Sustained socio-economic inequalities in hospital admissions for cardiovascular

events among people with diabetes in England

Running title: Socioeconomic inequalities in cardiovascular admissions in diabetes

Zainab Shather*1 MPH, Anthony A Laverty*1 MSc PhD, Alex Bottle1 MSc PhD,

Hilary Watt1 CStat MSc, Azeem Majeed MD1, Christopher Millett1 PhD, Eszter P

Vamos1 PhD

*These authors contributed equally to this work

1Public Health Policy Evaluation Unit, School of Public Health, Imperial College

London, London, UK

Word count of main text: 3,163 Abstract: 247.

Tables: 3 Figures: 1

Conflict of interests: AB reports grants from Dr Foster and Medtronic, outside the submitted work. No other authors declared conflict of interest.

Funding: The Public Health Policy Evaluation Unit at Imperial College London is supported by funding from the National Institute for Health Research’s School of Public Health Research programme.

Address for Correspondence:

Eszter P Vamos MD, MPH, FFPH, PhD

Senior Clinical Lecturer in Public Health

& Clinical Consultant in Public Health

Deputy Director of Public Health Policy Evaluation Unit

Public Health Policy Evaluation Unit, School of Public Health

Imperial College London

London W6 8RP, UK

1

Tel: 00 44 (0) 207 594 0817

Fax: 00 44 (0) 207 594 0854

Email: e.vamos@imperial.ac.uk

2

Abstract

Objective

This study aimed to determine changes in absolute and relative socio-economic

inequalities in hospital admissions for major cardiovascular events and procedures

among people with diabetes in England.

Methods

We identified all patients with diagnosed diabetes aged ≥45 years admitted to hospital

in England between 2004-2005 and 2014-2015 for acute myocardial infarction,

stroke, percutaneous coronary intervention (PCI) and coronary artery bypass graft

(CABG). Socio-economic status was measured using Index of Multiple Deprivation.

Diabetes-specific admission rates were calculated for each year by deprivation

quintile. We assessed temporal changes using negative binomial regression models.

Results

Most admissions for cardiovascular causes occurred among people aged ≥65 years

(71%) and men (63.3%), and the number of admissions increased steadily from the

least to the most deprived quintile. People with diabetes in the most deprived quintile

had 1.94-fold increased risk of acute myocardial infarction (95% CI 1.79-2.10), 1.92-

fold risk of stroke (95% CI 1.78-2.07), 1.66-fold risk of CABG (95% CI 1.50-1.74),

and 1.76-fold risk of PCI (95% CI 1.64-1.89) compared with the least deprived group.

Absolute differences in rates between the least and most deprived quintiles did not

significantly change for acute myocardial infarction (P=0.29) and were reduced for

stroke, CABG and PCI (by 17.5, 15 and 11.8 per 100,000 people with diabetes,

respectively, P≤0.01 for all).

3

Conclusions

Socio-economic inequalities persist in diabetes-related hospital admissions for major

cardiovascular events in England. Besides improved risk stratification strategies

considering socio-economically defined needs, wide-reaching population-based

policy interventions are required to reduce inequalities in diabetes outcomes.

4

Introduction

Cardiovascular disease remains a leading cause of morbidity and premature mortality

and is a substantial contributor to health inequalities globally1. The risk of morbidity

and death, particularly cardiovascular mortality, shows marked variations according

to socio-economic status, as measured by income, education, social class and area-

based deprivation indices2. Lower socio-economic status is also a powerful predictor

of higher incidence of Type 2 diabetes as well as its acute and long-term

complications3. This mirrors the socio-economic patterning of risk factors such as

poor diet, lack of exercise and smoking, psychological stress and access to and use of

evidence-based preventive and therapeutic interventions3,4. The fast-growing

prevalence of Type 2 diabetes with a disproportionate burden on disadvantaged

populations represents a serious concern for health systems and societies.

Epidemiological studies show that despite targeted interventions, inequalities in

coronary heart disease mortality in the general population have not only persisted but

widened since the 1970s in England and other developed countries2,5-7. Although

absolute inequalities in mortality narrowed over time as cardiovascular mortality has

fallen in all socio-economic groups, faster declines in more affluent groups have led

to an increase in relative socioeconomic inequalities in many countries5,7. Worryingly,

there is increasing evidence that health inequality gaps have been widening since

2010 in England alongside reductions in public service funding8.

