Download BITTERSWEET FINDINGS OF BLOOD GLUCOSE LEVELS IN

BITTERSWEET FINDINGS OF
BLOOD GLUCOSE LEVELS
IN 467,955 PATIENTS
IN PRIMARY CARE
MARCH 2015
Principal Investigator:
Associate Professor Melinda Carrington
Research Team:
Associate Professor Neale Cohen
Professor Simon Stewart
Dr Tracey Gerber
Funding: Sanofi Australia Pty Ltd provided funding for this study.
This company was not involved in the design, collection,
analysis or interpretation of the data, but they were given
the opportunity to review this Report prior to publication.
Any enquiries or comments about this publication should be directed to:
Centre of Research Excellence to Reduce Inequality in Heart Disease
Preventative Health
Baker IDI Heart and Diabetes Institute
PO Box 6492
Melbourne VIC 8008
T 1800 850 622
E [email protected]
W http://www.cre2rihd.org.au/
Suggested reference:
Carrington MJ, Stewart S, Gerber T and Cohen N.
Bittersweet findings of blood glucose levels in 467,955 patients in primary care.
June 2015, Baker IDI Heart and Diabetes Institute, Melbourne, Australia.
CONTENTS
List of Tables 2
List of Figures 2
Abbreviations 3
Executive Summary 4
Key findings 5
Introduction
What is diabetes?
How is diabetes diagnosed?
Diabetes prevalence in Australia
The effects of diabetes
Treatment for diabetes
The economic impact of diabetes in Australia
6
7
8
8
9
9
Objectives10
Methods
Data source11
Consent11
Data procedures and analyses11
Study cohort
General Practitioners12
Patients13
Results
Section 1: Blood sugar levels of first patient encounters
Part 1
General trends14
Trends over time14
Time trends according to gender16
Hypoglycaemic treatment17
Hypoglycaemic treatments prescribed according to class
17
Time trends in class of hypoglycaemic treatments prescribed
18
Trends according to hypoglycaemic treatment
Blood glucose levels over time according to hypoglycaemic treatment
Associated risk factors according to hypoglycaemic treatment
19
19
20
Part 2
Glycaemic control22
HbA1c levels according to hypoglycaemic treatment
22
Hypoglycaemic treatment according to HbA1c control
22
Associated risk factors according to HbA1c control
24
Section 2: Blood sugar levels of repeat patient encounters
Part 1
Patients never prescribed hypoglycaemic treatment
Part 2
Patients initiated on hypoglycaemic treatment
Patients initiated on insulin therapy after management with non-insulin therapeutics
28
29
31
Conclusions32
Limitations33
References34
Appendix35
1
LISTINGS
LIST OF TABLES
Table 1. Reference blood glucose and HbA1c levels for the detection and control of type 2 diabetes
7
Table 2. Number of men and women with HbA1c and/or fasting glucose measurements recorded per year
between 2005 and 2013
13
Table 3. Frequency of epochs within contiguous visit sequence lengths (2-18) that HbA1c levels are above 7.0%
28
Table 4. Percentage of patients who started hypoglycaemic treatment according to time therapy was
initiated and sequence length
29
LIST OF FIGURES
Figure 1. The prevalence of diagnosed diabetes in Australia from 1988 to 2013
8
Figure 2. Total primary care clinics and patients with fasting glucose or HbA1c monitoring per State and Territory
between 2005 and 2013
12
Figure 3. Total number of men and women with HbA1c or fasting glucose measurements recorded between
2005 and 2013 according to age
13
Figure 4. Average annual HbA1c levels between 2005 and 2013
14
Figure 5. Percentage of patients with elevated HbA1c levels between 2005 and 2013
14
Figure 6. Average annual fasting glucose levels between 2005 and 2013
15
Figure 7. Percentage of patients in each diabetes risk category according to fasting glucose levels between 2005 and 2013
15
Figure 8. Average annual HbA1c levels according to gender between 2005 and 2013
16
Figure 9. Average annual fasting glucose levels according to gender between 2005 and 2013
16
Figure 10. Percentage of patients prescribed hypoglycaemic treatments according to class
17
Figure 11. Percentage of patients prescribed hypoglycaemic treatments according to class between 2005 and 2013
18
Figure 12. Average annual HbA1c levels according to hypoglycaemic treatment between 2005 and 2013
19
Figure 13. Average annual fasting glucose levels according to hypoglycaemic treatment between 2005 and 2013
19
Figure 14. Average annual weight levels according to hypoglycaemic treatment between 2005 and 2013
20
Figure 15. Average annual systolic (upper trends) and diastolic (lower trends) blood pressure levels according
to hypoglycaemic treatment between 2005 and 2013
20
Figure 16. Average annual total cholesterol levels according to hypoglycaemic treatment between 2005 and 2013
21
Figure 17. Average annual eGFR levels according to hypoglycaemic treatment between 2005 and 2013
21
Figure 18. Percentage of patients who achieve glycaemic targets according to treatment
22
Figure 19. Percentage of patients who achieve the glycaemic target HbA1c of 7.0% according to treatment
between 2005 and 2013
23
Figure 20. Percentage of patients prescribed hypoglycaemic treatment according to class and achievement
of the glycaemic HbA1c goal of 7.0%
23
Figure 21. Average annual weight levels according to achievement of the glycaemic HbA1c goal of
7.0% between 2005 and 2013
24
Figure 22. Average annual blood pressure levels according to achievement of the glycaemic HbA1c goal of
7.0% between 2005 and 2013
24
Figure 23. Average annual total cholesterol levels according to achievement of the glycaemic HbA1c goal of
7.0% between 2005 and 2013
25
Figure 24. Average annual eGFR levels according to achievement of the glycaemic HbA1c goal of
7.0% between 2005 and 2013
25
Figure 25. Average HbA1c levels of contiguous 6 monthly visit data
26
Figure 26. Flowchart of contiguous 6 monthly visit data according to hypoglycaemic treatment
27
Figure 27. Average HbA1c levels that treatment was initiated according to gender
30
Figure 28. Class of hypoglycaemic treatments patients were initiated on31
2
ABBREVIATIONS
CVD
Cardiovascular Disease
eGFR
Estimated glomerular filtration rate
DPP4
Dipeptidyl peptidase 4
GP
General practitioner
GLP1
Glucagon like peptide 1
GPRN
General Practice Research Network
HbA1c
Glycated haemoglobin
HCN
Health Communication Network trading as MedicalDirector
3
EXECUTIVE
SUMMARY
This report is a retrospective
analysis of patient-based
electronic medical records
in primary care in Australia
between 2005 and 2013.
The main objective was
to examine trends in blood
glucose levels and the
prescription of hypoglycaemic
treatment. It provides current
data on the pattern of
surveillance of blood glucose
monitoring and management
of diabetes over 9 years,
in turn offering important
clinical and public health
messages for health
professionals and the
general public.
The importance of undertaking large
scale analyses to examine blood
glucose levels in the primary care
setting, whereby the majority of
management for diabetes occurs,
cannot be overstated. In the context
of escalations in blood glucose levels
in Australia over the past 20 years,
associated with increases in modifiable
risk factors such as body fatness, clinical
practice needs to be reviewed against
standards of diabetes prevention and
care. This is essential to reduce the
prevalence of diabetes, minimise diabetic
complications and prolong survival in
those already affected by this condition.
We analysed de-identified data for this
report from a representative sample
of primary care clinics using Medical
Director patient management software,
supplied by the Health Communication
Network© (HCN). A blood glucose
measurement (either fasting glucose
or HbA1c test result) was documented
for 467,955 patients (55% women, aged
55.4 ± 19.7 years) by 945 GPs from
320 primary care clinics across Australia
in the study period from 2005 to 2013.
4
KEY
FINDINGS
• Women were more likely to have
a blood sugar result recorded than
men, particularly in the 25 to 34 year
age group, perhaps reflecting
surveillance for gestational diabetes.
• The frequency of blood glucose
testing increased after 45 years
of age for both men and women.
