreement between patients with schizophrenia and providers on factors of antipsychotic medication adherence. Psychiatr Serv. 2006; 57 (8):1170-1178. (196) Rand CS, Wise RA. Measuring adherence to asthma medication regimens. Am J Respir Crit Care Med. 1994; 149 (2 Pt 2):S69-S76. (197) Rao JK. Communication interventions make a difference in conversations between physicians and patients: A systematic review of the evidence. Med Care. 2007; 45 (4):340-349. (198) Rathbun RC, Farmer KC, Stephens JR, Lockhart SM. Impact of an adherence clinic on behavioral outcomes and virologic response in treatment of HIV infection: a prospective, randomized, controlled pilot study. Clinical Therapeutics. 2005; 27 (2):199-209. (199) Raynor DK, Blenkinsopp A, Knapp P, Grime J et al. A system review of quantitative and qualitative research on the role and effectiveness of written information available to patients about individual medicines. Health Technol Assess. 2007; 11 (5):iii-98. (200) Razali SM, Hasanah C, I, Khan UA, Subramaniam M. Psychosocial interventions for schizophrenia. J Ment Health. 2000; 9 (3):283-289. (201) Reginster JY, Rabenda V, Neuprez A, Reginster JY et al. Adherence, patient preference and dosing frequency: understanding the relationship. [Review] [30 refs]. Bone. 2006; 38 (4 Suppl 1):S2-S6. Medicines concordance: full guideline DRAFT (July 2008) page 367 of 373 DRAFT FOR CONSULTATION (202) Reid M, Clark A, Murdoch DL, Morrison C et al. Patients strategies for managing medication for chronic heart failure. Int J Cardiol. 2006; 109 (1):6673. (203) Remien RH, Stirratt MJ, Dolezal C, Dognin JS et al. Couple-focused support to improve HIV medication adherence: a randomized controlled trial. AIDS. 2005; 19 (8):807-814. (204) Richards C, Coulter A. Is the NHS becoming more patient centred: trends from the national surveys of NHS patients in England 2002-07. 2007. Picker Institute. (205) Richter J, Eisemann M, Zgonnikova E. Doctors' authoritarianism in end-of-life treatment decisions. A comparison between Russia, Sweden and Germany. Journal of Medical Ethics. 2001; 27 (3):186-191. (206) Rickles NM, Svarstad BL, Statz-Paynter JL, Taylor LV et al. Pharmacist telemonitoring of antidepressant use: effects on pharmacist-patient collaboration. J Am Pharm Assoc. 2005; 45 (3):344-353. (207) Ross SE, Moore LA, Earnest MA, Wittevrongel L et al. Providing a webbased online medical record with electronic communication capabilities to patients with congestive heart failure: randomized trial.[see comment]. Journal of Medical Internet Research. 2004; 6 (2):e12. (208) Rudd P, Miller NH, Kaufman J, Kraemer HC et al. Nurse management for hypertension. A systems approach. Am J Hypertens. 2004; 17 (10):921-927. (209) Ruskin PE, Silver-Aylaian M, Kling MA, Reed SA et al. Treatment outcomes in depression: comparison of remote treatment through telepsychiatry to inperson treatment. Am J Psychiatry. 2004; 161 (8):1471-1476. (210) Ryan M, Farrar S. Using conjoint analysis to elicit preferences for health care. BMJ. 2000; 320 (7248):1530-1533. (211) Sackett DL, Gibson ES, Taylor DW. Randomised clinical trial of strategies for improving medication compliance in primary hypertension. Lancet. 1975; 1 (7918):1205-1207. (212) Sadik A, Yousif M, McElnay JC. Pharmaceutical care of patients with heart failure. Br J Clin Pharmacol. 2005; 60 (2):183-193. (213) Samuelson PA, William D Nordhaus. Economics. 18th edition. ed. McGrawHill; 2004. (214) Savage R, Armstrong D. Effect of a general practitioner's consulting style on patients' satisfaction: a controlled study. Br Med J. 1990; 301 (6758):968-970. (215) Schaffer SD, Tian L. Promoting adherence: effects of theory-based asthma education. Clin Nurs Res. 2004; 13 (1):69-89. Medicines concordance: full guideline DRAFT (July 2008) page 368 of 373 DRAFT FOR CONSULTATION (216) Schneider A, Wensing M, Quinzler R, Bieber C et al. Higher preference for participation in treatment decisions is associated with lower medication adherence in asthma patients. Patient Educ Couns. 2007; 67 (1-2):57-62. (217) Schneider PJ, Murphy JE, Pedersen CA. Impact of medication packaging on adherence and treatment outcomes in older ambulatory patients. J Am Pharm Assoc. 2008; 48 (1):58-63. (218) Scotto CJ. The lived experience of adherence for patients with heart failure. J Cardpulm Rehabil. 2005; 25 (3):158-163. (219) Segador J, Gil-Guillen VF, Orozco D, Quirce F et al. The effect of written information on adherence to antibiotic treatment in acute sore throat. International Journal of Antimicrobial Agents. 