While extensively studied in the general population, little is known about how people

with diabetes from different socioeconomic groups have benefited from reductions in

cardiovascular disease over the last decade. Few longitudinal studies have examined

5

the temporal relations between measures of socio-economic status and cardiovascular

outcomes in diabetes internationally3,9-18. Some, but not all, studies reported widening

socio-economic inequalities in people with diabetes9-11,19. Previous studies have been

limited by small sample sizes, short follow-up times or were conducted a long time

ago and most of them focused on cardiovascular mortality rather than morbidity.

The objective of this nationwide study was to describe hospital admissions for acute

myocardial infarction, stroke, percutaneous coronary intervention (PCI) and coronary

artery bypass graft (CABG) in people with diabetes by socio-economic deprivation

between 2004-2005 and 2014-2015 in England. We also assessed whether the

absolute and relative socio-economic gradient in study outcomes have changed among

people with diabetes during this 11-year period.

6

Materials and Methods

Hospital Episode Statistic (HES) is a national administrative database managed by the

NHS Digital that contain records of all in-patient and day-case admissions to NHS

hospitals across England, including private patients treated in NHS facilities20.

England offers universal health coverage with care free at the point of delivery. HES

is an administrative dataset used primarily for financial reimbursements of NHS

hospitals for the care they provide based on coding data, and reporting is mandatory.

We identified all people with diagnosed diabetes aged 45 years and above, who were

admitted to hospital between the financial years 2004-2005 and 2014-2015 in England

for acute myocardial infarction, stroke, CABG and PCI. Data extracted for each

admission included age, gender, registered family practice, principal diagnosis and up

to 19 secondary diagnoses on admission using the ICD-10 (International

Classification of Diseases) classification, up to 12 procedures and in-patient deaths.

Interventions were identified using the Office of Population Censuses and Surveys 4

(OPCS4) coding.

We included patients with both Type 1 diabetes (ICD-10 E10) and Type 2 diabetes

(ICD-10 E11), recorded in any diagnostic field. Cardiovascular admissions, identified

as principal diagnosis on admissions, included acute myocardial infarction (ICD-10

I21 and I22) and stroke (ICD-10 I60-I64). Revascularisation procedures were

identified using procedure codes in any field for PCI (OPCS4 K49-K50, K79) or

CABG (OPCS4 K40-K46).

7

We used the Index of Multiple Deprivation 2010 as a measure of socio-economic

status. The Index of Multiple Deprivation is the most commonly used method of

measuring deprivation in England at the small area level (Lower-layer Super Output

Area, an area covering 1,000 to 3,000 people or 400 to 1,200 households)21. We used

deprivation quintiles defined for the whole population of England (i.e. population of

England grouped into equal fifths according to deprivation scores). We assigned each

patient a deprivation quintile (1-least deprived, 5-most deprived) based on their family

practice postcode22.

Diabetes-specific cardiovascular admission rates were calculated by gender, four age-

bands (45-54, 55-64, 65-74 and >=75 years) and deprivation quintiles using the

number of admissions as numerator and the number of people with diabetes in

particular age, sex and deprivation bands as denominator. Data on the total number of

people with diagnosed diabetes in England for each study year were obtained from the

Quality and Outcomes Framework (QOF). QOF is a national primary care pay-for-

performance scheme introduced in the UK in 2004 that holds data on number of

diabetes patients for over 99% of family practices in England 23. As the age, gender

and deprivation distribution of people with diabetes are not available from QOF, we

used Health Survey for England (HSE) data between 2006 and 2014 and estimated the

number of people with diabetes within each age, sex and deprivation stratum24. For

those study years when HSE surveys did not include data on diabetes prevalence (in

2004, 2005, 2007 and 2008), we used 2006 HSE data similar to previous studies25,26.

Admission rates were directly standardized by age and gender using the first study

year’s population structure as reference. All event rates were expressed per 100,000

people with diabetes. In-patient case-fatality rates were calculated using the number

8

of in-patients deaths as numerator and the number of admissions by age, gender, and

deprivation quintile as the denominator.

Statistical analyses

Group differences were tested using Pearson’s Chi-square test for categorical

variables. For graphical presentation of age- and sex-standardized admission rates, we

used locally weighted scatterplot smoothing for each study outcome by deprivation

quintile over study years. We used two sets of analyses to estimate changes in

admission rates. First, we fitted linear regression models to estimate how differences

in absolute admission rates between the least and most deprived groups changed

during the study period. Second, we fitted negative binomial regression models for

each study outcome to estimate rate ratios for the deprivation groups using the least

deprived group as reference. We used age, gender, deprivation quintile, study year

and a quadratic term for study year as independent variables. We then tested for an

interaction between year and deprivation to assess changes in relative inequalities

over time. We also fitted negative binomial regression models to obtain rate ratios for

in-patient case-fatality. All statistical analyses were conducted using Stata version

14.0.