• Average blood sugar levels
(both HbA1c and fasting glucose)
showed no change between 2005 and
2013 and were consistently higher in
men than women.
• The proportion of patients with
elevated HbA1c levels above 7%,
irrespective of treatment, was around
40% in 2013 and decreased by
approximately 10% since 2005.
• Fasting glucose results identified
individuals as low risk (< 5.5 mmol/L),
increased risk (pre-diabetes; 5.6 to 6.9
mmol/L) or at very high risk of diabetes
(≥ 7.0 mmol/L) for 71%, 21% and 8%
of all measurements, respectively.
• In the 5% of patients with evidence
of hypoglycaemic prescription
information:
›› Biguanides and sulphonylureas
were the most commonly prescribed
non-insulin hypoglycaemic treatments
in 74% and 43% of individuals,
respectively, whilst insulin was
prescribed for 20%.
›› Prescriptions for biguanides and
insulin increased by 14% and 5%,
respectively, and sulphonylureas
decreased by 20% from 2005 to 2013.
›› The advent of DPP4 inhibitors saw
15% of patients on hypoglycaemic
agents prescribed this treatment
in 2013.
• Average blood sugar levels (both
HbA1c and fasting glucose) were
consistently higher in patients
prescribed insulin than non-insulin
hypoglycaemic treatment or those
not prescribed any treatment.
• Inability to achieve the HbA1c
glycaemic recommendation of 7.0%
was associated with increased weight
but no differences in total cholesterol,
eGFR or blood pressure levels from
2005 to 2013.
• Any form of hypoglycaemic treatment
(non-insulin or insulin) was associated
with increased weight and reduced
total cholesterol and eGFR levels
with marginally higher blood
pressure levels. The weight of those
prescribed treatment and the eGFR
levels of all individuals increased
from 2005 to 2013.
• There were no changes over time
in HbA1c levels [range 7.0% to 7.5%]
with more frequent medical evaluation,
albeit average values were higher in
patients with more regular evaluation.
• The HbA1c glycaemic goal
of 7.0% was:
›› Achieved for 43% of patients without
hypoglycaemic treatment and for
15% prescribed pharmacological
therapy.
›› Not achieved for 22% of patients
pharmacologically managed and
20% who were not prescribed
hypoglycaemic medication yet
who could benefit from treatment.
• From the pool of HbA1c measurements,
more patients who achieved the
HbA1c glycaemic target of 7.0%
over the study period (regardless
of treatment) was offset by less
patients above this target and not
pharmacologically managed, rather
than not being treated to target.
• Between 10% and 36% of patients who
were never prescribed hypoglycaemic
treatment were above the recommended
HbA1c level of 7.0% in at least half of
all visits.
• Average time to hypoglycaemic
treatment initiation was 18 ± 10
months [range 12 to 96 months]
at a HbA1c level of 7.4 ± 1.4 %
overall, but higher in men than women
(7.4 ± 1.4 % vs 7.3 ± 1.3 %). Initiation
of treatment was not influenced by
more ongoing contact with a GP.
• Biguanides and sulphonylureas were
the most commonly prescribed
non-insulin hypoglycaemic treatments
in 73% and 27.0% of newly treated
individuals, respectively, whilst insulin
was first recorded for 16% of patients.
• Average time from non-insulin
hypoglycaemic therapy to insulin
therapy was 26 ± 17 months at
a HbA1c level of 8.6% ± 1.6%.
• Insulin was 3 times more commonly
prescribed and biguanides less
commonly administered for patients
above the glycaemic goal of 7.0%
compared to patients who achieved
this target. More prescriptions for
sulphonylureas, thiazolidinediones
and DPP4 inhibitors were administered
in patients not achieving standard
glycaemic control.
5
INTRODUCTION
WHAT IS DIABETES?
Type 1 diabetes
Diabetes is a chronic disorder characterised
by high blood sugar (hyperglycaemia)1.
Sugar or glucose is a nutrient that cells
in our body use to produce energy.
Glucose is found in many foods including
bread, cereals, fruit, starchy vegetables,
milk, yoghurt and sweets. To absorb
glucose from the blood stream into cells,
a hormone named insulin is needed.
Insulin is produced in the pancreas and
when released, attaches to glucose.
This combination then moves into our cells.
In type 1 diabetes, the pancreas stops
insulin production. As insulin is required
for glucose uptake into the body’s cells,
low amounts of insulin (or none at all)
lead to hyperglycaemia. In order to
maintain adequate blood sugar levels,
regular monitoring and insulin injections
are required numerous times per day.
Individuals with diabetes do not produce
insulin in sufficient amounts leading to
excess blood glucose. When this condition
is not treated accordingly by diet and
lifestyle modification or pharmacological
medication (injections or non-insulin
agents), the result can be life threatening.
There are three main types of diabetes;
type 1, type 2 and gestational. The signs
and symptoms include:
• Excessive thirst
• Excessive and unexplained
weight loss (specific to type 1)
• Blurry vision
• Slow healing cuts
• Mood swings, headaches
and dizziness
• Leg cramps
• Feeling tired and lethargic
• Always feeling hungry
• Increased urination
Type 1 diabetes is an autoimmune
disease; cells in the immune system
(that usually protect the body from
disease and illness) begin to attack
the pancreas, destroying cells that
produce insulin, hence decreasing
insulin production and resulting in
hyperglycaemia. There is no definitive
cause of type 1 diabetes yet genetics,
viral or bacterial infection, autoimmune
reactions and toxic chemicals have been
proposed as potential possibilities.
Type 2 diabetes
In type 2 diabetes, the pancreas may be
able to produce insulin but in insufficient
amounts and hyperglycaemia occurs.
Type 2 diabetes is regularly treated
without medication and can be managed
by altering lifestyle via increasing
physical activity and adopting a healthy
diet. However some individuals with type
2 diabetes, particularly those with longer
duration diabetes, require medication.
Type 2 diabetes occurs when the cells
in the body become resistant to insulin,
this is called insulin resistance. Initially,
the pancreas reacts by over-producing
insulin, but over time cannot sustain this,
leading to a loss of insulin producing
cells and insulin deficiency. This means
that the pancreas works at a decreased
rate and is not able to make an adequate
supply of insulin for glucose uptake.
An individual is at higher risk of
developing type 2 diabetes under the
following conditions1:
• Family history of diabetes in a first
degree relative (genetic predisposition)
• Aged above 55 years
• Aged above 45 years and are
overweight or have hypertension
• Aged above 35 years and have
Aboriginal and/or Torres Strait Islander
ethnicity
• Personal history of gestational diabetes
• Delivered a baby weighing over
4.5 kgs (9 lbs)
• Personal history of polycystic
ovarian syndrome
Gestational diabetes
Gestational diabetes is diagnosed during
pregnancy (week 24 to 28) and often
disappears after childbirth. It can recur
in future pregnancies. In gestational
diabetes, the placenta produces hormones
to aid in pregnancy but which can make
cells insulin resistant. Most of the time,
the pancreas accounts for this and
makes more insulin, however in some
cases the pancreas cannot keep up and
glucose remains in the blood stream
resulting in hyperglycaemia1.
Gestational diabetes is more frequent in
woman with the following characteristics:
• Aged above 30 years
• Family history of diabetes in a first
degree relative
• Overweight or obese
• Aboriginal or Torres Strait Islander,
Indian, Vietnamese, Chinese, Middle
Eastern or Polynesian ethnic background
Gestational diabetes is managed by
adopting a healthy eating plan and
increasing physical activity with frequent
monitoring of blood glucose levels. In
some cases medication is necessary.
6
Presently, there are around 1.1 million people living
with diabetes in Australia. This is double the number of
Australians with diabetes since 1990, and the amount
is enduring with about 100,000 new cases of diabetes
diagnosed each year1.
HOW IS DIABETES
DIAGNOSED?
Diabetes is diagnosed (and monitored)
via a simple blood test to measure glucose
levels. This can be done in a number
of ways and the results compared to
reference levels (refer Table 1)2.
