Glycaemic control of Type 1 diabetes in clinical practice early in the 21st century: An international comparison

Abstract

Aims:
Improving glycaemic control in people with Type 1 diabetes is known to reduce complications. Our aim was to compare glycaemic control among people with Type 1 diabetes using data gathered in regional or national registries.

Methods:
Data were obtained for children and/or adults with Type 1 diabetes from the following countries (or regions): Western Australia, Austria, Denmark, England, Champagne-Ardenne (France), Germany, Epirus, Thessaly and Thessaloniki (Greece), Galway (Ireland), several Italian regions, Latvia, Rotterdam (The Netherlands), Otago (New Zealand), Norway, Northern Ireland, Scotland, Sweden, Volyn (Ukraine), USA and Wales) from population or clinic-based registries. The sample size with available data varied from 355 to 173 880. Proportions with HbA1c < 58 mmol/mol (< 7.5%) and ≥ 75 mmol/mol (≥ 9.0%) were compared by age and sex.

Results:
Data were available for 324 501 people. The proportions with HbA1c 58 mmol/mol (< 7.5%) varied from 15.7% to 46.4% among 44 058 people aged < 15 years, from 8.9% to 49.5% among 50 766 people aged 15-24 years and from 20.5% to 53.6% among 229 677 people aged ≥ 25 years. Sex differences in glycaemic control were small. Proportions of people using insulin pumps varied between the 12 sources with data available.

Conclusion:
These results suggest that there are substantial variations in glycaemic control among people with Type 1 diabetes between the data sources and that there is room for improvement in all populations, especially in young adults.

Full text

Research: Care Delivery
Glycaemic control of Type 1 diabetes in clinical practice
early in the 21st century: an international comparison
J. A. McKnight
1,
*
,a
, S. H. Wild
2,
*
,a
, M. J. E. Lamb
2,a
, M. N. Cooper
3,4
, T. W. Jones
3,4,5
,E.A.
Davis
3,4,5
, S. Hofer
6,b
, M. Fritsch
7,b
, E. Schober
7,b
, J. Svensson
8
, T. Almdal
9
, R. Young
10
,J.T.
Warner
11,c
, B. Delemer
12
, P. F. Souchon
13,d
, R. W. Holl
14,b
, W. Karges
15,b
, D. M. Kieninger
16,b
,
S. Tigas
17
, A. Bargiota
18
, C. Sampanis
19
, V. Cherubini
20,e
, R. Gesuita
21
, I. Strele
22
, S. Pildava
23
,
K. J. Coppell
24
, G. Magee
25
, J. G. Cooper
26
, S. F. Dinneen
27,28,f
, K. Eeg-Olofsson
29,g
, A.-M.
Svensson
30,g
, S. Gudbjornsdottir
29,30,g
, H. Veeze
31
, H.-J. Aanstoot
31
, M. Khalangot
32,h
,W.V.
Tamborlane
33,i
and K. M. Miller
34,i
on behalf of the
a
Scottish Diabetes Research Network
Epidemiology Group,
b
German/Austria DPV database,
c
National Pediatric Diabetes Audit and
the Royal College of Paediatrics and Child Health,
d
CARe
´DIAB Network,
e
RIDI Study Group,
f
Galway University Hospitals Department of Diabetes, Endocrinology and Metabolism,
g
National Diabetes Register in Sweden,
h
Ukrainian Diabetes Register Team and
i
T1D Exchange
Clinic Network
1
Metabolic Unit, Western General Hospital, Edinburgh and University of Edinburgh,
2
Centre for Population Health Sciences, University of Edinburgh, UK,
3
Telethon Kids Institute, The University of Western Australia, Perth,
4
Department of Endocrinology and Diabetes, Princess Margaret Hospital for Children, Perth,
5
School of Paediatrics and Child Health, The University of Western Australia, Perth, Australia,
6
Department of Pediatrics, Medical University of Innsbruck,
7
Department of Pediatrics, Medical University of Vienna, Austria,
8
Department of Pediatrics, Copenhagen University Hospital Herlev,
9
Department of Medicine F,
Copenhagen University Hospital, Hellerup, Denmark,
10
Salford Royal Foundation NHS Trust, Salford,
11
Children’s Hospital for Wales, Cardiff, UK,
12
Department
of Endocrinology, Diabetes and Nutrition, American Memorial Hospital, University Hospital of Rheims,
13
Department of Pediatrics, American Memorial Hospital,
University Hospital of Rheims, France,
14
Institute of Epidemiology and Medical Biometry, ZIBMT, University of Ulm,
15
Division of Endocrinology, RWTH Aachen
University,
16
Diabetes Division, Department of Paediatrics, Universita
¨tsmedizin Johannes Gutenberg Universita
¨t Mainz, Germany,
17
Department of Endocrinology,
University of Ioannina,
18
Department of Endocrinology and Metabolic Diseases, University of Thessaly,
19
Second Department of Internal Medicine, Aristotle
University of Thessaloniki, Hippokratio General Hospital, Thessaloniki, Greece,
20
Department of Women’s and Children’s Health, SalesiHospital, Ancona,
21
Centre
of Epidemiology and Biostatistics, Polytechnic University of Marche, Italy,
22
Department of Public Health and Epidemiology, Riga Stradins University, Riga,
23
The
Centre for Disease Prevention and Control of Latvia, Riga, Latvia,
24
Edgar Diabetes and Obesity Research, Department of Medicine, University of Otago, Dunedin,
New Zealand,
25
Daisy Hill Hospital, Newry, County Down, UK,
26
Norwegian Adult Diabetes Register, Noklus, Bergen, Norway,
27
Galway University Hospitals,
28
NUI Galway, Galway, Ireland,
29
Department of Medicine, Sahlgrenska University Hospital, University of Gothenburg,
30
Centre of Registers in Region
Vo
¨straGo
¨taland, Go
¨teborg, Sweden,
31
Diabeter, National Centre for Pediatric and Adolescent Diabetes, Rotterdam, the Netherlands,
32
Shupyk National Medical
Academy of Postgraduate Education and Komisarenko Institute of Endocrinology and Metabolism, Kiev, Ukrai ne,
33
Yale University, New Haven, CT and
34
Jaeb Centre
for Health Research, Tampa, FL, USA
Accepted 11 December 2014
Abstract
Aims Improving glycaemic control in people with Type 1 diabetes is known to reduce complications. Our aim was to
compare glycaemic control among people with Type 1 diabetes using data gathered in regional or national registries.
Methods Data were obtained for children and/or adults with Type 1 diabetes from the following countries (or regions):
Western Australia, Austria, Denmark, England, Champagne-Ardenne (France), Germany, Epirus, Thessaly and
Thessaloniki (Greece), Galway (Ireland), several Italian regions, Latvia, Rotterdam (The Netherlands), Otago (New
Zealand), Norway, Northern Ireland, Scotland, Sweden, Volyn (Ukraine), USA and Wales) from population or clinic-
based registries. The sample size with available data varied from 355 to 173 880. Proportions with HbA
1c
<58 mmol/
mol (<7.5%) and ≥75 mmol/mol (≥9.0%) were compared by age and sex.
Results Data were available for 324 501 people. The proportions with HbA
1c
58 mmol/mol (<7.5%) varied from
15.7% to 46.4% among 44 058 people aged <15 years, from 8.9% to 49.5% among 50 766 people aged 15–24 years
and from 20.5% to 53.6% among 229 677 people aged ≥25 years. Sex differences in glycaemic control were small.
Correspondence to: John A. McKnight. E-mail: john.mcknight@nhs.net
*Joint senior authors.
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK 1
DIABETICMedicine
DOI: 10.1111/dme.12676