9

Results

The baseline characteristics of people with diabetes over the age of 45 years admitted

to hospital for acute myocardial infarction, stroke, CABG and PCI in England

between 2004-2005 and 2014-2015 by socio-economic deprivation quintiles are

shown in Table 1. The number of admissions steadily increased across deprivation

groups from the least to the most deprived quintile for all outcomes. In 2004-2005,

approximately one-quarter of admissions occurred among patients in the most

deprived quintile, while 15% of admissions were from the most affluent groups for all

four outcomes (Table 1). We found a similar pattern when examining the proportion

of admissions further broken down by age and sex. The majority of admissions for all

cardiovascular causes occurred among people aged ≥65 years (71%) and men (63.3%)

across all quintiles. However, with increasing levels of deprivation, significantly

larger proportions of younger patients were represented among those affected by

cardiovascular disease (Table 1).

Absolute changes in admissions rates

Figure 1 shows locally weighted scatterplot smoothing curves for the age- and gender-

standardized admission rates for study outcomes between 2004-2005 and 2014-2015

in people with diabetes in England. A consistent inverse socio-economic patterning

was evident for all outcomes for all study years with increasing admission rates along

the deprivation quintiles from the least (quintile 1) to most deprived groups (quintile

5).

Admission rates for acute myocardial infarction decreased in all deprivation quintiles

between 2004-2005 and 2009-2010, followed by the flattening out of rates until the

10

end of the study period in all groups except for quintile 5 (Figure 1). In quintile 5

(most deprived group), admission rates for acute myocardial infarction increased

between 2010-11 and 2014-2015 from 147.8 to 163.3 per 100,000 people with

diabetes. Admission rates for stroke rose between 2004-2005 and 2014-2015 in all

socio-economic groups, except for the most deprived group where admission rates

gradually declined during the study period. While CABG rates declined in all

socioeconomic groups, there was a consistent parallel increase in admission rates for

PCI.

In linear regression models, between 2004-2005 and 2014-2015, the absolute

differences in admission rates between the least (quintile 1) and most deprived groups

(quintile 5) remained unchanged for acute myocardial infarction (P=0.29) and were

reduced significantly by 17.5 per 100,000 people with diabetes for stroke (P=0.002),

by 15 per 100,000 people with diabetes for CABG (P<0.001), and by 11.8 per

100,000 people for PCI (P=0.03).

Relative differences in admission rates

Between 2004-2005 and 2014-2015, admission rate ratios for acute myocardial

infarction, stroke, CABG and PCI showed a statistically significant socioeconomic

gradient, with admission rate ratios steadily increasing with increasing levels of

deprivation (Table 2, Model 1). People with diabetes in the most deprived quintile

(quintile 5) were approximately twice as likely to be admitted to hospital for

cardiovascular as those in the least deprived quintile (quintile 1), with rate ratios of

1.94 for acute myocardial infarction, 1.92 for stroke, 1.66 for CABG, and 1.76 for

PCI admissions (P< 0.001 for all) (Table 2).

11

The interaction term between year and deprivation quintiles did not reach statistical

significance for any of the outcomes. This reflects that deprived groups experienced

similar proportional changes in study outcomes as the most affluent group over time

(Table 2, Model 2).

In-patient case-fatality

From 2004-2005 to 2014-2015, there was a reduction in in-patient case-fatality rates

for all outcomes except for PCI (Table 1). In negative binomial regression models, in-

patient case-fatality rates for acute myocardial infarction, CABG and PCI were higher

in the most deprived group (quintile 5) compared with quintile 1 (most affluent)

(Table 3, Model 1). However, only associations for PCI reached statistical

significance with the most deprived group experiencing 24% higher in-patient case-

fatality compared with the least deprived group. By contrast, for stroke admissions,

the most deprived group had significantly lower case-fatality rates compared with the

most affluent group. Trends in in-patient case-fatality did not statistically significantly

differ between the most affluent and other deprivation groups (Table 3, Model 2).