• Fasting blood glucose levels:
Measured after an overnight fast,
a simple blood sample is taken to
determine the amount of glucose
in the blood.
• HbA1c concentration: A blood
sample is collected (does not require
overnight fasting) to measure HbA1c a protein that is reflective of the
average blood glucose concentration
over an 8-12 week period.
• Glucose tolerance test:
After consuming a drink loaded with
glucose, blood samples are measured
over a period of time (2 hours) to
determine the rate at which the
glucose is taken into the cells.
In addition to blood glucose monitoring,
symptom assessment is recommended
to identify typical symptoms such as
increased thirst or hunger, frequent
urination or slow healing wounds.
Table 1. Reference blood glucose and HbA1c levels for the detection
and control of type 2 diabetes
Diagnostic criteria for diabetes
Blood glucose (spot sample)
Ideal
Pre-diabetes
Diabetes
Fasting
Below 5.5 mmol/L
5.6 to 6.9 mmol/L
7.0 mmol/L and higher
HbA1c (8-12 week period)
Ideal
Pre-diabetes
Diabetes
Below 5.7% (39 mmol/mol)
5.7 to 6.4% (39 to 46 mmol/mol)
6.5% (48 mmol/mol) and higher
For individuals with diabetes
Pre-prandial
Post-prandial
Blood glucose (spot sample)
3.9 to 7.2 mmol/L
Less than 10 mmol/L
Target HbA1c
More stringent
Below 6.5% (48 mmol/mol),
if achievable without adverse effects
Reasonable
Below 7.0% (53 mmol/mol)
Less stringent
Below 8.0% (64 mmol/mol),
if lower targets cause difficulty to achieve
2
Source: American Diabetes Association
Pre-diabetes
Pre-diabetes is characterised by impaired
fasting glucose or impaired glucose
tolerance. Blood glucose levels are
higher than normal in pre-diabetes but
are not high enough for a diagnosis of
diabetes. To prevent the development
of type 2 diabetes, treatment involves
lifestyle modification, predominantly
to reduce body weight.
7
Every day, 280 Australians develop
diabetes amounting to over 100,000
people diagnosed per year 1.
DIABETES PREVALENCE
IN AUSTRALIA
THE EFFECTS OF DIABETES
Figure 1 shows the increasing prevalence
in diabetes from 1988 to 2013 which has
almost quadrupled within this 20 year
span. More recently however, there is a
plateau in prevalence rates of diabetes
from 2008 to 20133. Approximately 1.1
million Australians already have diabetes
consisting of 120,000 individuals with
type 1, 956,000 with type 2 and 23,600
with gestational diabetes1. This may
be a significant underestimate however
because a large number of individuals
are living with diabetes but are unaware
of it due to the progressively slow onset
and the fact that it is largely asymptomatic;
there is 1 undiagnosed diabetes case for
every three diagnosed cases4. Hence,
the total number of Australians with
diabetes, including those with pre-diabetes,
is approximately 3.2 million. Lifestyle
measures have been shown to prevent
the progression to diabetes in people
with pre-diabetes by up to 58.0%1.
Diabetes can predispose individuals
to a number of dangerous and life
threatening conditions. Diabetes shares
a range of risk factors with cardiovascular
disease (CVD) such as obesity, high
blood pressure and raised lipids
(e.g. cholesterol and triglyceride) due
to poor diet and physical inactivity.
Over half of sufferers of diabetes (58.0%)
are diagnosed with CVD3. Diabetes can
directly damage vital organs; damage
to the kidney (nephropathy), nerves
(neuropathy) and eyes (retinopathy)
can be affected, particularly if diabetes
is left untreated. In 2007/2008, over
95,000 Australians experienced loss
of vision due to diabetes3 and
approximately 5,700 adults with
diabetes received treatment for
end stage kidney disease.
In 2010, approximately 7,750 Australians
died from diabetes or a diabetes-related
cause, accounting for 5.4% of all deaths,
55% of whom were men. By 2033, it is
predicted that more than 3.4 million
Australians will have diabetes at a cost
of $8 billion5, greatly increasing the
burden imposed by diabetes, with a
profound impact on our health care
system.
Per cent
5
4
3
2
1
0
1988
1993
1998
2003
2008
2013
3
Source: Australian Institute of Health and Welfare
Figure 1. The prevalence of diagnosed diabetes in Australia from 1988 to 2013
8
Treatment for diabetes is concerned with
reducing blood glucose levels by changing
dietary and lifestyle behaviours
and/or with medications6.
TREATMENT FOR DIABETES
2. Medication
1. Lifestyle behaviours
There are different types of medication
used to treat diabetes that differ in
respect to their action.
• Increasing physical activity:
At least 30 minutes of moderate
physical activity should be performed
on most, if not all days of the week.
A total of at least 150 minutes per
week is recommended.
• Making healthy nutrition choices:
Diet should follow a normal and
healthy eating plan. The Australian
Dietary Guidelines provide information
on healthy eating for adults7. To ensure
a healthy lifestyle, adults need to
achieve and maintain a healthy weight
by being active and choosing nutritious
food and drink to meet energy needs.
A wide variety of food from the following
5 groups should be consumed:
I. Vegetables
II. Fruit
III. Grains, mostly wholegrain
and high fibre
IV. Lean meats, poultry, fish,
eggs, tofu, nuts, seed and legumes
V. Reduced fat milk, cheese
and yoghurt
Foods with high salt content, added
sugar and high saturated fat content
should be avoided and alcohol intake
restricted.
• Biguanides: Reduce the amount of
glucose that is released from the liver
and slows glucose absorption in the
small intestine, preventing blood
glucose from rising.
• Sulphonylureas: Stimulate the
release of insulin from the pancreas.
This leads to an increased amount of
glucose taken into the cells, lowering
blood glucose.
• Thiazolidineodiones: Reduce the
amount of glucose released from the
liver and improves the action of insulin
on our cells, resulting in lowered
blood glucose.
THE ECONOMIC IMPACT
OF DIABETES IN AUSTRALIA
In 2008/2009, 2.3% (estimated at
$1,507 million) of the total allocated
health expenditure was utilised to treat
and manage diabetes. The majority
of costs were spent in primary care,
24% on out of hospital medical services,
33% on hypoglycaemic medication
and 43% on hospital admitted
patients3. General Practitioners (GP),
endocrinologists, psychologists,
exercise physiologists, diabetes nurses
and diabetes educators8 are the first
step in seeking diabetes treatment and
management, and do not come cheaply.
• Dipeptidyl Peptidase 4 (DPP-4)
inhibitors: Elevate the release of a
hormone within the gut that stimulates
the pancreas to release insulin.
• Alpha Glucosidase Inhibitor:
Decrease the rate of carbohydrate
digestion in the intestine, reducing
the rate at which glucose enters the
blood stream.
• Glucagon Like Peptide 1 (GLP1)
agonists: Stimulates the pancreas
to release insulin and restricts the
release of glucose into the blood
stream by the liver. Additionally, this
medication group can delay the
stomach from emptying, reducing
appetite and preventing blood
glucose from rising after meals.
• Metiglinide: Increases the release of
insulin from the pancreas resulting in
greater uptake of glucose into our cells.
• Insulin: Aids in glucose absorption
from the blood stream into cells. Insulin
attachment to insulin receptors is
required to bring glucose into the cell.
9
OBJECTIVES
OBJECTIVES
Given the escalation in blood glucose
levels in Australia over the past 20 years,
the main objective of this Report was to
examine trends in blood glucose and
HbA1c levels and treatment in patients
attending primary care, whereby the
majority of management for the condition
occurs. It provides current data on the
pattern of surveillance of blood glucose
monitoring and management of diabetes
for the nine years beginning from
January 2005 to December 2013. Many
clinical and public health messages for
health professionals and the broader
community can be gleaned from these
analyses.