Proportions of people using insulin pumps varied between the 12 sources with data available.
Conclusion These results suggest that there are substantial variations in glycaemic control among people with Type 1
diabetes between the data sources and that there is room for improvement in all populations, especially in young adults.
Diabet. Med.00, 000–000 (2015)
Introduction
It is 20 years since the publication of the Diabetes Control
and Complications Trial (DCCT), which proved beyond
doubt that a lower HbA
1c
in people with Type 1 diabetes
reduces the risk of the development and progression of early
microvascular and neuropathic complications [1,2]. Long-
term follow-up studies of this patient cohort in the Epide-
miology of Diabetes Interventions and Complications (EDIC)
study confirmed the benefits of optimal metabolic control in
reducing the risk of late macrovascular complications [3].
Consequently, management guidelines from many national
and international diabetes associations (e.g. the Scottish
Intercollegiate Guideline Network Management of Diabetes
guideline 116 [4], the American Diabetes Association (ADA)
Standards of Medical Care in Diabetes [5] and International
Society of Pediatric and Adolescent Diabetes [6]) recommend
a target HbA
1c
of 48–58 mmol/mol (6.5–7.5%) for most
people with Type 1 diabetes.
Since publication of DCCT there have been a number of
changes to the treatment of Type 1 diabetes. These include
the release of new insulin analogues designed to have a more
‘physiological’ effect, education programmes primarily tar-
geted at adults with Type 1 diabetes such as dose adjust-
ment for normal eating (DAFNE) [7] and increased use of
improved approaches to delivering continuous subcutaneous
insulin infusion (CSII). Despite these developments, a
number of national and international audits suggest that
the target HbA
1c
is not being achieved in many people with
Type 1 diabetes. These audits have focused mainly on the
paediatric population. For example, mean HbA
1c
has been
shown to vary between 67 and 74 mmol/mol (8.3 and
8.9%) in children attending centres in Europe, Japan,
Australia and North America (n=2873) [8,9]; between
51 and 66 mmol/mol (6.8 and 8.2%) in clinics in Sweden
(n=2280) [10]; and between 62 mmol/mol (7.8%) in
Australia and 91 mmol/mol (10.5%) in Indonesia for clinics
in Asia and the Western Pacific region (n=2312) [11].
Failure to meet recommended standards for control of
Type 1 diabetes among young people has been reported
from the USA [12]. An analysis of trends in metabolic
control in children and adolescents in Germany and Austria
revealed a mean decrease of HbA
1c
from 72 to 65 mmol/
mol (8.7% to 8.1%) [13]. We have not found any similar
published work comparing glycaemic control in representa-
tive adult populations with Type 1 diabetes from different
countries. For example, the EURODIAB IDDM Complica-
tions Study, published 20 years ago, was based on a sample
of 120–140 people from each of 26 centres in 16 European
countries that had attended a hospital clinic in the previous
year, stratified by age into three groups (15–29, 30–44 and
45–60 years), sex and duration in three groups (1–7, 8–14
and 15+years) with an overall response rate of 73% so is
unlikely to be representative [14].
As a first step towards obtaining a broad international
assessment of current treatment outcomes in children and
adults with Type 1 diabetes, colleagues with access to
diabetes registries in 19 countries in Australasia, Europe
and North America agreed to collaborate in a study to
examine and compare HbA
1c
levels in patients enrolled in
their registries.
Methods
A formal collaboration of registries, led by two of the
authors (JAM and SHW) was developed through the
European Diabetes Epidemiology Group, personal contacts
and meetings at international conferences.
A standardized data format was agreed with collaborators
who supplied descriptive data and counts of patients within
HbA
1c
categories by sex, age at date of data extraction, and,
where available, CSII use, from each dataset. Mean or
median HbA
1c
measurement over the country-specific look-
back period was used for each individual in order to
minimize seasonal variation where possible, however, some
countries provided data based on the most recent HbA
1c
value. The periods studied were between 2010 and 2013
(except the New Zealand data, which were from 2005). Data
What’s new?
•We present HbA
1c
data from registries in 19 different
countries describing control in 324 501 people with
Type 1 diabetes, across all age groups.
•These data are the best representation of diabetes care
available and therefore describe the ‘state of the art’.
•We show clearly that Type 1 diabetes control is not as
good as suggested in guidelines, but that some health-
care systems appear to result in better control than
others.
•These data present a challenge to diabetes services.
Leaders in diabetes units/service can compare their local
data to our data and encourage improvement.
2
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK
DIABETICMedicine HbA
1c
in Type 1 diabetes: an international comparison J. A. McKnight et al.

sources were classified as national population-based registers
if >70% of people with Type 1 diabetes in the country were
included, regional population-based registers if >70% of
people with Type 1 diabetes in one or more regions were
included or clinic-based registries. Details for each dataset
are given in the Appendix (Table A1).
Table 1 Characteristics of country populations of >100 individual
Country or region Data source N
Percentage
male
Median mmol/mol
(25th and 75th
centile)
Median %
(25th and
75th centile)
Missing
HbA
1c
(%)
Age <15 years
England National population 15 959 52 68 (61; 81) 8.4 (7.7; 9.6) 22.8
Germany National population 14 838 51 59 (52; 68) 7.6 (6.9; 8.4) 1.4
Denmark National population 1 499 50 64 (57; 72) 8.0 (7.4; 8.7) 15
Austria National population 1 008 55 59 (52; 68) 7.6 (6.9; 8.4) 2.2
Wales National population 975 52 68 (61; 81) 8.4 (7.7; 9.6) 10.4
Latvia National population 281 54 79 (65; 97) 9.4 (8.1; 11.0) 11
Italy (Valle D’Aosta, Liguria,
Marche, Abruzzo, Calabria,
Toscana; provinces: Bari-Foggia,
Messina)
Regional population 972 52 60 (53; 66) 7.6 (7.0; 8.2) 13
Western Australia Regional population 604 51 61 (55; 68) 7.7 (7.2: 8.4) 0
USA Clinic 10 870 52 67 (61; 78) 8.3 (7.7; 9.3) 0.9
Netherlands (Rotterdam) Clinic 506 51 63 (56; 69) 7.9 (7.3; 8.5) 0
France (Champagne Ardenne) Clinic 214 53 64 (59; 72) 8.1 (7.5; 8.7) 0
Age 15–24 years
England and Wales National population 20 939 55 76 (63; 92) 9.1 (7.9; 10.6) 26.8
Sweden*National population 3 856 56 66 (57; 77) 8.2 (7.4; 9.2) 3.3
Scotland National population 3 579 55 78 (66; 93) 9.3 (8.2; 10.7) 11
Denmark National population 2 575 54 69 (60; 81) 8.5 (7.6; 9.6) 5.8
Northern Ireland National population 708 52 73 (60; 81) 8.8 (7.6; 9.6) 25
Latvia National population 400 58 77 (62; 98) 9.2 (7.8; 11.1) 25
Italy (Valle D’Aosta, Liguria,
Marche, Abruzzo, Calabria,
Toscana; provinces: Bari-Foggia,
Messina)
Regional population 1 012 55 62 (55–70) 7.8 (7.2; 8.6) 13
Norway Regional population 520 52 67 (58; 79) 8.3 (7.5; 9.4) 3.7%
Western Australia Regional population 369 55 67 (57; 79) 8.3 (7.4; 9.4) 0
France (Champagne Ardenne) Regional population 206 54 69 (61; 85) 8.5 (7.7; 9.9) 0
Ukraine (Volyn) Regional population 191 49 59 (50; 64) 7.5 (6.7; 8.1) 7.8
Germany Clinic 7 764 55 65 (54; 78) 8.0 (7.1; 9.3) 2.0
USA Clinic 7 189 52 68 (58; 81) 8.4 (7.5; 9.6) 1.1
The Netherlands (Rotterdam) Clinic 684 50 66 (57; 75) 8.2 (7.4; 9.0) 0
Austria Clinic 575 50 64 (54; 76) 8.0 (7.1; 9.1) 2.5
Ireland (Galway) Clinic 198 54 77 (66; 91) 9.2 (8.2; 10.5) 20
Age 25 +years
England and Wales National population 144 840 57 67 (57; 79) 8.3 (7.4: 9.4) 15.0
Scotland National population 20 958 57 69 (60; 81) 8.5 (7.7: 9.5) 6.0
Sweden National population 20 613 55 64 (57; 72) 8.0 (7.4: 8.7) 2.3
Denmark National population 18 648 57 63 (55; 73) 7.9 (7.2: 8.8) 0.5
Northern Ireland National population 4 425 54 67 (58; 79) 8.3 (7.5: 9.4) 32
Norway Regional population 3 258 53 63 (56; 72) 7.9 (7.3: 8.7) 5.4
Latvia National population 1 267 53 67 (57; 79) 8.3 (7.4: 9.4) 26
Ukraine (Volyn) Regional population 511 58 57 (50; 67) 7.4 (6.7: 8.3) 39
France (Champagne Ardenne) Regional population 443 51 64 (56; 75) 8.0 (7.3: 9.0) 0
Italy (Valle D’Aosta, Liguria,
Marche, Abruzzo, Calabria,
Toscana; provinces: Bari-Foggia,
Messina)
Regional population 336 58 58 (52–68) 7.5 (6.9-8.4) 11
New Zealand (Otago) Regional population 300 55 67 (58; 77) 8.3 (7.5; 9.2) 0
USA Clinic 7 461 46 59 (52; 67) 7.5 (6.9; 8.3) 1.8
Germany Clinic 5 302 53 57 (49; 69) 7.4 (6.6; 8.5) 7.0
Ireland (Galway) Clinic 927 57 67 (58; 77) 8.3 (7.5; 9.2) 31
Greece (Epirus, Thessaly and
Thessaloniki)
Clinic 188, 167 49, 43 60 (52; 68) 7.6 (6.9; 8.4) 0
Austria Clinic 216 46 66 (56; 75) 8.2 (7.3; 9.0) 6.5
The Netherlands (Rotterdam) Clinic 129 47 58 (52; 65) 7.5 (6.9; 8.1) 0
Nincludes people with missing data.
*
Data for Sweden are for 18-24 year olds
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK 3
Research article DIABETICMedicine

