Download A Pragmatic Approach to Implementing Best Practices for Clinical

Journal of the American Medical Informatics Association
Volume 14
Number 1
25
Jan / Feb 2007
JAMIA
Focus on Computerized Provider Order Entry (CPOE)
Viewpoint Paper 䡲
A Pragmatic Approach to Implementing Best Practices for
Clinical Decision Support Systems in Computerized Provider
Order Entry Systems
PETER A. GROSS, MD, DAVID W. BATES, MD, MSC
Abstract
Incorporation of clinical decision support (CDS) capabilities is required to realize the greatest
benefits from computerized provider order entry (CPOE) systems. Discussions at a conference on CDS in CPOE
held in San Francisco, California, June 21–22, 2005 produced several papers in this issue of JAMIA. The first paper
reviews CDS for electronic prescribing within CPOE systems;1 the second describes current controversies
regarding creation, maintenance, and uses of CPOE order sets for CDS;2 and the third presents issues related to
certification as a potential means of validating CPOE systems for widespread use.3 This manuscript summarizes
all of the discussions at the meeting and provides a pragmatically oriented view of how to implement CPOE with
CDS.
䡲 J Am Med Inform Assoc. 2007;14:25–28. DOI 10.1197/jamia.M2173.
Introduction
Implementation of clinical decision support (CDS) within
computerized provider order entry (CPOE) systems requires
ongoing difficult and expensive endeavors. Project leaders
must coordinate the efforts of clinical, technical, and human
factors engineering experts over long periods of time. Institutional justification for installing CPOE systems and adding
CDS is often lacking, with the result that only about 10% of
U.S. hospitals had fully adopted CPOE by 2002.4 One
justifiable reason for moderate to large medical centers to
adopt CPOE with CDS is to reduce substantially the high
number of medication errors that regularly occur.5 While
CPOE systems with relatively modest levels of clinical
decision support can reduce serious medication error rates
Affiliations of the authors: Department of Medicine, Hackensack
University Medical Center (PAG), Hackensack, NJ; UMDNJ-New
Jersey Medical Center (PAG), Newark, NJ; Department of Medicine,
Brigham and Women’s Hospital and Harvard Medical School
(DWB), Boston, MA.
Supported by unconditional educational grants from the Frost
Outcomes Fund at the Hackensack University Medical Center
Foundation and the Department of Informatics, Partners Healthcare, The BD Foundation, and Siemens Medical Solutions. Other
support was received from the Joint Commission on Accreditation
of Healthcare Organizations, The Medical Society of New Jersey,
and the University of Medicine and Dentistry of New Jersey.
The authors thank all the participants in this conference, which was
held in San Francisco, CA, June, 2005.
Correspondence and reprints: Peter A. Gross, MD, Department of
Internal Medicine, Hackensack University Medical Center, Hackensack, NJ 07601; Tel: (201) 996-3500; Fax: (201) 996-3505; e-mail:
⬍[email protected]⬎.
Received for review: 06/08/06; accepted for publication: 10/16/06.
substantially,1,6 some errors will still occur. More advanced
CDS capabilities in CPOE may include computer-based
guidance for usual and maximum medication doses, and
dose adjustments for compromised renal function or advancing age. Advanced CDS within CPOE prescribing can
be highly beneficial.1,7,8
Fortunately, implementing many of the best safety-related
CDS interventions in CPOE may be easy to accomplish. For
example, computerization of prescribing with CDS enables
pediatric dosing based on age and weight.1 Limiting drug
and dosage choices can also reduce unnecessary variation.
Likewise, selecting evidence-based, disease-specific order
sets for patient care can make the “right orders for the
patient” the default.2 However, designing a complex yet
flexible, protocol-based, safe dosing approach to chemotherapy remains a relatively unsolved problem in most institutions implementing CDS within CPOE. This should serve as
a reminder that CPOE with CDS has not yet fully addressed
all medication dosing problems. CPOE has many easy-toimplement benefits outside the medication and safety
realms. For example, CPOE systems can help clinicians to
make more cost– effective use of the clinical laboratory,3 to
deliver clinical knowledge for educational purposes at the
point of care,3 and to implement clinical guidelines and
order sets.2
Technical Issues Related to CPOE and CDS
Discussed at the Meeting
Presenters at the meeting (see Appendix, available online at
www.jamia.org) included an international array of CPOE
and CDS experts. The conference addressed many issues
related to medication ordering and safety. A key priority
when ordering new medications is reconciling the history of
26
GROSS, BATES, Best Practices for Clinical Decision Support Systems in CPOE Systems
drug allergies electronically.1 While allergies from prior
admissions should be available for new admissions, the
current lack of nationally accepted standards for recording
medication allergies records makes long-term maintenance
and cross-institutional interchange of such information challenging. Discussants at the meeting noted that, in addition to
recording allergy triggering agents—which should include
not only medications but also non-drug allergies such as
latex, intravascular contrast, and foods—the nature of the
allergic reactions should also be recorded. A reliable and
auditable means of removing incorrect allergies from a
patient’s electronic record must exist.1 Common cross-sensitivities with related drugs should be considered, although
carte blanche acceptance of all possible cross-sensitivities
can contribute to excessive false-positive allergy alerts. Contraindications for drug use should require a strong evidence
base before being incorporated into a CDS system.