12

Discussion

In England between 2004-2005 and 2014-2015, diabetes-related hospital admission

rates for major cardiovascular causes showed a marked socio-economic gradient with

consistently higher admission rates with increasing levels of deprivation. Absolute

inequalities in admission rates remained unchanged for acute myocardial infarction

and were decreased for stroke, PCI and CABG during the study period. However,

people in different deprivation groups experienced similar proportional changes in

outcomes, leaving relative inequalities among people with diabetes unchanged for the

study period. People with diabetes in the most deprived group had 1.9-fold increased

admission rates for acute myocardial infarction and stroke, 1.8-fold increased

admission rates for PCI and stroke and 1.7-fold increased CABG admission rates

when compared with the least deprived areas. In-patient case-fatality was

approximately 6% lower for stroke and 24% higher for PCI in the most deprived

group compared with the most affluent.

Cardiovascular and related mortality have fallen substantially during the past few

decades in England and other Western countries25,27-29. Epidemiological studies have

shown that up to 75% of reductions in coronary heart disease mortality can be

explained by population-level improvements in major risk factors, predominantly

smoking, blood pressure and cholesterol, while 25-50% can be attributed to

improvements in the quality and accessibility of health services6,27,28. These advances

are, however, disproportionately spread across socio-economic groups and the slower

pace of decline in coronary heart disease mortality in the most deprived areas are

likely to mirror corresponding patterns of key risk factors30. Furthermore, adverse

trends in physical activity, obesity and Type 2 diabetes, all of which are more likely to

13

affect the socio-economically disadvantaged, offset some of the favorable changes

seen in cardiovascular outcomes30.

While widely investigated in the general population, among people with diabetes data

on recent changes in socio-economic inequalities in major cardiovascular events are

scarce internationally. Most previous studies focused on the association between

socio-economic status and all-cause or coronary heart disease mortality and few have

investigated specific cardiovascular events such as stroke19,31. This is an important

knowledge gap as people with diabetes increasingly survive cardiovascular events and

mortality itself does not provide accurate reflection on its burden15. A study from the

US reported that educational disparities in self-reported history of cardiovascular

remained uniform between 1997 and 2005 among people with diabetes, whilst

inequalities widened amongst people without diabetes9. Previous cohort studies

demonstrated a 1.5- to 3-fold increased risk of fatal and non-fatal cardiovascular

disease among lower socio-economic groups compared with more privileged groups

with Type 1 and Type 2 diabetes16,17,32.

We found that, unexpectedly, the most deprived group had lower in-patient case-

fatality related to stroke admissions compared with the most affluent group. This

finding was not consistent with findings for other study outcomes. Patients admitted

for stroke living in most deprived areas were younger compared with more affluent

groups and although our analyses were adjusted for age, they may have differed in

other risk factors that could explain this finding. It is also possible that most deprived

groups are over-represented among those with fatal stroke who do not reach hospital

and therefore, fewer patients at high risk of mortality contribute to in-patient case-

fatality rates.

14

Aggressive management of major cardiovascular risk factors aligned with socio-

economically defined needs are likely to reduce inequalities in cardiovascular

outcomes. Risk stratification to identify patients with diabetes for more intensive

cardiovascular prevention has assumed increasing importance recently33. Because

coronary heart disease risk in Type 2 diabetes is not similar to the risk of patients with

a previous cardiovascular event in all instances, diabetes is no longer considered a

coronary heart disease risk equivalent by many scientific groups34. Given the highly

heterogeneous risk profile of patients with diabetes, risk stratification strategies

including cardiovascular risk score estimations bear increasing importance35. By

contrast to some predictive models used in the general population, these strategies do

not include measures of socio-economic status among people with diabetes35. Whether

inclusion of deprivation improves cardiovascular risk estimations and risk

stratification for people with diabetes warrants further research.

England has a long history of concerted efforts to assess, understand and reduce

health inequalities36. The NHS provides free care at the point of use and substantial

investments have been made during the past decades to improve the quality of chronic

disease management and reduce variations in care23. Besides these efforts, important

population-wide measures were introduced to improve risk factors across the

population such as smoke-free legislation, other tobacco control policies and

measures to reduce other cardiovascular risk factors (poor diet, hypertension and

physical activity)36,37. While interventions targeting individuals at high risk of

cardiovascular disease are essential, generally, such interventions may benefit

population subgroups differently due to differences in access, uptake and compliance,

potentially resulting in widening socio-economic inequalities38. Therefore, these need

15

to be complemented by wide-reaching population-level strategies to reduce risk

factors across the whole population such as legislative and fiscal policies38.