10
METHODS
DATA SOURCE
De-identified data (removal of patient’s
name, address, telephone number,
Medicare number or any other
information that could “reasonably”
identify the person) for this report were
supplied by the Health Communication
Network© (HCN). HCN are the providers
of Medical Director, the leading clinical
software product in the Australian health
market that is used by more than 17,000
clinicians and 85% of computerised GPs
(refer http://www.hcn.com.au/Products/
Medical+Director). The data from a
representative sample of GP clinics
using Medical Director are stored in
a longitudinal patient-based database;
currently, this dataset stores more than
32 million patient encounters, over 30
million scripts for approximately 3 million
unique patients from 1,100 GPs.
Participating GP practices are members
of the General Practice Research Network
(GPRN) that have consented to take part
in HCN research and development
activities. Practices of the GPRN are
provided with a data collection software
tool which encrypts and de-identifies all
data prior to it been sent on a weekly
basis to the HCN. The data collected
are available to third party buyers and
are also analysed by HCN to undertake
research to improve patient care.
CONSENT
The information provided by the HCN
is collected on accordance with privacy
guidelines (refer http://www.hcn.com.au/
Privacy). The HCN provides an
information statement policy for display
at the point of care and a leaflet outlining
the possible uses of the information and
also the de-identification process for the
consumer (refer Appendix 1). Patients
and clinicians are able to opt out and
refuse the use of any personal health
information for research and statistical
purposes by selecting a box in the data
collection software. No patient can be
identified from any data collected by
the GPRN and HCN does not sell
GP/specialists’ personal details.
The use of these data was approved by
the Human Research Ethics Committee
at the Alfred Hospital, Melbourne, Australia
(Project number 22/14).
DATA PROCEDURES
AND ANALYSES
Data were cleaned to remove outliers
and extreme scores before linking
variables for analyses (e.g. GP and
patient demographics with clinical and
prescription information). All individuals
with acceptable (after data cleaning)
HbA1c or fasting glucose measurements
within the study period and who could be
identified according to age and sex were
considered for inclusion in this report.
Only patients over 18 years of age were
included in analyses.
HbA1c was the main variable of interest
(supported by fasting blood glucose
measurements). Clinical data at the
same date of visit was included such
as anthropometric measurements
(body weight) and biomedical
measurements (blood pressure, total
cholesterol and estimated glomerular
filtration rate [eGFR]).
Prescription data was reviewed to establish
patients who were on hypoglycaemic
treatment which was then coded
according to class of action. Patients
indicated as being on hypoglycaemic
treatment was documented by evidence
of a hypoglycaemic prescription on
or before the date of a blood glucose
result (for Section 1 analyses) or within
6 monthly epochs in any sequence of
epochs (Section 2 analyses), otherwise
hypoglycaemic therapy was not confirmed.
There are two main sections to the report
that vary based on the data selected for
analyses, with relevant sub-analyses
contained within each section:
SECTION 1: Blood sugar levels
of first patient encounters
This section uses only one measurement
(the first) for any individual patient in the
data set. For patients who had more than
one recorded blood glucose result, only
the initial result was used and their
remaining data were set aside for analyses
to be undertaken as part of Section 2.
The key elements of Section 1 are:
Part 1 – General trends in blood
glucose/HbA1c levels and associated
anthropometric and biomedical risk
factors according to gender and
treatment; diagnostic criteria and
hypoglycaemic treatment prescriptions
Part 2 – Glycaemic control and
treatment and associated risk
factor control
SECTION 2: Blood sugar levels of
repeat patient visits
This section uses all measurements for
an individual patient whereby there are
at least two HbA1c results separated
by 6 months. Results are shown in 6
monthly epochs for the longest available
sequence length, up to a maximum of
9 years (i.e. 18 sets of 6 monthly blocks).
The key elements of Section 2 are:
Part 1 – Patients never prescribed
hypoglycaemic treatment.
Part 2 – Patients who are initiated
on hypoglycaemic treatment or insulin
therapy.
11
STUDY
COHORT
GENERAL PRACTITIONERS
Over the 9 years commencing in 2005,
a blood glucose measurement (either
fasting glucose or HbA1c test result) was
recorded for a total of 467,955 patients
(55% women, aged 55.4 ± 19.7 years)
by 945 GPs from 320 primary care
clinics across Australia. Figure 2 shows
the distribution of results and the number
of clinics who provided the measurements
for each State and Territory in Australia.
Clinics = 3
Patients
GLU= 3,312
HbA1c= 392
Clinics = 14
Patients
GLU= 29,484
HbA1c= 3,855
Clinics = 30
Patients
GLU= 65,927
HbA1c= 6,662
Clinics = 50
Patients
GLU= 95,996
HbA1c= 7,691
Clinics = 129
Patients
GLU= 115,169
HbA1c= 15,270
Clinics = 78
Patients
GLU= 102,060
HbA1c= 13,838
Clinics = 3
Patients
GLU= 5,177
HbA1c= 317
Clinics = 13
Patients
GLU= 31,237
HbA1c= 2,696
Figure 2. Total primary care clinics and patients with fasting glucose or HbA1c
monitoring per State and Territory between 2005 and 2013
12
PATIENTS
Raw data of 1,516,347 blood sugar
readings (HbA1c or fasting glucose) from
487,638 patients between 2005 and 2013
were received. Following the exclusion
of patients with missing age and sex
information and removal of outliers or
extreme HbA1c and/or fasting glucose
measurements, there were a total of
1,296,396 blood sugar readings from
467,955 unique individuals during the
study period.
Table 2 shows the number of men
and women who had either HbA1c or
glucose measured according to year
of visit. Within the entire dataset, there
was a predominance of fasting glucose
(n=448,362) compared to HbA1c
(n=50,721) measurements. A total of
31,128 had both a fasting glucose and
HbA1c result recorded on the same
date of visit; both of these values were
used in Section 1 of this report. Another
417,234 individuals had only a fasting
glucose result and 19,593 had only a
HbA1c measurement the first time they
appeared in the dataset.
Table 2. Number of men and women with HbA1c and/or fasting glucose
measurements recorded per year between 2005 and 2013
Year
HbA1c (n=50,721)
Men
Women
Total
Glucose (n=448,362)
Men
Women
Total
2005
5,798
4,742
10,540
23,149
25,399
48,548
2006
3,352
3,053
6,405
27,722
33,283
61,005
2007
3,340
3,013
6,353
28,960
36,224
65,184
2008
3,271
3,064
6,335
27,804
33,725
61,529
2009
3,086
2,762
5,848
22,778
28,097
50,875
2010
2,544
2,325
4,869
20,173
25,720
45,893
2011
2,173
1,895
4,068
20,654
25,642
46,296
2012
1,963
1,820
3,783
17,835
22,592
40,427
2013
1,286
1,234
2,520
12,499
16,106
28,605
60000
Men
Women
Number of patients
50000
Of the total sample, 55% (n=256,038)
were women. Figure 3 shows the number
of men and women who had any blood
sugar assessment between 2005 and
2013 according to age. As shown, there
was a progressively greater number of
blood sugar results with increasing age,
with more frequent testing (approximately
80,000 tests per age decade) occurring
after 45 years of age. In nearly all age
groups, excluding 55 to 74 year olds,
there were more women than men who
had a blood sugar result recorded.
40000
30000
20000
10000
0
18-24
25-34
35-44
45-54
55-64
65-74
75 and over
Age (years)
Figure 3. Total number of men and women with HbA1c or fasting glucose
measurements recorded between 2005 and 2013 according to age
13
RESULTS
SECTION 1
PART 1
BLOOD SUGAR LEVELS
OF FIRST PATIENT
ENCOUNTERS
GENERAL TRENDS
TRENDS OVER TIME
In this first set of analyses, only the
earliest HbA1c and fasting glucose
result(s) that was recorded for an
individual patient was used. In instances
where patients had repeat blood sugar
measurements taken between 2005
and 2013, the initial recorded result
was included for analyses and any
additional measurements were set
aside for analyses to be undertaken
in Section 2 of this Report.