Meeting participants also discussed CPOE-based drug– drug
interaction checking. While important, if not implemented
correctly, this CPOE feature often leads to alerts for so many
low-clinical-relevance interactions that clinician– users ignore the most critical interaction alerts due to “information
overload” or “inability to recognize the needle in the haystack.”1 Not only should CDS–CPOE checking be done for
drug– drug interactions, but also for duplicate drugs, drug–
disease, drug–lab, and drug–pregnancy issues.1 Drug interactions should also be classed by severity. Alerts should
present users with the ability to order available alternatives.1
The very few truly life-threatening drug interactions (i.e.,
those that should never be given together) should generate
a “hard stop” response in the CPOE system so that an order
for the combination will not be implemented; potentially
serious reactions should have an interruptive action and
require a reason for override; and, interactions with less
serious potential consequences might be presented through
a non-interruptive mode.1
The process of drug ordering can also be simplified and
made safer by setting up predetermined lists with patientappropriate doses within order sets.1,2 Another advantage of
these predetermined order sets is that they can be updated
and distributed more easily on the computer than by paper.
Schedules for periodic review of predetermined lists should
be established and adhered to by appropriately qualified
clinical expert maintainers.
Meeting participants noted that tying a reliable, up-to-date
electronic patient record system to the CPOE system can
facilitate pre- and post-admission testing, and facilitate
accuracy of medication reconciliation now mandated by the
Joint Commission on Accreditation of Healthcare Organizations (JCAHO). Within CPOE systems, CDS elements should
promote related laboratory test monitoring for common
drugs such as statins, oral anticoagulants, ACE inhibitors,
and others.
An important topic discussed at the meeting was the optimal method for providing patients with computer-stored
educational information about their prescriptions. While
many individuals now seek information on the Internet,
they may often not find credible advice, or even worse, find
incorrect information. Physicians today may communicate
with their patients by e-mail to answer questions or provide
advice on the need for booster vaccines or other medical
care. Eventually, the patient’s electronic medical record will
be available online. As the public becomes more health care
savvy, the Institute of Medicine’s (IOM) goal of providing
patient-centered care may become more of a reality.
Making CPOE and CDS safer and more user-friendly will
demand that human factor engineering principles be integrated into systems. Human perception, memory, reliability,
and other aspects of cognitive psychology should be carefully considered. These are discussed in detail below.
When fully implemented as part of a complete electronic
medical record (EMR) system, well-designed CPOE systems
may produce an extraordinary effect on care provider satisfaction.9 Valuable time is lost at present when clinicians
must “chase down” paper charts, physically visit the radiology department to view films, and engage in “telephone
tag” to obtain simple advice from colleagues. To paraphrase
what Confucius said three millennia ago, the essence of
knowledge is having it available to apply—in real time.
While most of the discussion during the meeting focused on
the inpatient setting, participants recognized that many
CPOE–CDS applications must also function in the outpatient setting. While the decision support needed differs
somewhat, most of the key principles remain the same.
Issues of how to plan and manage care optimally across the
inpatient– outpatient continuum represent another unsolved
challenge to current CDS applications within CPOE systems.
Finally, the meeting discussed the importance of standards
in facilitating CPOE and CDS implementations. A standardized medical terminology for ordering and results has to be
developed and maintained. Relevant standards activity exists at the national level, with references at government Web
sites such as the National Library of Medicine, at standards
organizations such as HL-7, and at professional organizations’ Web sites, such as AMIA and HIMSS.