Strengths and limitations

To our knowledge, this is the first nationwide study describing recent trends of

hospital admissions for selected cardiovascular causes in people with diabetes by

socio-economic status in England. Our data cover the entire population of England

treated in NHS hospitals and given that our study outcomes require hospital

admissions, our data are likely to provide an accurate reflection on diabetes-related

cardiovascular event rates. The diabetes denominator data was obtained from national

diabetes registers with the participation of over 99% of family practices.

This study does have limitations. We used routinely collected data and we were

unable to directly assess misclassification and miscoding. However, HES is an

administrative dataset and the reimbursement of NHS hospitals is directly determined

by coded data, and therefore, there is a strong financial incentive for providers to

accurately code clinical information. Routinely collected clinical data are nationally

audited regularly and a systematic review has evaluated its accuracy as high for both

diagnoses and procedures39. Our study only included cardiovascular outcomes that

resulted in hospital admissions and does not capture fatal cardiovascular cases, such

as those who died before reaching hospital. Furthermore, we could not stratify our

analyses by type of diabetes because national prevalence data according to types were

not available for the study years. We assigned an Index of Multiple Deprivation score

to individual patients based on their family practice postcode as individual-level

measures of socio-economic status (e.g. education, income, etc.) are not routinely

collected in UK family practice. Postcode-based deprivation scores are available for

16

each patient and are generally accepted as good proxies for individual-level

deprivation40,41. We used data from Health Survey for England to ascertain the age,

gender and deprivation distribution of diabetes in England. For 2004, 2005 and 2007

and 2008 when age-, gender- and deprivation-specific data were not available we used

data from 2006, the mid-term of this 5-year period25,26.

Conclusion

This nationwide study covering the entire population in England, where universal

health coverage provides free care for all, showed a persisting socio-economic

gradient in hospital admissions for major cardiovascular causes among people with

diabetes during the past decade. These findings highlight that social factors are still

not adequately addressed in health policy. Besides risk stratification strategies that

consider the socio-economic needs of individuals with diabetes, wide-reaching

population-based policy interventions are required that alter risk factors across the

population to reduce inequalities in diabetes outcomes.

Contributions

ZS researched the data and wrote the manuscript. AL researched the data and wrote

the manuscript. AB researched the data and wrote the manuscript. HW researched the

data and wrote the manuscript. AM reviewed/edited the manuscript and revised it for

important intellectual content. CM wrote the manuscript and revised it for important

intellectual content. EV conceived the idea, researched the data and wrote the

manuscript.

17

References

1. Roth, GA, Johnson, C, Abajobir, A, et al. Global, Regional, and National

Burden of Cardiovascular Diseases for 10 Causes, 1990 to 2015. J Am Coll

Cardiol 2017; 70(1): 1-25.

2. Jemal, A, Ward, E, Anderson, RN, Murray, T, Thun, MJ. Widening of

socioeconomic inequalities in U.S. death rates, 1993-2001. PLoS One 2008; 3(5):

e2181.

3. Agardh, E, Allebeck, P, Hallqvist, J, Moradi, T, Sidorchuk, A. Type 2

diabetes incidence and socio-economic position: a systematic review and meta-

analysis. Int J Epidemiol 2011; 40(3): 804-18.

4. Wild, S, Macleod, F, McKnight, J, et al. Impact of deprivation on

cardiovascular risk factors in people with diabetes: an observational study.

Diabet Med 2008; 25(2): 194-9.

5. Bajekal, M, Scholes, S, O'Flaherty, M, Raine, R, Norman, P, Capewell, S.

Unequal trends in coronary heart disease mortality by socioeconomic

circumstances, England 1982-2006: an analytical study. PLoS One 2013; 8(3):

e59608.

6. Bajekal, M, Scholes, S, Love, H, et al. Analysing recent socioeconomic

trends in coronary heart disease mortality in England, 2000-2007: a population

modelling study. PLoS Med 2012; 9(6): e1001237.

7. Mackenbach, JP, Kulhanova, I, Artnik, B, et al. Changes in mortality

inequalities over two decades: register based study of European countries. BMJ

2016; 353: i1732.

8. Department of Health: Annual report and accounts 2016 to 2017.

Available at https://www.gov.uk/government/publications/department-of-health-

annual-report-and-accounts-2016-to-2017 (Last accessed: 4 June 2018).

18

9. Dray-Spira, R, Gary, TL, Brancati, FL. Socioeconomic position and

cardiovascular disease in adults with and without diabetes: United States trends,

1997-2005. J Gen Intern Med 2008; 23(10): 1634-41.