Statistical analyses
All analyses were conducted in three age strata: <15 years,
15–24 years and ≥25 years. Where data were available for
fewer than 100 people from a given country or region these
were excluded from the analysis. Descriptive statistics were
reported for each country or region. Proportions of individuals
with HbA
1c
<58 mmol/mol (<7.5%), 58–74 mmol/mol
(7.5–8.9%) and ≥75 mmol/mol (≥9%) were calculated by
age, sex and source of data (populationor clinic). Multivariable
logistic regression was used to estimate odds ratios for HbA
1c
<58 mmol/mol (<7.5%) compared with HbA
1c
≥58 mmol/
mol (≥7.5%) by age stratum, sex, and country within a single
model containing all variables. Comparison groups for each
variable were based on the largest subgroup (i.e. ≥25 years of
age, male and England and Wales). Percentages of individuals
using insulin pumps (CSII) were reported, where available, by
sex and country or region within the three age strata.
Results
HbA
1c
data were obtained from 324 501 people with Type 1
diabetes from Western Australia, Austria, Denmark, Eng-
land, Champagne-Ardenne in France, Germany, Epirus,
Thessaly and Thessaloniki in Greece, Galway in Ireland,
several Italian regions, Latvia, Rotterdam in The Nether-
lands, Otago in New Zealand, Norway, all adult hospital
clinics in Northern Ireland, Scotland, Sweden, Volyn in
Ukraine, selected clinics in the USA and Wales. Data were
available for people aged <15 years from 11 countries or
regions, aged 15–24 years from 16 countries or regions and
aged ≥25 years from 17 countries or regions. The charac-
teristics of the populations are shown in Table 1, stratified by
age category, by data source and in decreasing order of
population size. The total sample size varied from 355
(Greek clinic data) to 213 960 (English and Welsh National
Diabetes Audit data), although the sample size reduced after
exclusion of missing data and restriction to country or region
and age-specific samples of 100 or more people. Chi-squared
tests comparing differences in HbA
1c
distribution by age, sex
and data source all yielded statistically significant results at
P<0.0001 as a consequence of the large sample size.
Median HbA
1c
levels by age group and country are shown
in Table 1. Median HbA
1c
varied between 57 and 79 mmol/
mol (7.4% and 9.4%) for different countries and age groups.
In <15 year olds, median HbA
1c
was highest in Latvian
children and lowest in Austrian, German and Italian
children, whereas in the other two age strata, median HbA
1c
was highest in Scottish individuals and lowest in Ukrainian
individuals. The overall proportions with HbA
1c
≥58 mmol/
mol (≥7.5%) were lower in data from clinics than from
population-based data sources (66.2% vs. 73.0%). The
highest proportion with an HbA
1c
≥58 mmol/mol
(≥7.5%) was among Irish individuals aged 15–24 years
(91.1%) and the lowest was among German individuals aged
≥25 years (46.4%). The highest proportion with HbA
1c
≥75 mmol/mol (≥9%) was found amongst Latvian individ-
uals aged <15 years (58.0%) and the lowest proportion
with HbA
1c
≥75 mmol/mol (≥9%) was among Dutch
individuals in the ≥25 years age group (7.8%).
The distribution of HbA
1c
levels by age category and sex in
the total population is shown in Table 2 and by country or
region, age group and data source in Fig. 1. The 15–24-year-
old age group had lower proportions of people with HbA
1c
<58 mmol/mol (<7.5%) than the other age groups and
proportions of people with HbA
1c
<58 mmol/mol (<7.5%)
were lower among girls/women than boys/men in all age
groups. The multivariable regression model yielded odds
ratios (OR; 95% confidence intervals [CI]) for HbA
1c
<58 mmol/mol (<7.5%) of 0.58 (0.56–0.60) and 0.51
(0.50–0.52) for people aged <15 and 15–24 years compared
with people aged ≥25 years, respectively; of 0.89 (0.88–
0.91) for females compared with males; and of between 0.67
(0.66–0.70) for Scotland and 3.43 (3.32–3.55) for Germany
compared with England and Wales. After applying target
HbA
1c
levels recommended by different diabetes associa-
tions, we found that 7.1% of people in the whole database
had an HbA
1c
<48 mmol/mol (<6.5%), 8.7% had an
HbA
1c
of 48–53 mmol/mol (6.5–6.9%) and 12.3% had an
HbA
1c
of 53–58 mmol/mol (7.0–7.4%).
There was wide variation in the use of CSII between the
populations studied (Table 3). This may due to selection bias
in the clinic populations, although there remains wide
variation (1–93%) in the use of CSII reported by popula-
tion-based registers. In the majority of age and country
groups, a higher proportion of females than males used CSII.
Discussion
We have described patterns of glycaemic control in a large
number of people with Type 1 diabetes from 19 different
countries or regions across the world. These data demon-
strate that the majority of people with Type 1 diabetes have
higher HbA
1c
levels than recommended in guidelines based
on the published evidence [4–6] and are therefore at
increased risk of microvascular and macrovascular compli-
cations than people with better glycaemic control. The
limitations of the existing data for making comparisons
between countries are discussed further below.
The data also indicate that HbA
1c
levels are higher in those
aged 15–24 years than among other age groups across many
countries. There are abundant data that illustrate the special
challenges in managing teenagers and young adults with
Type 1 diabetes, including the insulin resistance of puberty
and the emotional lability that characterizes that developmen-
tal stage. However, there is an emerging body of evidence that
young adults also face problems with controlling their diabe-
tes. Adults in their early twenties have many changes in
circumstances associated with leaving the parental home and
developing a career, and these changes may make the
4
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK
DIABETICMedicine HbA
1c
in Type 1 diabetes: an international comparison J. A. McKnight et al.