Based on 50,721 HbA1c measurements, Figure 4 shows that average yearly
HbA1c levels remained unchanged at around 7.0% between 2005 and 2013.
HbA1c
10
9
HbA1c (%)
8
7
6
5
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 4. Average annual HbA1c levels between 2005 and 2013
Figure 5 shows, for three different glycaemic control targets, a general trend for a
reduction in the proportion of individuals with elevated HbA1c levels between 2005
to 2013, becoming evident after 2011.
Over 60% of patients had a HbA1c level greater than 6.5% in 2005 compared to
approximately 50% in 2013. Likewise, just under half of patients had a HbA1c value
above 7.0% in 2005 compared to under 40% in 2013. The percentage of patients with
elevated HbA1c above 8.0% remained unchanged at approximately 20% between
2005 and 2013.
100
HbA1c ≥ 6.5%
HbA1c ≥ 7.0%
HbA1c ≥ 8.0%
90
Elevated HbA1c (%)
80
70
60
50
40
30
20
10
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 5. Percentage of patients with elevated HbA1c levels between 2005 and 2013
14
Fasting glucose
Based on 448,362 fasting blood glucose records, Figure 6 shows no change in yearly
glucose measurements between 2005 and 2013, which remained at an average level
of 5.5 mmol/L.
10
Fasting glucose (mmol/L)
9
8
7
6
5
4
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 6. Average annual fasting glucose levels between 2005 and 2013
According to diagnostic thresholds2, Figure 7 shows a consistent trend over time
in the proportion of patients (71%) with a fasting glucose measurement who were low
risk for diabetes (< 5.5. mmol/L). Exactly 21% of patients were identified at increased
risk in the pre-diabetes range (5.6 to 6.9 mmol/L) and for 8.0%, diabetes may be
considered by a fasting blood glucose of over 7.0 mmol/L.
100
Glu ≤ 5.5 mmol/L
Glu 5.6 - 6.9 mmol/L
90
Glu ≥ 7.0 mmol/L
80
Patients (%)
70
60
50
40
30
20
10
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 7. Percentage of patients in each diabetes risk category according
to fasting glucose levels between 2005 and 2013 Note: Medical histories are not captured in the clinical management software used by GPs.
Therefore it is unknown from the available data who, with a fasting glucose result, has a diagnosis
of diabetes to be able to distinguish between diagnostic testing and glycaemic control.
15
RESULTS
SECTION 1
PART 1
TIME TRENDS ACCORDING TO GENDER
HbA1c
Figure 8 shows that average yearly HbA1c levels were relatively unchanged between
2005 and 2013 for both men and women; they remained at around 7.0% but were
consistently higher for men than women.
10
Men (n=26,813)
Women (n=23,908)
9
HbA1c (%)
8
7
6
5
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 8. Average annual HbA1c levels according to gender between 2005 and 2013
Fasting glucose
Consistent with observed trends in HbA1c levels according to gender, Figure 9
illustrates that there was no change over time in fasting blood glucose results
between 2005 and 2013 for men or women. Men had higher levels than women
(5.7 mmol/L vs. 5.3 mmol/L) for all years.
10
Men (n=201,574)
Women (n=246,788)
Fasting glucose (mmol/L)
9
8
7
6
5
4
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 9. Average annual fasting glucose levels according to gender between 2005 and 2013
16
HYPOGLYCAEMIC TREATMENT
HYPOGLYCAEMIC TREATMENTS PRESCRIBED
ACCORDING TO CLASS
Any hypoglycaemic treatment was prescribed for 24,534 individuals upon initial
presentation (5.2% of total cohort). However it is likely that only people with a HbA1c
measurement (i.e. 50,721), that is used to assess the effectiveness of management
on glycaemic control, had been diagnosed with diabetes so the percentage
would be 48.4%.
Figure 10 shows that non-insulin hypoglycaemic medication was prescribed more
frequently than insulin. The most common class used to treat diabetes were biguanides
(74%) followed by sulphonylureas (43%). Less commonly prescribed non-insulin
hypoglycaemic treatments were thiazolidinediones and DPP4 inhibitors prescribed
for 8.0% and 3% of patients, respectively. Insulin was prescribed for 20% of all patients
being treated.
100
90
80
Patients (%)
70
60
50
40
30
20
10
ide
Me
tig
lin
P1
GL
nh
i bi
tor
P4
Alp
ha
G
luc
Th
iaz
os
ida
oli
din
se
I
ed
DP
ion
es
rea
ylu
on
Su
lph
ua
n
Big
Ins
uli
n
ide
0
Hypoglycaemic treatment
Figure 10. Percentage of patients prescribed hypoglycaemic treatments according to class
17
RESULTS
SECTION 1
PART 1
TIME TRENDS IN CLASS OF HYPOGLYCAEMIC
TREATMENTS PRESCRIBED
Figure 11 highlights the change over time in the proportion of patients prescribed
hypoglycaemic treatments. There was an increase from 66% to 80% in the number of
patients prescribed biguanides. Conversely there was a 20% reduction in sulfonylurea
prescriptions from 49% to 29%. The number of patients prescribed thiazolidinediones
doubled from 5% to 10% in 2005 to 2010, before dropping to 4% in 2013. The advent
of DPP4 inhibitors in 2009 saw a rise to 15% by 2013 of patients prescribed this class
of treatment. Insulin showed a slight increase of 5% to 21% from 2005 to 2013.
All other forms of treatment were rarely prescribed at any time over the study period.
Insulin
Biguanide
Sulphonylurea
Thiazolidinediones
DPP4
Alpha Glucosidase Inhibitor
GLP1
Metiglinide
100
90
80
Patients (%)
70
60
50
40
30
20
10
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 11. Percentage of patients prescribed hypoglycaemic treatments according
to class between 2005 and 2013
18
TRENDS ACCORDING TO HYPOGLYCAEMIC TREATMENT
BLOOD GLUCOSE LEVELS OVER TIME ACCORDING
TO HYPOGLYCAEMIC TREATMENT
Figure 12 shows that the average HbA1c levels of those prescribed insulin (8.4%) was
significantly higher than those prescribed non-insulin hypoglycaemic treatment (7.5%)
or not taking any hypoglycaemic therapy (6.8%). There was some variation in HbA1c
levels over time as a function of treatment; average HbA1c levels were higher in 2011
in people prescribed insulin and in 2008 for people not on any treatment for diabetes.
Otherwise, there was no change in HbA1c levels within groups between 2005 and 2013.
10
Treatment, insulin (n=3,934)
Treatment, non-insulin (n=16,465)
No treatment (n=32,082)
HbA1c (%)
9
8
7
6
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 12. Average annual HbA1c levels according to hypoglycaemic treatment
between 2005 and 2013
Figure 13 shows that the fasting glucose levels of those prescribed insulin
(10.1 mmol/L) was significantly higher than those prescribed non-insulin hypoglycaemic
treatment (8.7 mmol/L) or not taking hypoglycaemic therapy (5.3 mmol/L). Aside from
an increase in fasting glucose levels in 2010 in people prescribed insulin, there was
no change in blood sugar levels within groups from 2005 to 2013.
12
Treatment, insulin (n=3,242)
Treatment, non-insulin (n=15,023)
No treatment (n=431,546)
Fasting glucose (mmol/L)
11
10
9
8
7
6
5
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 13. Average annual fasting glucose levels according to hypoglycaemic
treatment between 2005 and 2013 19
RESULTS
SECTION 1
PART 1
ASSOCIATED RISK FACTORS ACCORDING
TO HYPOGLYCAEMIC TREATMENT
Weight
As shown in Figure 14, weight was significantly lower in untreated individuals (81 kg)
compared to those prescribed either non-insulin or insulin hypoglycaemic medication
(average 90 kg). Hypoglycaemic therapy was also associated with a heavier weight
over time, ranging from approximately 87 kg in 2005 to 90 kg in 2013.