2005; 26 (1):56-61. (220) Shi L, Hudges M, Yurgin N, Boye KS. Impact of dose frequency on compliance and health outcomes: A literature review (1966-2006). Expert Review of Pharmacoeconomics and Outcomes Research. 2007; 7 (2):187-202. (221) Shields CG, Epstein RM, Fiscella K, Franks P et al. Influence of accompanied encounters on patient-centeredness with older patients. The Journal of the American Board of Family Practice / American Board of Family Practice. 2005; 18 (5):344-354. (222) Sidat M, Fairley C, Grierson J. Experiences and perceptions of patients with 100% adherence to highly active antiretroviral therapy: a qualitative study. Aids Patient Care STDS. 2007; 21 (7):509-520. (223) Sookaneknun P, Richards RM, Sanguansermsri J, Teerasut C. Pharmacist involvement in primary care improves hypertensive patient clinical outcomes. Annals of Pharmacotherapy. 2004; 38 (12):2023-2028. (224) Stason WB. Compliance, quality of life, and cost effectiveness. Curr Hypertens Rep. 1999; 1 (6):471-474. (225) Stevenson FA. General practitioners' views on shared decision-making: a qualitative analysis. Patient Educ Couns. 50, 291-293. 2003. (226) Stewart A, Noakes T, Eales C, Shepard K et al. Adherence to cardiovascular risk factor modification in patients with hypertension. Cardiovascular Journal of Southern Africa. 2005; 16 (2):102-107. (227) Stone VE, Jordan J, Tolson J, Miller R et al. Perspectives on adherence and simplicity for HIV-infected patients on antiretroviral therapy: self-report of the relative importance of multiple attributes of highly active antiretroviral therapy (HAART) regimens in predicting adherence. J Acquir Immune Defic Syndr. 2004; 36 (3):808-816. (228) Strang JS, Falloon I-RH, Moss HB. The effects of family therapy on treatment compliance in schizophrenia. Psychopharmacol Bull. 1981; 17 :87-88. Medicines concordance: full guideline DRAFT (July 2008) page 369 of 373 DRAFT FOR CONSULTATION (229) Strydom A, Forster M, Wilkie BM, Edwards C et al. Patient information leaflets for people with learning disabilities who take psychiatric medication. British Journal of Learning Disabilities. 2001; 29 (2):72-76. (230) Sturgess IK, McElnay JC, Hughes CM, Crealey G. Community pharmacy based provision of pharmaceutical care to older patients. Pharm World Sci. 2003; 25 (5):218-226. (231) Taylor CT, Byrd DC, Krueger K. Improving primary care in rural Alabama with a pharmacy initiative. Am J Health Syst Pharm. 2003; 60 (11):11231129. (232) Theunissen-Nicolet CM. Manipulation of patient-provider interaction: Discussing illness representations or action plans concerning adherence. Patient Educ Couns. 2003; 51 (3):247-258. (233) Thompson AG. The meaning of patient involvement and participation in health care consultations: a taxonomy. Social Science & Medicine. 2007; 64 (6):1297-1310. (234) Thomson RG, Eccles MP, Steen IN, Greenaway J et al. A patient decision aid to support shared decision-making on anti-thrombotic treatment of patients with atrial fibrillation: randomised controlled trial. Quality & Safety in Health Care. 2007; 16 (3):216-223. (235) Tortolero-Luna-Guillermo B-TG-J. Relationship between English Language Use and Preferences for Involvement in Medical Care among Hispanic Women. Journal of Women's Health. 2006; 15 (6):774-785. (236) Towle A. Physician and patient communication skills: competencies for informaed shared decision-making. 1997. Vancouver, Canada, University of British Columbia. (237) Trevena LJ, Davey HM, Barratt A, Butow P et al. A systematic review on communicating with patients about evidence. J Eval Clin Pract. 2006; 12 (1):13-23. (238) Tsuyuki RT, Fradette M, Johnson JA, Bungard TJ et al. A multicenter disease management program for hospitalized patients with heart failure.[see comment]. Journal of Cardiac Failure. 2004; 10 (6):473-480. (239) Tuckett D, Boutlon M, Olson C, Williams A. Meetings between experts: an appraoch to sharing ideas in medical consultation. 1985. Tavistock Publishers. (240) Tuldrà A, Fumaz CR, Ferrer MJ, Bayés R et al. Prospective randomized twoArm controlled study to determine the efficacy of a specific intervention to improve long-term adherence to highly active antiretroviral therapy. J Acquir Immune Defic Syndr. 