Discussion of Non-technical Issues Related to CDS
and CPOE
Mastery of available clinical information is no longer
possible, if it ever was in the first place. With over 20,000
biomedical journals and more than 300,000 randomized
controlled clinical trials, availability of electronic search
resources is mandatory. Online medical, surgical, pediatric, and nursing textbooks are just the beginning.
Guidelines and options for order sets are now available
electronically. “Point-of-care” information can potentially
make care safer and improve its quality. However, how
does an institution or the nation get from here to there?
Technology is only part of the solution—the socio-technical issues are at least as important.1,10 Care providers at
the level of individuals and institutions are slow to
change. Clinical leaders in organizations must pro-actively take responsibility for making iterative improvements. Those implementing systems must monitor them
for adverse effects on both patient care and on caregivers’
workflows. Similar monitoring must occur to ensure that
desired goals are achieved. Continuous refinement based
on ongoing end-user feedback must guide progress. Some
key metrics of technological systems that influence clinicians’ ease of use, and happiness, include: easy access to
Journal of the American Medical Informatics Association
Volume 14
workstations; point-of-care entry and viewing of documentation, including availability of wireless mobile carts
for “rounding” activities; speed and timeliness of the
CPOE system screen flips (no delays longer than a few
seconds); and, secure remote access to clinical systems
from office or home. Taking care of these technical issues
and clinical work flow challenges will eliminate most
barriers to using the system.
Poor planning and implementation by administration or
by the clinical leadership or the information technology
group can result in poor performance or even failures.11
Collaboration requires use of the golden rule and adherence to the “three legged stool” model—where operations, clinicians, and information technology personnel
work together to build trust. All need to focus on continuous modification and feedback, and must operate as a
learning organization.
In determining the overall success of the project, the
processes of planning for and executing a CPOE and CDS
implementation can be as important as the selection of the
systems to install. Implementation of CPOE and CDS
systems should generally be done step-wise— unit-byunit within a hospital— because the house-wide “big
bang” approach carries serious risks. During the implementation, having adequate numbers of computer workstations placed at easy-to-access locations is critical—
planners must determine beforehand the maximum
number of clinician users likely to be active at one time on
each ward or in each clinic, and accommodate this scenario. Similarly, one must plan to replace outdated workstations and networking gear on a regular, rotating basis,
with a planned equipment life cycle of 3– 4 years at most.
Overall, experience from many sites suggests a preferred
order of implementation for CPOE as part of an EMR
system that works well.12 Institutions should begin with
“basics” such as automating laboratory and radiology
results reporting, followed by a pharmacy system that
helps to manage orders with built-in CDS. Addition to the
general EMR of online transcriptions of history and
physical and operative notes and electronic signatures
comes next. Nursing documentation and the electronic
medical administration record (eMAR) may optionally
follow next, or occur after CPOE is implemented. Only
when the building blocks are in place, and clinician– users
are comfortable with technology, should the arduous task
of CPOE and CDS implementation occur gradually.
When implementing systems, it is important to keep in mind
the Ten Commandments for Effective Clinical Decision
Support published by Bates et al.:13
1. Speed is everything—this is what information system
users value most.
2. Anticipate needs and deliver in real time— deliver information when needed.
3. Fit into the user’s work flow—integrate suggestions
with clinical practice.
4. Little things can make a big difference—improve usability to “do the right thing.”
5. Recognize that physicians will strongly resist stopping— offer alternatives rather than insist on stopping
an action.
Number 1
Jan / Feb 2007
27
6. Changing direction is easier than stopping— changing
defaults for dose, route, or frequency of a medication
can change behavior.
7. Simple interventions work best—simplify guidelines by
reducing to a single computer screen.
8. Ask for additional information only when you really
need it—the more data elements requested, the less
likely a guideline will be implemented.
9. Monitor impact, get feedback, and respond—if certain
reminders are not followed, readjust or eliminate the
reminder.
10. Manage and maintain your knowledge-based systems—
both use of information and currency of information
should be carefully monitored.
The financial commitment required to implement CPOE and
CDS is considerable. Cost savings for the total system will
mean increased costs for some subsystems, especially given
the current reimbursement scheme. While it is the “right”
thing to do from the societal perspective, it also has to make
sense for the hospital, and the Chief Financial Officer has to
be convinced. Inadequate documentation of the patient’s
diagnoses represents one of the main tripping points in
maximizing compliance with external quality and reimbursement agencies. If electronic systems can capture correct
diagnoses during CPOE or CDS to improve compliance with
billing directives, it is possible to improve case-mix determinations and reimbursement. Ideally, such systems can
help delivered care, documented care, and coded care to
become the same.