100
Treatment, insulin (n=1,465)
Treatment, non-insulin (n=7,606)
No treatment (n=101,654)
95
Weight (kg)
90
85
80
75
70
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 14. Average annual weight levels according to hypoglycaemic treatment
between 2005 and 2013
Blood pressure
Figure 15 shows that systolic blood pressure levels were slightly lower in non-treated
individuals (131 mmHg) over the period 2005 to 2013 but was similar for people
prescribed either non-insulin on insulin hypoglycaemic therapy (approximately
135 mmHg). Diastolic blood pressure was similar in all three groups, ranging
between 76 and 79 mmHg for insulin and non-treated patients, respectively.
Treatment, insulin (n=2,962)
Treatment, non-insulin (n=14,743)
No treatment (n=239,297)
150
140
Blood pressure (mmHg)
130
120
110
100
90
80
70
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 15. Average annual systolic (upper trends) and diastolic (lower trends) blood
pressure levels according to hypoglycaemic treatment between 2005 and 2013 20
Cholesterol
As shown in Figure 16, total cholesterol levels were higher in untreated individuals
(5.1 mmol/L) compared to those prescribed either non-insulin or insulin hypoglycaemic
medication (average 4.5 mmol/L) for the duration of the time period. There was no
change in cholesterol levels over time or within treatment groups.
Total cholesterol (mmol/L)
6
Treatment, insulin (n=3,477)
Treatment, non-insulin (n=15,837)
No treatment (n=372,519)
5
4
3
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 16. Average annual total cholesterol levels according to hypoglycaemic
treatment between 2005 and 2013
eGFR
As shown in Figure 17, eGFR levels were higher in untreated individuals compared
to individuals prescribed non-insulin or insulin hypoglycaemic medication which each
showed similar levels over the study duration. In all groups, there was an increasing
trend in eGFR levels from 2005 to 2013 which was a greater difference (16 ml/min per
1.73 m2) in those not taking hypoglycaemic treatment compared to those prescribed
non-insulin agents (11 ml/min per 1.73 m2) or insulin (8 ml/min per 1.73 m2) agents.
Treatment, insulin (n=3,696)
Treatment, non-insulin (n=15,641)
No treatment (n=380,018)
100
eGFR (ml/min per 1.73 m2 )
95
90
85
80
75
70
65
60
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 17. Average annual eGFR levels according to hypoglycaemic treatment
between 2005 and 2013
21
RESULTS
SECTION 1
PART 2
GLYCAEMIC CONTROL
HbA1c LEVELS ACCORDING TO HYPOGLYCAEMIC TREATMENT
Figure 18 shows the proportion of patients who achieve HbA1c glycaemic targets
according to hypoglycaemic treatment2. More stringent targets prove more difficult
to achieve as evidenced by 56.6%, 42.3% and 22.5% of patients above the glycaemic
goals of ≥ 6.5%, ≥ 7.0% and ≥ 8.0%, respectively. To illustrate using a reasonable
HbA1c goal of 7.0%, 2 in every 5 patients achieved this goal without pharmacological
treatment (white segment) and a further 15% were effectively managed with
hypoglycaemic medication (red segment). Approximately 1 in 5 patients
who were prescribed treatment did not attain the recommended level of 7.0%
(light blue segment). Another 1 in 5 individuals with a HbA1c level ≥ 7.0% were
not pharmacologically managed and could benefit from treatment.
A: Glycaemic target 6.5%
B: Glycaemic target 7.0%
C: Glycaemic target 8.0%
No treatment + above 8%
(n=5,299: 10.4%)
No treatment + above 7%
(n=10,335: 20.4%)
No treatment + above 6.5%
(n=14,626: 28.8%)
No treatment + below 6.5%
(n=17,456: 34.4%)
Treated + above 6.5%
(n=14,118: 27.8%)
No treatment + below 7%
(n=21,747: 42.9%)
Treated + above 7%
(n=11,094: 21.9%)
No treatment + below 8%
(n=26,783: 52.8%)
Treated + above 8%
(n=6,124: 12.1%)
Treated + below 8%
(n=12,515: 24.7%)
Treated + below 7%
(n=7,545: 14.9%)
Treated + below 6.5%
(n=4,521: 8.9%)
Figure 18. Percentage of patients who achieve glycaemic targets according to treatment
HYPOGLYCAEMIC TREATMENT ACCORDING TO HbA1c CONTROL
Figure 19 shows a small increase from 2005 to 2013, albeit with some year-to-year
variability in the proportion of people below the HbA1c glycaemic target of 7.0% who
are not prescribed hypoglycaemic therapy (41% to 46%; white stack) and who achieve
this target with medication (12% to 17%; red stack). Therefore, there was approximately
10% improvement over time in the percentage of patients who achieved the HbA1c
glycaemic target of 7.0%. More people who achieve the recommendations means less
people who are above target; there was a reduction in the proportion of patients above
the HbA1c goal of 7.0% who were not pharmacologically managed (28% in 2005 to
17% in 2013, dark blue stack) with no change in the percentage of individuals treated
but not at target (19% to 20%; light blue stack).
22
100
No treatment + below 7%
On treatment + below 7%
On treatment + above 7%
No treatment + above 7%
Patients (%)
80
60
40
20
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 19. Percentage of patients who achieve the glycaemic target HbA1c of 7.0%
according to treatment between 2005 and 2013
Figure 20 identifies that individuals above the glycaemic goal of 7.0% were almost
3 times more likely to be prescribed insulin (28.0% vs. 10%). Around 10% more
patients who were effectively managed were prescribed biguanides (79% vs 71%).
More patients above the recommended level of 7.0% were prescribed sulphonylureas
(47.0% vs. 39%), thiazolidinediones (10% vs. 5%) and DPP4 inhibitors (4% vs. 2%).
100
HbA1c ≥ 7% (n = 11,094)
HbA1c < 7% (n=7,545)
90
80
Patients (%)
70
60
50
40
30
20
10
e
P1
ni d
gli
Me
ti
ibi
nh
se
I
os
ida
GL
tor
P4
DP
din
oli
iaz
Alp
ha
Gl
uc
Th
Su
lph
on
ed
ion
es
yl u
rea
ide
Big
ua
n
I ns
uli
n
0
Hypoglycaemic treatment
Figure 20. Percentage of patients prescribed hypoglycaemic treatment according
to class and achievement of the glycaemic HbA1c goal of 7.0%
23
RESULTS
SECTION 1
PART 2
ASSOCIATED RISK FACTORS ACCORDING TO HbA1c CONTROL
Weight
Figure 21 shows that weight was consistently lower in patients who achieved
the glycaemic goal of 7.0% (86 kg) compared to above this level (90 kg). Similar
to observed trends in weight according to hypoglycaemic treatment, there was
a gradual incline in weight from 2005 to 2013.
100
HbA1c < 7% (n=8,625)
HbA1c ≥ 7% (n = 6,315)
95
90
Weight (kg)
85
80
75
70
65
60
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 21. Average annual weight levels according to achievement of the glycaemic
HbA1c goal of 7.0% between 2005 and 2013
Blood pressure
From 2005 to 2013, Figure 22 shows that blood pressure levels were similar for
individuals above or below the glycaemic goal of 7.0% (approximately 136/79 mmHg)
with no change over time.
HbA1c < 7% (n=17,243)
HbA1c ≥ 7% (n = 12,473)
150
140
Blood pressure (mmHg)
130
120
110
100
90
80
70
60
50
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 22. Average annual blood pressure levels according to achievement
of the glycaemic HbA1c goal of 7.0% between 2005 and 2013
24
Cholesterol
Figure 23 shows that from 2005 to 2013, total cholesterol levels did not change
over time and were similar for individuals above or below the glycaemic goal of 7.0%
(approximately 4.7 mmol/L).
6
HbA1c < 7% (n=23,762)
Total cholesterol (mmol/L)
HbA1c ≥ 7% (n = 16,948)
5
4
3
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 23. Average annual total cholesterol levels according to achievement
of the glycaemic HbA1c goal of 7.0% between 2005 and 2013
eGFR
Figure 24 shows that eGFR levels were similar for individuals above or below the
glycaemic goal of 7.0% (approximately 79 ml/min per 1.73 m2), with an approximate
10 ml/min per 1.73 m2 increase overall from 2005 to 2013.