2000; 25 (3):221-228. (241) Turner BJ. Adherence to antiretroviral therapy by human immunodeficiency virus- infected patients. J Infect Dis. 2002; 185 (Suppl 2):S143-S151. Medicines concordance: full guideline DRAFT (July 2008) page 370 of 373 DRAFT FOR CONSULTATION (242) Urien AM, Guillen VF, Beltran DO, Pinzotas CL et al. Telephonic back-up improves antibiotic compliance in acute tonsillitis/pharyngitis. International Journal of Antimicrobial Agents. 2004; 23 (2):138-143. (243) Urquhart J. Patient non-compliance with drug regimens: measurement, clinical correlates, economic impact. European heart journal. 1996; 17 Suppl A :8-15. (244) van S, Nyamathi A, Carpio F, Pearce D et al. Effects of a treatment adherence enhancement program on health literacy, patient-provider relationships, and adherence to HAART among low-income HIV-positive Spanish-speaking Latinos. Aids Patient Care STDS. 2005; 19 (11):745-759. (245) van-Dam HA. Provider-patient interaction in diabetes care: Effects on patient self-care and outcomes A systematic review. Patient Educ Couns. 2003; 51 (1):17-28. (246) Vick S, Scott A. Agency in health care. Examining patients' preferences for attributes of the doctor-patient relationship. J Health Econ. 1998; 17 (5):587605. (247) Vinter RN, Petricek G, Katic M. Obstacles which patients with type 2 diabetes meet while adhering to the therapeutic regimen in everyday life: Qualitative study. Croat Med J. 2004; 45 (5):630-636. (248) Vivian EM. Improving blood pressure control in a pharmacist-managed hypertension clinic. Pharmacotherapy. 2002; 22 (12):1533-1540. (249) Vrijens B, Belmans A, Matthys K, de K et al. Effect of intervention through a pharmaceutical care program on patient adherence with prescribed once-daily atorvastatin. Pharmacoepidemiology & Drug Safety. 2006; 15 (2):115-121. (250) Wagner GJ, Kanouse DE, Golinelli D, Miller LG et al. Cognitive-behavioral intervention to enhance adherence to antiretroviral therapy: a randomized controlled trial (CCTG 578). AIDS. 1909; 2006 Jun 12. (9). (251) Weber R, Christen L, Christen S, Tschopp S et al. Effect of individual cognitive behaviour intervention on adherence to antiretroviral therapy: prospective randomized trial. Antiviral Therapy. 2004; 9 (1):85-95. (252) Wetzels R, Harmsen M, van WC, Grol R et al. Interventions for improving older patients' involvement in primary care episodes. Cochrane Database of Systematic Reviews. 2007;(4 (CD004273) Source notes: Update Software online or CD-ROM updated quarterly ISSN: 1469-493X.). (253) Wetzels R, Wensing M, van WC, Grol R. A consultation leaflet to improve an older patient's involvement in general practice care: A randomized trial. Health Expect. 2005; 8 (4):286-294. (254) Weymiller AJ, Montori VM, Jones LA, Gafni A et al. Helping patients with type 2 diabetes mellitus make treatment decisions: statin choice randomized trial. Arch Intern Med. 2007; 167 (10):1076-1082. Medicines concordance: full guideline DRAFT (July 2008) page 371 of 373 DRAFT FOR CONSULTATION (255) Wilkinson CR, Williams M. Strengthening patient-provider relationships. Lippincott's case management : managing the process of patient care. 2002; 7 (3):86-99. (256) Wills CE, Holmes RM. Patient comprehension of information for shared treatment decision making: State of the art and future directions. Patient Educ Couns. 2003; 50 (3):285-290. (257) Wilson HS, Hutchinson SA, Holzemer WL. Reconciling incompatibilities: a grounded theory of HIV medication adherence and symptom management. Qual Health Res. 2002; 12 :1309-1322. (258) World Health Organization. Adherence to long-term therapies; evidence for action. WHO. 2003 (259) Wyatt GE, Longshore D, Chin D, Carmona JV et al. The efficacy of an integrated risk reduction intervention for HIV-positive women with child sexual abuse histories. AIDS & Behavior. 2004; 8 (4):453-462. (260) Xiong W, Phillips MR, Hu X, Wang R et al. Family-based intervention for schizophrenic patients in China. A randomised controlled trial. Br J Psychiatry. 1994; 165 (2):239-247. (261) Zermansky AG, Petty DR, Raynor DK, Lowe CJ et al. Clinical medication review by a pharmacist of patients on repeat prescriptions in general practice: A randomised controlled trial. Health Technol Assess. 2002; 6 (20). (262) Zhang M, Wang M, Li J, Phillips MR. Randomised-control trial of family intervention for 78 first-episode male schizophrenic patients. An 18-month study in Suzhou, Jiangsu. Br J Psychiatry Suppl. 1994; 165 (Suppl 24):96-102. Medicines concordance: full guideline DRAFT (July 2008) page 372 of 373 DRAFT FOR CONSULTATION Appendices Please note the appendices are available as separate files. Medicines concordance: full guideline DRAFT (July 2008) page 373 of 373