Conclusion
Clinical decision support provides value when a CPOE
system is in place. Implementing CPOE alone is not enough.
The conference pointed out that barriers may seem formidable,
but that an approach can succeed that elicits leadership
support early, identifies and involves strong champions, and
involves a committed administration that provides adequate
financial support. Attention to detail is critical to provide
appropriate support for clinicians’ work flows, adequately
fast system response times, and near-perfect system reliability. National CDS resources are not uniformly available at
present, so institutions must rely on themselves and on
variable offerings from commercial vendors. During such
complex implementations, institutions must remain cognizant of, and carefully monitor, errors that these systems can
cause as well as the errors that they prevent.14-16
Finally, implementing institutions must reach an internal
consensus regarding which decision support features,
“rules,” and alerts to use, and must determine the most
appropriate place to insert them into CPOE systems. Human
factors engineering experience can help to decide the best
ways to implement CDS. Decision support can become
outdated if it is not kept current using credible, evolving
biomedical information. The clinician– users of such systems
must actively provide feedback that system maintainers
rapidly and regularly address in order to attain an optimal
approach to using CDS in CPOE.
References y
1. Kuperman G, Bobb A, Payne T, et al. Medication-related clinical
decision support in computerized provider order entry systems:
A review. J Am Med Inform Assoc. 2007;14:29 – 40.
28
GROSS, BATES, Best Practices for Clinical Decision Support Systems in CPOE Systems
2. Bobb A, Payne T, Gross P. Viewpoint: Controversies surrounding use of order sets for clinical decision support in computerized provider order entry. J Am Med Inform Assoc. 2007;14:
41–7.
3. Classen D, Avery A, Bates D. Evaluation and certification of
computerized provider order entry systems. J Am Med Inform
Assoc. 2007;14:48 –55.
4. Ash JS, Gorman PN, Seshadri V, Hersh WR. Computerized
physician order entry in U.S. hospitals: results of a 2002 survey.
J Am Med Inform Assoc. 2004;11:95–9.
5. Bates DW, Boyle DL, Vander Vliet MB, Schneider J, Leape L.
Relationship between medication errors and adverse drug
events. J Gen Int Med. 1995;10:199 –205.
6. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized
physician order entry and a team intervention on prevention of
serious medication errors. JAMA. 1998;280:1311– 6.
7. Chertow GM, Lee J, Kuperman GJ, et al. Guided medication
dosing for inpatients with renal insufficiency. JAMA. 2001;286:
2839 – 44.
8. Peterson JS, Kuperman GJ, Shek C, Patel M, Avorn J, Bates DW.
Guided prescription of psychotropic medications for geriatric
inpatients. Arch Intern Med. 2005;165:802–7.
9. Mekhjian HS, Kumar RR, Kuehn L, et al. Immediate benefits
realized following implementation of physician order entry at
an academic medical center. J Am Med Inform Assoc. 2002;9:
529 –39.
10. Ash JS, Stavri PZ, Kuperman GJ. A consensus statement on
considerations for a successful CPOE implementation. J Am
Med Inform Assoc. 2003;10:229 –34.
11. Campbell EM, Sittig DF, Ash JS, Guappone KP, Dykstra RH.
Types of unintended consequences related to computerized
provider order entry. J Am Med Inform Assoc. 2006;13:
547–56.
12. Garets D, Davis M. EMRs versus EHRs: Concepts as different as
apples and oranges at least deserve separate names. Health
Inform. 2005(Oct);22:53– 4.
13. Bates DW, Kuperman GJ, Wang S, et al. Ten commandments for
effective clinical decision support: making the practice of evidence-based medicine a reality. J Am Med Inform Assoc.
2003;10:523–30.
14. Koppel R, Metlay JP, Cohen A, et al. Role of computerized
physician order entry systems in facilitating medication errors.
JAMA. 2005;293:1197–203.
15. Chaudhry B, Wang J, Wu S, et al. Systematic review: impact of
health information technology on quality, efficiency, and costs
of medical care. Ann Intern Med. 2006;144:742–52.
16. Colpaert K, Claus B, Somers A, Vandewoude K, Robays H,
Decruyenaere J. Impact of computerized physician order entry
on medication prescription errors in the intensive care unit: a
controlled cross-sectional trial. Crit Care. 2006;10:R21 [Epub
ahead of print].