100
HbA1c < 7% (n=22,585)
HbA1c ≥ 7% (n = 16,625)
eGFR (ml/min per 1.73 m2 )
90
80
70
60
50
40
0
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 24. Average annual eGFR levels according to achievement of the glycaemic
HbA1c goal of 7.0% between 2005 and 2013
25
RESULTS
SECTION 2
In this section, all available HbA1c
measurements for an individual patient
were averaged into 6 monthly epochs
from the index date of visit. Patients who
had two or more contiguous blocks with
HbA1c measurements between 2005
and 2013 are included for analyses
and their longest sequence selected for
investigation. For example, if a patient
had 2 visits 6 months apart in 2008 and
5 visits every 6 months from 2010, the
second sequence of 5 visits would
be chosen to analyse.
Figure 25 shows minimal incremental benefit in HbA1c levels with more ongoing
medical evaluation in primary care. HbA1c levels were lowest in people with fewer
contiguous visits and highest in people who had 8 years of visit data (shown as
16 epochs). The range of HbA1c results was confined to between 7.0% and 7.5%.
9
8
HbA1c (%)
BLOOD SUGAR LEVELS
OF REPEAT PATIENT
ENCOUNTERS
7
6
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Contiguous visits (6 monthly epochs)
Figure 25. Average HbA1c levels of contiguous 6 monthly visit data
26
The flow chart in Figure 26 describes patients with contiguous 6 monthly visit data
and their treatment plans according to the timing of prescriptions for non-insulin and
insulin hypoglycaemic treatment. Of a total of 37,044 patients with contiguous data,
half were initially prescribed such therapy (i.e. epoch 1 representing 0-6 months).
Of the individuals not prescribed treatment at the start of their data series (n=18,479),
approximately two thirds (n=11,517) never started hypoglycaemic medication
(light grey text box, refer Part 1 analyses) and one third were initiated on treatment
(red text boxes, refer Part 2 analyses). All individuals prescribed hypoglycaemic
treatment, evident at either their first (n=18,565) or subsequent GP encounters
(n=6,962), with continuing HbA1c evaluation were reviewed for medication type
(light blue text boxes, also refer Part 2 analyses).
Patients with contiguous visits
37,044
Prescribed hypoglycaemic
medication at Visit 1
18,565 (50.1%)
Not prescribed hypoglycaemic
medication at Visit 1
18,479 (49.9%)
Insulin
3,118
No subsequent prescription
for hypoglycaemic medication
11,517
Oral
17,007
Subsequently prescribed
hypoglycaemic medication
6,962
Insulin
1,106
Noninsulin
6,136
Prescribed hypoglycaemic
medication
25,527
Removed
(no follow-up visits)
1,594
Prescribed hypoglycaemic
medication with follow-up
23,933
Insulin ± noninsulin
3,953
Non-insulin
only
19,980
Start insulin
2,018
Figure 26. Flowchart of contiguous 6 monthly visit data according to hypoglycaemic treatment
27
RESULTS
SECTION 2
PART 1
PATIENTS NEVER PRESCRIBED HYPOGLYCAEMIC TREATMENT
For 11,517 individuals who are never prescribed hypoglycaemic medication,
Table 3 identifies the number of epochs within contiguous visit sequence lengths that
the glycaemic goal of 7.0% is not achieved. It can be seen that the majority of people
consistently achieved this goal and were rarely above the recommended level.
Only a small proportion of patients, ranging from 10% to 36%, were above the
recommended HbA1c level of 7.0% in at least half of all visits (grey shaded area).
Table 3. Frequency of epochs within contiguous visit sequence lengths (2-18) that HbA1c levels are above 7.0%
Contiguous visits (6 monthly epochs)
Number
epochs
elevated
(n)
0
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
(3344)
(3161)
(1411)
(1095)
(685)
(565)
(346)
(240)
(211)
(139)
(86)
(92)
(42)
(37)
(14)
(31)
(18)
2584 2567 1084
826
479
392
234
150
117
73
50
46
18
15
4
16
10
1
318
157
106
74
38
47
29
22
23
9
6
7
6
4
2
5
1
2
442
165
74
48
39
26
12
12
5
8
5
6
1
3
2
2
2
272
43
51
34
15
14
9
5
7
4
2
2
1
0
0
0
104
28
18
15
7
7
4
1
0
3
0
0
0
1
0
68
27
19
4
7
5
0
4
5
2
2
0
0
0
50
13
5
1
4
4
2
3
0
0
1
0
0
38
5
4
4
4
0
3
2
0
0
1
2
36
11
7
1
0
1
0
0
0
2
0
17
5
3
0
1
2
2
2
1
0
32
13
3
1
1
4
0
0
0
16
4
4
2
0
0
0
0
8
6
1
1
0
0
0
4
0
2
1
1
0
5
0
1
1
0
3
1
1
0
0
0
0
0
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
enduring
elevation*
2
13 % 14 % 10 % 13 %
14 % 15 % 14 % 17 % 25 %
29 % 17 % 22 % 26 % 32 % 36 % 13 % 17 %
*Grey shading; indicative of approximately 50% of visits with elevated HbA1c levels ≥7.0%
28
RESULTS
SECTION 2
PART 2
PATIENTS INITIATED ON HYPOGLYCAEMIC TREATMENT
Table 4 shows the proportion of individuals who start hypoglycaemic medication
(n=6,962) and the timing of initiation within contiguous visit sequence lengths.
To illustrate, 69% of individuals with 3 contiguous visits are initiated on hypoglycaemic
medication after epoch 2 (i.e. 1 year).
The average time to hypoglycaemic treatment initiation was 18 ± 10 months
[range 12 to 96 months], with no difference between men and women. This did not
vary by the frequency of contact with GPs over the years; the majority of patients were
prescribed hypoglycaemic medication within the first 18 months (epoch 3), with little
evidence of slighter longer delays with more frequent visitation. Overall, 80% of all
patients had begun therapy by this time.
Table 4. Percentage of patients who started hypoglycaemic treatment according to time therapy was initiated and sequence length
Contiguous visits (6 monthly epochs)
Time
treatment
initiated
(6 month 2
%
epoch)
2
100
3
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
%
%
%
%
%
%
%
%
%
%
%
%
%
%
%
%
69
61
59
53
48
51
47
39
42
40
41
41
27
35
48
27
31
24
22
23
21
20
16
19
15
11
13
11
14
17
2
16
15
11
12
14
10
11
11
11
10
11
6
14
12
5
7
8
6
6
6
10
12
9
8
8
6
10
3
11
5
6
6
7
5
7
4
4
5
11
6
2
0
5
5
4
5
4
5
2
5
6
2
9
4
2
2
3
2
6
6
3
2
5
0
9
11
3
3
3
8
4
5
7
2
2
4
3
3
6
3
4
8
8
4
4
2
3
3
4
0
11
2
2
2
1
1
4
0
2
3
3
1
1
1
2
2
2
1
0
0
9
1
0
7
0
0
2
3
0
0
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
TOTAL (n)
0
907
1207
912
852
642
519
444
325
253
206
167
152
124
85
65
46
56
29
RESULTS
SECTION 2
PART 2
Independent of ongoing management, the average HbA1c level when treatment
was first prescribed was 7.4 ± 1.4 %. It was significantly higher in men than women
(7.4 ± 1.4 % vs 7.3 ± 1.3 %, p<0.01) regardless of the delay to initiate treatment
(see Figure 27).
10
Men
Women
HbA1c (%)
9
8
7
6
5
0
2
3
4
5
6
7
8
Treatment initiation time (6 monthly epochs)
Figure 27. Average HbA1c levels that treatment was initiated according to gender
30
Figure 28 shows that upon initiation, the most commonly prescribed hypoglycaemic
medication was biguanide therapy in 73% of all patients. Sulphonylureas were
prescribed for 27% of individuals and insulin for 16%.