management of their diabetes more difficult. The transition
from paediatric to adult care usually occurs between 15 and
25 years of age, and if not well managed, can result in erratic
care and poorer outcomes [15]. A more detailed analysis of
control within this age category comparing different social
structures might provide further insights and inform strategies
to support people with Type 1 diabetes at that stage of life.
We identified significant differences in diabetes control
between countries or regions. The different approaches to
defining Type 1 diabetes might have resulted in a greater
proportion of people with Type 2 diabetes being included in
some populations, which may introduce bias towards better
glycaemic control. The inclusion of data for women with
diabetes during pregnancy and differing proportions of
pregnant women in the different datasets may also influence
comparisons between datasets. In future comparisons, it
would also be helpful to explore the extent to which
differences between countries might be explained by differ-
ences in duration of diabetes or the prevalence of complica-
tions of diabetes, but duration data were incomplete and
complication data were not available for this analysis. This
simple analysis was undertaken without additional funding
to provide information for subsequent applications to
support more detailed data collection and analysis.
Further possible explanations for differences between
countries or regions include, for example, differences in
diabetes education, frequency of contact between patient
and clinical teams, ethnic/racial/cultural diversity, the impact
of low socio-economic status, rural/urban differences linked to
distance between home and the diabetes centre, access to
specialist care and treatments influenced by funding decisions
made by governments or private insurance policies, training of
diabetes specialists and general practitioners and differences in
HbA
1c
treatment goals recommended by national or local
guidelines. Among the countries that provided national
population data there were no obvious associations between
quality of glycaemic control and health service type, defined as
automatic health coverage funded by taxes (as in Denmark,
Norway, Sweden and the UK), social health insurance (as
available in Austria and Germany), gross domestic product
(GDP) per capita which varied from $23 028 for Latvia to
$65 461 for Norway or proportion of GDP spent on health
care, which varied from 6.0% for Latvia to 12.4% for The
Netherlands based on World Bank 2013 data.
Another limitation of the study is that, because of selection
bias, HbA
1c
levels might be even higher than we have reported.
For example, levels of control in clinic-based populations may
be better than in the wider populations from which they were
drawn and this may influence country or regional compari-
sons. In addition, population-based samples are subject to
potential selection bias if ascertainment of cases of Type 1
diabetes and HbA
1c
data are incomplete. For example, the
estimated case ascertainment varies from 99% in Scotland
with <10% missing HbA
1c
,to~18% ascertainment and 5%
missing HbA
1c
in the Norwegian sample. We have, however,
Table 2 Frequency (%) of people in different HbA
1c
categories by age and sex
Age group
<15 years 15–24 years ≥25 years Total
HbA
1c
category Male Female Male Female Male Female Male Female
<58 mmol/mol (<7.5%) 6 891 (30) 6 261 (29) 6 451 (24) 4705 (20) 38 183 (30) 28 033 (28) 51 525 (29) 38 999 (27)
58–74 mmol/mol (7.5–8.9%) 10 453 (46) 9 760 (46) 9 746 (36) 8 410 (36) 53 044 (41) 41 531 (41) 73 243 (41) 59 701 (41)
≥75 mmol/mol (≥9%) 5 388 (24) 5 305 (25) 11 274 (40) 10 179 (44) 37 556 (29) 31 328 (31) 54 218 (30) 46 812 (32)
Total 22 732 (100) 21 326 (100) 27 471 (100) 23 294 (100) 128 783 (100) 100 892 (100) 178 986 (100) 145 512 (100)
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK 5
Research article DIABETICMedicine

made a judgement as to whether each dataset adequately
represents an overall population based on whether it covers
>70% of people with Type 1 diabetes and have defined them
as such. We recognize that some bias is likely to remain within
these population samples, particularly when they represent a
subset of one or more regions within a country. For example, it
seems likely that, within any population, those with poor
control who do not engage with healthcare services would be
least well represented within our data. We encourage all
healthcare systems to develop methods to monitor key aspects
of diabetes care for their whole population with diabetes as this
is the only way to truly measure this robustly.
Another potential source of bias is the measurement
method for HbA
1c
. The International Federation of Clinical
(a) (b)
(c) (d)
(e) (f)
FIGURE 1 Proportions of individuals in each HbA
1c
category in each of the three age strata by population or clinic data source: <15 years old (a,
population; b, clinic); 15–24 years old (c, population; d, clinic); ≥25 years old (e, population; f, clinic). Eng&Wales, England and Wales; W.Aust,
Western Australia; (R), one or more regions within a country (see text for details); NL, The Netherlands; N. Ireland, Northern Ireland; Otago NZ,
region of Otago in New Zealand.
6
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK
DIABETICMedicine HbA
1c
in Type 1 diabetes: an international comparison J. A. McKnight et al.

Chemists (IFCC) standardization or a similar quality assur-
ance scheme was used in all of the healthcare systems
involved in this comparison, but it is still possible that results
from different populations are not directly comparable.
However, it seems unlikely that any variation would account
for the major differences between countries that we
observed. In addition, for audit purposes, the HbA
1c
values
in the first 3 months after diagnosis of diabetes may be
discounted because they do not reflect quality of diabetes
care. It is likely that the proportion of people with newly
diagnosed diabetes varied between registries and this
accounts for some of the differences between data sources.
Different summary measures of HbA
1c
were used for
different populations, but again it seems unlikely that this
would account for all the differences between populations.
Our analysis was based on three predefined age groups
reflecting the organization of services in differentcountries. The
age distribution within the >25 years age group varied in
different populations and this might have some influence on
overall comparisons. The proportion of people with HbA
1c
<58 mmol/mol (<7.5%) increased with age >39 years (from
25% among 40–49-year-olds to 34% in people aged
>60 years; n=57 387 and 59 038, respectively). The reasons
for this are unclear, but might relate to survival bias or more
aggressive glucose management of people with complications.
Some of the potential biases in our comparisons between
countries or regions have been addressed by the Hvidøre
Study group. In addition to the international comparisons of
glycaemic control performed by the Hvidøre Study group
mentioned previously, it identified variations in mean HbA
1c
levels measured at a central laboratory of between 7.4% and
9.2% for 2093 11–18-year-olds attending clinics in 21
different centres during 2005 [16]. The variation persisted
after adjustment for age, duration of diabetes, type of insulin
regime, daily dose of insulin and BMI. In a previous analysis,
the Hvidøre Study group found that differences in overall
control between centres persisted 3 years after the feedback
of results to the centres [17]. The Hvidøre Study group also
noted slightly higher HbA
1c
results in females compared with
males for 11–18-year-olds [17]. We have identified statisti-
cally significant differences in glycaemic control between
males and females of all ages, but the clinical significance of
this is unclear.
The current report is the first very large international study
of glycaemic control in people with Type 1 diabetes covering
all age groups. It is important to note that, despite the
limitations in the comparisons, the individual HbA
1c
values
summarized in this analysis are used by clinicians to support
decision-making in their practice and it is clear that there is
scope for improvement in the glycaemic control of Type 1
diabetes in all countries. In common with other large-scale
epidemiological comparisons, this work generates important
questions that require further investigation. Why do the
majority of people with Type 1 diabetes in all populations
included still have high-risk glucose levels if the findings
cannot be explained by bias? Is it predominantly a funda-
mental limitation of the available therapies or are optimal
treatments and systems of care not being deployed? Is it
possible to collect further data to explain the between-
country differences in HbA
1c
levels and use the information
to improve outcomes for patients?
We hope that these data will act as a stimulus to all
countries to critically evaluate obstacles to better treatment
outcomes and, where modifiable, to take corrective action.
The findings of this study should also encourage individual
practitioners and clinics to regularly review the HbA
1c
profile
of their local population and to institute quality improve-
ment programmes where needed to optimize glucose control
Table 3 Frequency (%) of insulin pump users, by country or region, sex and age group
<15 years 15–24 years ≥25 years
Country Data source Male Female Male Female Male Female
Western Australia Regional population 105 (34) 119 (40) 58 (29) 78 (47) ––
Austria Clinic 204 (37) 215 (48) 65 (22) 99 (35) 83 (83) 96 (83)
Denmark National population 391 (58) 441 (65) 281 (20) 317 (27) 475 (4) 764 (10)
France (Champagne Ardenne) Regional population 92 (83) 91 (93) 59 (59) 41 (50) 37 (22) 49 (34)
Germany Clinic 2940 (39) 3 153 (44) 1246 (29) 1339 (38) 544 (19) 760 (30)
Greece
(Epirus, Thessaly and
Thessaloniki)
Clinic –– –– 16 (10) 32 (17)
Italy (Valle D’Aosta, Liguria,
Marche, Abruzzo, Calabria,
Toscana; provinces:
Bari-Foggia, Messina)
Regional population 109 (22) 110 (24) 77 (14) 85 (19) 30 (15) 35 (25)
Latvia National population 23 (15) 15 (12) 19 (8) 10 (6) 7 (1) 19 (3)
Netherlands (Rotterdam) Clinic 188 (73) 179 (72) 175 (51) 205 (60) 41 (67) 54 (79)
Norway Clinic –– 90 (34) 74 (30) 430 (26) 458 (31)
Sweden National population –– 540 (26) 516 (31) 1727 (16) 2153 (25)
U.S.A. Clinic 2533 (45) 2523 (48) 1788 (48) 1793 (52) 1783 (52) 2622 (65)
–, no data available.
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK 7
Research article DIABETICMedicine

consistent with best achievable long-term health for their
patients.
Funding sources
This work has had local funding support for registries, but
there was no specific funding for this work.
Competing interests
None declared.
Acknowledgements
John McKnight and Sarah Wild are guarantors for this work.
The Austrian and the German DPV group thank all
participating centres for providing anonymized longitudinal
data. The documentation system and data management are
provided by Katharina Fink, Andreas Hungele and Esther
Molz. The DPV-Initiative is funded by the Competence Net
Diabetes Mellitus (FKZ 01GI1106).
Data from the NPDA for England and Wales were
submitted on behalf of the Royal College of Paediatrics
and Child Health. We thank analysts Andrew Hayton and
Swarna Khare for analytical help with the England and
Wales data.
Statistical help with the data from France was provided by
Alpha Diallo.
Ioanna Zografou, MD, Barbara Nikolaidou, MD, MSc also
contributed to the collection of the information in Greece.
We would like also to thank the collaborators to the RIDI
study group, from north to south Italy: Marisa Bechaz (Valle
d’Aosta region), Silvano Piffer, Vittoria Cauvin (Trento
province), Giuseppe D’Annunzio (Liguria region), Lorenzo
Iughetti, Barbara Predieri (Modena province), Lorenzo Lenzi,
Sonia Toni (Firenze province), Antonio Iannilli, Chiara
Giorgetti (Marche region), Stefano Tumini (Abruzzo region),
Mariella Bruzzese, Franco Mamm
ı (Calabria region), Fortu-
nato Lombardo, Giuseppina Salzano (Messina province)
Elvira Piccinno, Federica Ortolani, Maurizio Delvecchio
(Bari-Foggia provinces).
The Otago regional data were available thanks to the
former Otago Diabetes Team who established the regional
register, participating GPs, practice nurses, and diabetic
patients, Project Nurses Karen Anderson, Ruth Gardener,
Claire Lamb and Pauline Mumm for assisting with data
collection and data entry, and the following former organi-
zations who provided funding: Southern Regional Health
Authority, Health Funding Authority, Otago District Health
Board and Otago Diabetes Research Trust.
Members of the Galway University Hospitals (GUH)
Department of Diabetes, Endocrinology and Metabolism
include Dr Marcia Bell, Prof Fidelma Dunne, Dr Francis
Finucane, Prof Timothy O’Brien and Dr Esther O’Sullivan.
Data were extracted from the GUH Diabetes Clinical
Information System (DIAMOND). We are grateful to Ms
Marie Gately, Diabetes Centre Administrator for mainte-
nance of the database and Mr Vincent Jordan, Programme
Manager, ICT Implementation Services, Merlin Park, Gal-
way for assistance with data extraction.
The Scottish data were available for analysis by members of
the Scottish Diabetes Research Network (SDRN) thanks to the
hard work and dedication of health service staff across
Scotland who enter the data and people and organizations
(the Scottish Care Information-Diabetes Collaboration Steer-
ing Group, the Scottish Diabetes Group, the Scottish Diabetes
Survey Group, the managed clinical network managers and
staff in each health board) involved in setting up, maintaining
and overseeing the Scottish Care Information-Diabetes Col-
laboration, the national electronic diabetes register for Scot-
land. We thank Andy Judson and John Kernthaler for
performing the data extraction and Lesley Gardner of the
University of Edinburgh for submitting the manuscript.
Ukrainian diabetes register team members were Volodymir
Kovtun, IT Ing., Larisa Djankarashvili, MD, Liudmila
Koreniuk, MD.
References
1 The Diabetes Control and Complications Trial Research Group.
The effect of intensive treatment of diabetes on the development
and progression of long-term complications in insulin-dependent
diabetes mellitus. N Engl J Med 1993; 329: 977–986.
2 The Diabetes Control and Complications Trial/Epidemiology of
Diabetes. Interventions and Complications Research Group. Reti-
nopathy and nephropathy in patients with type 1 diabetes four
years after a trial of intensive therapy. N Engl J Med 2000; 342:
381–389.
3 Nathan DM, Cleary PA, Backlund JY, Genuth SM, Lachin JM,
Orchard TJ et al. Intensive diabetes treatment and cardiovascular
disease in patients with type 1 diabetes. N Engl J Med 2005; 353:
2643–2653.
4 Scottish Intercollegiate Guidelines Network. Management of Dia-
betes: A National Clinical Guideline. Edinburgh: Scottish Intercol-
legiate Guidelines Network, 2010. Report No. 116.
5 Standards of medical care in diabetes–2013. Diabetes Care 2013;
36(Suppl. 1): S11–S66.
6 IDF/ISPAD. 2011 Global Guideline for Diabetes in Childhood and
Adolescence. Available at https://www.ispad.org/resource-type/id-
fispad-2011-global-guideline-diabetes-childhood-and-adolescence
Last accessed 1 December 2013.
7 DAFNE Study Group. Training in flexible, intensive insulin
management to enable dietary freedom in people with type 1
diabetes: dose adjustment for normal eating (DAFNE) randomised
controlled trial. BMJ 2002 October 5; 325: 746.
8 Mortensen HB, Hougaard P. Comparison of metabolic control in a
cross-sectional study of 2,873 children and adolescents with IDDM
from 18 countries. The Hvidøre Study Group on Childhood
Diabetes. Diabetes Care 1997; 20: 714–720.
9 Mortensen HB, Robertson KJ, Aanstoot HJ, Danne T, Holl RW,
Hougaard P et al. Insulin management and metabolic control of
type 1 diabetes mellitus in childhood and adolescence in 18
countries. Hvidøre Study Group on Childhood Diabetes. Diabet
Med 1998; 15: 752–759.
10 Hanberger L, Samuelsson U, Lindblad B, Ludvigsson J. A1C in
children and adolescents with diabetes in relation to certain clinical
8
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK
DIABETICMedicine HbA
1c
in Type 1 diabetes: an international comparison J. A. McKnight et al.