100
90
80
Patients (%)
70
60
50
40
30
20
10
tig
lin
ide
P1
ida
Me
ibi
Inh
se
GL
tor
P4
DP
din
oli
os
iaz
Alp
ha
Gl
uc
Th
Su
lph
on
ed
ylu
ion
es
rea
e
nid
ua
Big
Ins
uli
n
0
Hypoglycaemic treatment
Figure 28. Class of hypoglycaemic treatments patients were initiated on PATIENTS INITIATED ON INSULIN THERAPY AFTER
MANAGEMENT WITH NON-INSULIN THERAPEUTICS
As summarised in Figure 26 (light blue text boxes), a total of 2,018 patients (56% men)
were initiated on insulin after ongoing management with non-insulin hypoglycaemic
treatments. The average time from non-insulin to insulin hypoglycaemic therapy was
26 ± 17 months. One third of individuals begun insulin therapy after 12 months, two
thirds by 24 months and 4 in 5 patients were prescribed this treatment after 36 months
of being prescribed non-insulin treatment.
The average HbA1c level at the time insulin was first prescribed following management
with non-insulin hypoglycaemic treatments was 8.6% ± 1.6% and was significantly
higher in men than women (8.6% ± 1.7% vs 8.5% ± 1.6%).
31
CONCLUSIONS
This is the first report of its kind to
evaluate patterns of blood glucose levels
and pharmacological management
of patients attending primary care in
Australia. The data from almost half a
million individuals managed by nearly
1000 GPs in the 9 years beginning from
2005 showed evidence of improvements
in HbA1c results, with about a 10%
reduction in individuals above the
glycaemic goal of 7.0%. Hyperglycaemia
remains to be a concern for clinicians,
patients, scientists and the general
public however, with 30-40% of people
presenting to primary care with blood
glucose assessment having values
above recommended limits for reduced
risk of diabetes (or complications
thereof). There appears to be plenty of
opportunity to improve glycaemic control
with half of patients above target likely
to benefit from starting treatment or
up-titration of treatment.
For decades, the reported prevalence of
diabetes and pre-diabetes has been on
the rise. This has provided the impetus
for increased surveillance of glucose
levels in primary care, supported by
government initiatives and primary care
items to proactively detect and manage
the condition. This undoubtedly explains
our findings in relation to increasing
surveillance with age, particularly from
middle age and onwards.
As with our other reports focussing on
blood pressure and lipid trends and
management in primary care9, 10, we
observed no improvements in average
blood glucose levels over the period
2005 to 2013. These consistent findings
are a reminder of the challenge to
improve antecedent risk factor levels at
the whole population level. Within this
patient cohort, fasting glucose levels
suggested that 8.0% were consistent
with a diagnosis of diabetes and a further
21% with abnormal glucose levels had
high risk of developing the condition.
The management of diabetes remains
dynamic in terms of the observed trends
in pharmacological management and
in subsequent glycaemic control. For
example, we observed an encouraging
trend in achieving the HbA1c goal of
7.0% that coincided with the transition
from older to newer therapeutics from
2011, with only nominal changes in
insulin therapy.
Despite some encouraging trends,
these data are a reminder of the clinical
challenges health professionals face in
preventing and managing diabetes. The
proportion of individuals, either being
actively treated for hyperglycaemia or
not, remains stubbornly high at around
40%. Of these individuals, the rule of half
applies; half were not on treatment and
could benefit from therapy, whilst the
other half were treated but did not
achieve target HbA1c levels and could
benefit from more proactive management.
More frequent patient/GP encounters
did not appear to result in better
glycaemic control. This may reflect
the well described and common
phenomenon of treatment inertia11.
It may also indicate the possibility that
more complex patients attend more
regularly and are harder to achieve
glycaemic control. Therefore, as with the
primary care management of elevated
blood pressure9, there appears to be
clear potential to firstly initiate treatment
earlier and apply more structured and
intensive management to achieve control
in those with persistently elevated blood
glucose levels12. Further investigation of
the individual and physician factors
influencing clinical management of
diabetes in primary care need to be
undertaken, including duration of
diabetes, frequency and severity of
hypoglycaemic events, weight gain and
other side effects of medication. This
would include exploring the application
of reimbursement items (now a major
government expenditure) to promote
best practice in diabetes prevention
and management.
32
LIMITATIONS
This report is a large and unique body
of evidence however consideration must
be given to limitations of data of this kind.
As mentioned in our previous reports9,10
the data were collected as part of routine
clinical practice and not in a systematic
and prospective manner. Specifically,
diabetes status was unknown to be able
to distinguish between measures of
surveillance/screening and glycaemic
control. Furthermore, due to the
de-identified nature of data (both from
a patient and GP perspective), we are
unable to determine if individual patients
consulted another GP and therefore
included twice in the data-set. In addition
to having to apply a series of conservative
assumptions to standardise between
individual comparisons (for example we
only accepted records where the age
and gender of individuals were clearly
identifiable), we have no way of verifying
the veracity of individual data. Moreover,
these data describe a specific primary
care patient cohort and caution should
be applied when making extrapolations
(i.e. beyond within cohort comparisons)
to the wider patient population being
managed within primary care in Australia
and, indeed, the wider population.
In summary, this report provides
bittersweet findings - pleasing in respect
to improved HbA1c levels mixed with
still unacceptably high blood glucose
readings and room for improvement
using new and potentially effective
treatments in combination with more
proactive management. Blood sugar
levels remain at historically high values
necessitating a more intensive approach
to lower blood glucose to levels to
prevent the progression of pre-diabetes
to overt diabetes and to reduce risk of
complications of this insidious condition. 33
REFERENCES
1. Diabetes Australia. Accessed 25/09/2014, from http://www.diabetesaustralia.com.au
2. American Diabetes Association. Accessed on 25/09/2014, from http://www.diabetes.org/
3. Australian Institute of Health and Welfare. Australia’s health 2010. Canberra, ACT, Australia: AIHW cat no. AUS 122
4. Australia Bureau of Statistics. Prevalence of diabetes. Diabetes in Australia: A Snapshot, 2007-08. Accessed on 25/09/2014,
from http://www.abs.gov.au/ausstats/[email protected]/mf/4820.0.55.001
5. Goss, J., (2008). Projection of Australian health care expenditure by disease, 2003 to 2033. Canberra, ACT, Australia,
AIHW cat. no. HWE 43.
6. Royal Australian College of General Practitioners & Diabetes Australia. (2014). General practice management of type 2 diabetes
2014-2015. Accessed on 25/09/2014 from http://www.racgp.org.au/your-practice/guidelines/diabetes/
7. National Health and Medical Research Council (2013). Australian Dietary Guidelines. Canberra: National Health and
Medical Research Council.
8. Department of Health and Aging. Primary Care. (2011). Accessed on 25/09/2014 from http://www.health.gov.au/
9. Carrington, MJ., Jennings, G. & Stewart, S. (2013). Pressure points in primary care: blood pressure and management
of hypertension in 532 050 patients from 2005 to 2010. Journal of Hypertension, 31, 1265-1271.
10. Carrington, MJ. & Stewart, S. (2011). Australia's cholesterol crossroads: An analysis of 199,331 GP patient records.
Melbourne, VIC, Australia: Baker IDI Heart and Diabetes Institute.
11. Argawal, R., Bills, J., Hecht, T. & Light, R. (2011). Role of home blood pressure monitoring in overcoming therapeutic inertia
and improving hypertension control: a systematic review and meta-analysis. Hypertension, 57, 29-38.
12. Stewart, S., Carrington, MJ., Swemmer, C., Anderson, C., Kurstjens, N., Amerena, J., Brown, A., Burrel, L., de Looze, F., Harris,
M., Hung, J., Krum, H., Schlaich, M., Stocks, NP. & Jennings, G. (2012). Effect of intensive structured care on individual blood
pressure targets in primary care: multicentre randomised controlled trial. British Medical Journal, 345, e7156.
34
APPENDIX
35