parameters: the Swedish Childhood Diabetes Registry SWEDI-
ABKIDS. Diabetes Care 2008; 31: 927–929.
11 Craig ME, Jones TW, Silink M, Ping YJ. Diabetes care, glycemic
control, and complications in children with type 1 diabetes from
Asia and the Western Pacific Region. J Diabetes Complications
2007; 21: 280–287.
12 Wood JR, Miller KM, Maahs DM, Beck RW, DiMeglio LA,
Libman IM et al. Most youth with type 1 diabetes in the T1D
Exchange Clinic Registry do not meet American Diabetes Associ-
ation or International Society for Pediatric and Adolescent Diabetes
clinical guidelines. Diabetes Care 2013; 36: 2035–2037.
13 Rosenbauer J, Dost A, Karges B, Hungele A, Stahl A, Bachle C
et al. Improved metabolic control in children and adolescents
with type 1 diabetes: a trend analysis using prospective multi-
center data from Germany and Austria. Diabetes Care 2012; 35:
80–86.
14 The EURODIAB IDDM Complications Study Group. Microvascu-
lar and acute complications in IDDM patients: the EURODIAB
IDDM Complications Study. Diabetologia 1994; 37: 278–285.
15 Lyons SK, Libman IM, Sperling MA. Diabetes in the adolescent:
transitional issues. J Clin Endocrinol Metab 2013; 98: 4639–4645.
16 de Beaufort CE, Swift PG, Skinner CT, Aanstoot HJ, Aman J,
Cameron F et al. Continuing stability of center differences in
pediatric diabetes care: do advances in diabetes treatment improve
outcome? The Hvidøre Study Group on Childhood Diabetes.
Diabetes Care 2007; 30: 2245–2250.
17 Danne T, Mortensen HB, Hougaard P, Lynggaard H, Aanstoot HJ,
Chiarelli F et al. Persistent differences among centers over 3 years in
glycemic control and hypoglycemia in a study of 3,805 children and
adolescents with type 1 diabetes from the Hvidøre Study Group.
Diabetes Care 2001; 24: 1342–1347.
18 McKnight JA, Morris AD, Cline D, Peden N, Fischbacher C, Wild
S. Implementing a national quality assurance system for diabetes
care: the Scottish Diabetes Survey 2001–2006. Diabet Med 2008;
25: 743–746.
19 Beck RW, Tamborlane WV, Bergenstal RM, Miller KM, DuBose
SN, Hall CA. The T1D Exchange clinic registry. J Clin Endocrinol
Metab 2012; 97: 4383–4389.
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK 9
Research article DIABETICMedicine

Appendix 1
Table A1 Description of data sources in each population
Country
Sample
representativeness Time Value
Sample
size/age
HbA
1c
method
alignment
Validation of Type 1
diabetes diagnosis
Australia (Western) >99% case ascertainment
verified
July 2012 to
June 2013
Mean value 973 under 18
years old
IFCC Clinical and
antibody status
Austria
DPV-Initiative
Austria–voluntary
clinic participation
>80%
children, clinic-based adults
1 September 2010 to
1 December 2011
Mean 1 852 (all ages) DCCT Clinical diagnosis
by diabetes
specialists, based
on insulin requirement
and b–cell antibodies
Denmark
(DanDiabKids
DVDDB =Danske
Voksen Diabetes
Database
Estimated 99.9% completeness
children
92% complete for adult
population
1 March 2011 to
28 February 2012
Single value 4 000 [0–18(+)]
20 274
IFCC Algorithm
based on clinical record,
age at diagnosis and
treatment patterns
England Adult data received (secondary
care and (87.9% of all GP
practices in England and Wales)
January 2011 to
March 2012
Single, most
recent value
173 880 IFCC Based on Read
codes (routine
data registration)
France Data from CAReDIAB network
in Champagne Ardenne
1 January 2011 to
31 Dec 2012
Mean 863
Estimate 70%
of those aged
<15 years and
30–40% of those
aged >15 years
Not
aligned
Clinical history
and antibodies
(for some patients)
by an endocrinologist
Germany
DPV-Initiative
Germany
>94% children, clinic-based adults
Adult data from 105 treatment
centres
1 September 2010 to
1 December 2011
Mean 28 664 (all ages) DCCT All diagnoses by
diabetes specialists
Greece Patients attending two university
hospital clinics in Epirus and
Thessaly, as well as a hospital
clinic in Thessaloniki
October 2011 to
October 2013
Latest HbA
1c
value
Only those with
results included;
355 in total
IFCC Diagnosis based
on endocrinologist
opinion from clinical
history, age at diagnosis,
treatment pattern and
antibody screen if
appropriate
Italy
Eight local registries
(whole regions:
Valle D’Aosta, Liguria,
Marche, Abruzzo,
Calabria,
Toscana; provinces:
Bari-Foggia, Messina)
Population –estimated 99%
completeness for children
September 2010 to
August 2011
Mean 2 664 aged
<55 years
IFCC Algorithm based
on clinical
record, age at
diagnosis and
treatment patterns
Latvia
Centre for Disease
National population 1 January 2010 to
31 December 2011
Mean but usually
only one valued
2 556
(all ages)
IFCC In children, diagnosis
must be confirmed in
a children’s clinical
10
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK
DIABETICMedicine HbA
1c
in Type 1 diabetes: an international comparison J. A. McKnight et al.

Table A1 (Continued)
Country
Sample
representativeness Time Value
Sample
size/age
HbA
1c
method
alignment
Validation of Type 1
diabetes diagnosis
Prevention
and Control (CDPC)
university hospital.
Distinction between diabetes
types may be less
accurate in adults because
there can be atypical cases.
There are some built-in
‘rules’ in the register
system as well, e.g.
patient must be on insulin
therapy. Only those with
diagnosis <30 years are
included in the data
The Netherlands 1 468 patients who receive
their regular diabetes care
at our centre
Last available value in
2013 and not more
than 1 year ago
(~96% in last 3 months)
Last value 1 319
(all ages)
IFCC Type 1 diagnosis based on
clinical data (symptoms,
DKA, pH, bicarbonate,
dehydration) +antibody
status (GAD, IA2, if these
two are negative: ICA)
New Zealand Otago region ~83% complete 2005 Most recent 385 (all ages),
but infants and
children are
under-represented
IFCC Checking hospital diabetes
clinic medical records (age
at diagnosis, antibody results,
clinical management patterns)
Northern Ireland Patients registered on specialist
clinic database.
All patients with Type 1 should be
referred to this service
2012 Most recent value.
If no value in 2012,
then recorded as
missing
5 133 aged >
18 years
IFCC Specialist clinic diagnosis of
Type 1 diabetes confirmed at
registration on database
system
Norway Voluntary participation. Estimated
completeness 18% (adult)
1 September 2010 to
31 December 2011
Most recent 3 778 adults IFCC Based on the diagnosis recorded
by the hospital doctor/GP. Use
of GAD, IA2 and C-peptide
relatively common
Republic of Ireland Galway region, clinic
database data
1 January 2012 to
11 December 2013
Mean 1 125
Missing data if
no attendance and
therefore no
HbA
1c
taken
IFCC Specialist endocrinologist
formally assess diagnosis at
entry into database
Scotland National population. Estimated
99.5% completeness (adults only)
1 December 2009 to
31 November 2011
Mean 28 063 (all ages) IFCC Algorithm based on
clinical record,
age at diagnosis
and treatment
patterns
Sweden Approximately 70% of the national
Type 1 diabetes population
2011 Most recent
HbA
1c
25 088 aged ≥
18 years
IFCC Treatment with insulin only
and onset age of diabetes
<30 years
Ukraine 99% completeness
region
2011 Mean 1 056 all ages IFCC Algorithm based on clinical
record,
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK 11
Research article DIABETICMedicine

Table A1 (Continued)
Country
Sample
representativeness Time Value
Sample
size/age
HbA
1c
method
alignment
Validation of Type 1
diabetes diagnosis
age at diagnosis (<30 years)
and treatment patterns (insulin)
USA
67 clinics
Sample –Type 1 diabetes
patients
who consented to enrol
in the T1D Exchange
Clinic Registry
September 2010 to
July 2012
Mean over year
prior to enrolment
25 833 aged <1
to 93 years
IFCC Clinical diagnosis of presumed
autoimmune Type 1 diabetes
and either islet cell antibodies
present or if antibodies
were negative or unknown,
then insulin must have
been started at or shortly
after diagnosis and used
continually thereafter
(except in the case of a
pancreas or islet cell
transplant)
Appendix 2
Narrative description of data sources
Australia
Australian paediatric data are for the period 1 July 2012 to
30 June 2013 and are based on a population-based register
that is estimated to be >99% complete for children
<16 years old in Western Australia, which has a population
of 2.43 million. Type 1 diabetes status is defined on the basis
of clinical criteria and antibody status.
Austria
Twenty-two Austrian centres (16 paediatric/6 adult) partic-
ipated in the analysis. The diabetes centres are all hospital-
based outpatient settings and use the German DPV system
(described below). The participating centres cover ~70% of
paediatric patients with Type 1 diabetes. The participating
adult centres use the DPV documentation system primarily
for pump users.
Denmark
In Denmark, >98% of patients with Type 1 diabetes are
cared for in hospital outpatient clinics. It is mandatory for
these clinics to report data on a yearly basis to the Danish
Adults Diabetes Registry (DVDD). Information obtained was
collected in the period 1 March 2011 to 28 February 2012. It
is estimated that the data provided covers around 92% of the
adult Danish Type 1 diabetes population.
DanDiabKids was initiated in 1996 as a tool for quality
assessment in diabetes care in children and adolescents.
National results are published in yearly reports. The register
is web-based and 19 centres annually report clinical data
obtained during outpatient visits and send one HbA
1c
sample
for central assessment. Included data are from status visits in
2010. Reporting to the register is mandatory. The ascertain-
ment rate is >95%.
England and Wales
The National Diabetes Audit (NDA) and the National
Paediatric Diabetes Audit (NPDA) are commissioned by the
Healthcare Quality Improvement Partnership as part of the
National Clinical Audit and Patient Outcomes Programme
following advice to the English and Welsh Departments
of Health from the National Advisory Group on Clinical
Audit and Enquiries (http://www.hscic.gov.uk/nda and
http://www.rcpch.ac.uk/child-health/standards-care/clinical-
audit-and-quality-improvement/national-paediatric-diabetes-
audi).
The NDA is managed by the Health and Social Care
Information Centre in partnership with Diabetes UK and
12
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK
DIABETICMedicine HbA
1c
in Type 1 diabetes: an international comparison J. A. McKnight et al.

Public Health England. Data are collected annually from the
Electronic Records of General Practices (88% in 2012) and
specialist services (58% in 2012) including type of diabetes,
age, sex, year of onset and latest HbA
1c
. In 2012, HbA
1c
data
were available for 86% of 213 960 patients with Type 1
diabetes.
The NPDA is managed by the Royal College of Paediatrics
and Child Health (RCPCH), which collects data annually
from 182 paediatric diabetes centres delivering care for
children and young people with diabetes. Data included in
this report are taken from the audit year 1 April 2011 to 31
March 2012 during which time 177 centres participated and
data are available on ~25 000 patients. NPDA data were
used separately for England and Wales for patients aged
<15 years and NDA data for England and Wales combined
were used for other age groups.
France
Data originate from Rheims and the CAR
eDIAB (diabetes
network in Champagne Ardenne) 2011–2012 results for
patients with Type 1 diabetes. The sample represents at least
70% of patients aged <15 years and 30–40% of the older
patients with Type 1 diabetes in the region. Data are
collected at every visit or hospitalization on CAR
eDIAB
shared web medical files. Diagnosis of Type 1 diabetes was
made by endocrinologists (adult and paediatric) at Rheims
hospital on the basis of clinical criteria and antibody status.
Germany
The DPV is a computer-based longitudinal documentation
system for diabetes patients with all types of diabetes and in all
age groups. The electronic health record is documented and
stored locally in each participating treatment centre. Twice a
year, anonymized data are transmitted to the central admin-
istrative unit in Ulm, Germany, where they are aggregated into
a cumulative database. Potentially incorrect data are reported
back to the participating centres for correction or confirma-
tion. The resulting data pool is used for quality management
(benchmarking reports twice yearly) and patient-centred
analyses (for a list of publications, see http://www.d-p-v.eu).
By March 2012, 300 specialized German diabetes centres had
participated in the analysis, 105 treatment centres for adult
diabetes patients and 195 paediatric centres. For patients with
Type 1 diabetes aged <15 years, 94% of all estimated
German patients are part of the registry. The DPV initiative
is funded by the German Federal Ministry of Education and
Research as part of the competence net diabetes mellitus.
Greece
Data were obtained from the regions of Epirus and Thessaly
in Central Greece (total population of 1 066 000). The
sample population was based on patients with Type 1
diabetes and available HbA
1c
data within the previous
24 months (October 2011 to October 2013), who were
actively followed up at the two major university hospitals of
Central Greece (University Hospital of Ioannina, Epirus and
University Hospital of Larisa, Thessaly).
Data were also obtained from the region of Thessaloniki in
North Greece (total population of 1 100 000). The sample
population was extracted from the diabetes centre database
and included patients with Type 1 diabetes and available
HbA
1c
data within the previous 36 months. Diagnosis of
Type 1 diabetes was made on the basis of clinical criteria
and/or the need for insulin treatment within 1 year of the
diagnosis of diabetes.
Italy
Data on mean values of HbA
1c
were extracted for the 12-
month period prior to 30 August 2011 from eight local
registries (whole regions: Valle D’Aosta, Liguria, Marche,
Abruzzo, Calabria, Toscana; provinces: Bari-Foggia, Mes-
sina). The registry for Type 1 diabetes in Italy (RIDI) is a
coordinating centre of registries that prospectively collects
data on newly diagnosed patients under 15 years of age.
Local population-based registries operate at the region or
province level. Electronic clinical records are stored locally.
The register is estimated to be 99% complete for children
and 93% complete for adults.
Latvia
Latvian data come from the population-based Diabetes
Register, which is the part of the Register of Patients with
Particular Diseases. The current owner of this system is the
Centre for Disease Prevention and Control of Latvia. The
Diabetes Register was set up in 1997. Information about
diabetes patients, at an individual level, is provided by family
doctors or endocrinologists; it must be updated at least once
per year. Although the completeness of the register has not
been formally evaluated and, therefore, no results published,
since 2009, in order to ensure completeness, the register data
are regularly compared with electronic records of the
dispensed reimbursed medications (insulin prescriptions are
100% reimbursed) and, consequently, prescribing physicians
are contacted if no matches are found between register and
dispensing data. Mean HbA
1c
from 1 January 2010 to 31
December 2011 was estimated. The case definition of Type 1
diabetes was based on clinical diagnosis (as stated in the
register) plus age at diagnosis <30 years.
The Netherlands
Data were provided by Diabeter, a national centre with a
focus on paediatric and adolescent diabetes. Diabeter has
three locations in the Netherlands, which serve as primary
referral centres for all children and adolescents (aged 0–
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK 13
Research article DIABETICMedicine

18 years) with newly diagnosed Type 1 diabetes in an area
with 1 600 000 inhabitants. Twenty per cent of the patients
submitted in this dataset have received diabetes care at
Diabeter since the onset of their Type 1 diabetes. In addition,
Diabeter is a secondary referral centre for Type 1 diabetes.
Eighty per cent of the patients submitted in this dataset were
diagnosed in another clinic before they were referred to
Diabeter. All adult patients were diagnosed at other clinics.
Type 1 diabetes was diagnosed based on clinical features
plus the presence of autoantibodies. Patients without
autoantibodies, with clinical characteristics such as low
insulin dose or with characteristics to suggest other diagnoses
(e.g. monogenetic diabetes, Type 2) are not included.
New Zealand
Data for 2005 were extracted from the Otago regional
diabetes register, which was established in 1998 to monitor
diabetes care in the Otago region, New Zealand. Individual
data relating to process and outcome measures were
collected annually from 93% of GPs in the region. The most
recent HbA
1c
value for the calendar year was recorded.
Infants and children were under-represented. Adult diabetes
hospital outpatient and retinal screening records were
accessed to check data as necessary, e.g. diabetes type, and
to search for missing data.
Northern Ireland
Clinic-based adult data were obtained from each of the five
Northern Ireland healthcare trusts through audit of their
hospital diabetes databases. This represents ~95% of all
adult patients in Northern Ireland with Type 1 diabetes –
who are predominantly looked after in secondary care. Data
were collated for the year 2012 and diabetes type in the
registries is determined clinically by individual physicians.
Norway
Data were extracted for the 15-month period prior to
December 2011 from the National Diabetes Register, which
includes data from 22 of 40 hospitals where the majority of
adults with Type 1 diabetes receive care so the data are
deemed to be population based. Data collection was started
in 2009 and some hospitals only joined late in 2011. The
estimated coverage of the register for this period is 18%.
Republic of Ireland
Data are from a single centre (Galway University Hospitals).
The hospital provides diabetes care to a catchment area of
~250 000 individuals. Data were extracted from a clinical
information system, DIAMOND (Hicom, Brookwood, UK),
used to capture all outpatient clinical encounters involving
patients with diabetes. Data were extracted on 11 December
2013 and cover a 23-month period. There is no formal
protocol for validation of the diagnosis of Type 1 diabetes.
All new patients would have been formally reviewed by a
consultant endocrinologist at their first visit. During this
visit, the classification of the patient’s diabetes would be
discussed and recorded.
Scotland
Population-based data were obtained from the 2011 extract
of the Scottish Care Information –Diabetes Collaboration
dataset. This register, now containing data for ~99% of
individuals with diagnosed diabetes in Scotland, has
been in place since 2001 and is populated by daily
downloads from primary and secondary databases [18]
and http://www.diabetesinscotland.org.uk/Publications.aspx?
catId=3. A large proportion of individuals aged <15 years
were missing an HbA
1c
measurement so data for this age
group were excluded from the analysis.
Sweden
Population-based data were obtained from the Swedish
National Diabetes Register (NDR). The Swedish NDR was
initiated in 1996 as a tool for quality assurance in diabetes
care. National results are published in yearly reports and the
register is administrated from Centre of Registers in Region
V€
astraG€
otaland, Gothenburg, Sweden (www.NDR.nu)
Reporting to the register is not mandatory, but all hospital
diabetes outpatient clinics and primary healthcare centres are
encouraged to do so. Currently 100% of hospital outpatient
clinics and 95% of primary healthcare centres participate.
More than 346 000 patients were reported to the register in
2012, representing ~85% of adult patients with diabetes in
Sweden. Annual reporting to the NDR is carried out by
trained physicians and nurses via the Internet or via clinical
records databases, with information collected during patient
visits at hospital outpatient clinics and primary healthcare
centres nationwide. All included patients have agreed by
informed consent to register before inclusion. In Sweden,
almost all patients with Type 1 diabetes receive their
treatment at hospital outpatient clinics. We estimate that
~90% of all adult patients with Type 1 diabetes, diagnosed
by clinicians, were reported to the NDR in 2011. However,
for this study, we included patients aged 18 years and older
using an epidemiological definition of Type 1 diabetes with
onset age <30 years and insulin treatment alone, represent-
ing ~70% of all adult Type 1 diabetes patients in the
register. We used the last recorded HbA
1c
value for each
individual between 1 January 2011 and 31 December 2011.
Ukraine
Data were provided from a territorial register of one of
Ukraine’s western regions –Volyn. Type 1 diabetes was
14
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK
DIABETICMedicine HbA
1c
in Type 1 diabetes: an international comparison J. A. McKnight et al.

defined as individuals receiving insulin and who developed
diabetes before the age of 30 years. Several physicians from
provincial subregions of Volyn also include children with
diabetes in their registers. Volyn’s adult insulin-treated
patient data should be considered complete with an esti-
mated coverage of 99% of the population, for whom data
completeness is 83%.
USA
The T1D Exchange Clinic Network includes 70 US-based
paediatric and adult endocrinology clinics. A registry of
individuals with Type 1 diabetes commenced enrolment in
September 2010 [19]. Although the registry data are
collected from a large number of individuals with Type 1
diabetes across the USA, it is not population based. Partic-
ipation in the registry is predicated on being treated by an
endocrinologist. Each clinic received approval from an
institutional review board. Informed consent was obtained
according to institutional review board requirements from
adult participants and parents/guardians of minors; assent
from minors was obtained as required. To be enrolled in the
clinic registry, an individual must have a clinical diagnosis of
presumed autoimmune Type 1 diabetes and either islet cell
antibodies present or if antibodies were negative or
unknown, then insulin must have been started at or shortly
after diagnosis and used continually thereafter (except in the
case of a pancreas or islet cell transplant). The T1D
Exchange differs from some other registries in that the
database includes information that was obtained from a
comprehensive questionnaire that was completed by the
participant or the parent of a child participant, as well as
data that were extracted from the participant’s medical
record, as previously described [19].
ª2014 The Authors.
Diabetic Medicine ª2014 Diabetes UK 15
Research article DIABETICMedicine