10. Koskinen, SV, Martelin, TP, Valkonen, T. Socioeconomic differences in

mortality among diabetic people in Finland: five year follow up. BMJ 1996;

313(7063): 975-8.

11. Forssas, E, Keskimaki, I, Reunanen, A, Koskinen, S. Widening

socioeconomic mortality disparity among diabetic people in Finland. Eur J Public

Health 2003; 13(1): 38-43.

12. Jackson, CA, Jones, NR, Walker, JJ, et al. Area-based socioeconomic status,

type 2 diabetes and cardiovascular mortality in Scotland. Diabetologia 2012;

55(11): 2938-45.

13. Roper, NA, Bilous, RW, Kelly, WF, Unwin, NC, Connolly, VM. Excess

mortality in a population with diabetes and the impact of material deprivation:

longitudinal, population based study. BMJ 2001; 322(7299): 1389-93.

14. Chaturvedi, N, Jarrett, J, Shipley, MJ, Fuller, JH. Socioeconomic gradient in

morbidity and mortality in people with diabetes: cohort study findings from the

Whitehall Study and the WHO Multinational Study of Vascular Disease in

Diabetes. BMJ 1998; 316(7125): 100-5.

15. Newton, JN, Briggs, AD, Murray, CJ, et al. Changes in health in England,

with analysis by English regions and areas of deprivation, 1990-2013: a

systematic analysis for the Global Burden of Disease Study 2013. Lancet 2015;

386(10010): 2257-74.

16. Secrest, AM, Costacou, T, Gutelius, B, Miller, RG, Songer, TJ, Orchard, TJ.

Associations between socioeconomic status and major complications in type 1

19

diabetes: the Pittsburgh epidemiology of diabetes complication (EDC) Study.

Ann Epidemiol 2011; 21(5): 374-81.

17. Rawshani, A, Svensson, A, Rosengren, A, Eliasson, B, Gudbjornsdottir, S.

Impact of socioeconomic status on cardiovascular disease and mortality in 24,947

individuals with type 1 diabetes. Diabetes Care 2015; 38: 1518 - 27.

18. Gnavi, R, Canova, C, Picariello, R, et al. Mortality, incidence of

cardiovascular diseases, and educational level among the diabetic and non-

diabetic populations in two large Italian cities. Diabetes Res Clin Pract 2011;

92(2): 205-12.

19. Lipscombe, LL, Austin, PC, Manuel, DG, Shah, BR, Hux, JE, Booth, GL.

Income-related differences in mortality among people with diabetes mellitus.

CMAJ 2010; 182(1): E1-E17.

20. National Health Service Digital: Hospital Episode Statistics. Available at:

http://content.digital.nhs.uk/hes (Last accessed: 4 June 2018).

21. Department for Communities and Local Government: English indices of

deprivation 2015. Available at:

https://www.gov.uk/government/statistics/english-indices-of-deprivation-2015

(Last accessed: 4 June 2018).

22. Strong, M, Maheswaran, R, Pearson, T. A comparison of methods for

calculating general practice level socioeconomic deprivation. Int J Health Geogr

2006; 5: 29.

23. National Health Service Digital: Quality and Outcomes Framework.

Available at http://content.digital.nhs.uk/qof (Last accessed: 4 June 2018).

24. UK Government Digital Service: Health Survey for England. Available at

https://data.gov.uk/dataset/health_survey_for_england. (Last accessed: 4 June

2018).

20

25. Vamos, EP, Millett, C, Parsons, C, Aylin, P, Majeed, A, Bottle, A.

Nationwide study on trends in hospital admissions for major cardiovascular

events and procedures among people with and without diabetes in England,

2004-2009. Diabetes Care 2012; 35(2): 265-72.

26. Laverty, AA, Bottle, A, Kim, SH, et al. Gender differences in hospital

admissions for major cardiovascular events and procedures in people with and

without diabetes in England: a nationwide study 2004-2014. Cardiovasc Diabetol

2017; 16(1): 100.

27. Unal, B, Critchley, JA, Capewell, S. Explaining the decline in coronary

heart disease mortality in England and Wales between 1981 and 2000.

Circulation 2004; 109(9): 1101-7.

28. Ford, ES, Ajani, UA, Croft, JB, et al. Explaining the decrease in U.S. deaths

from coronary disease, 1980-2000. N Engl J Med 2007; 356(23): 2388-98.

29. Booth, GL, Kapral, MK, Fung, K, Tu, JV. Recent trends in cardiovascular

complications among men and women with and without diabetes. Diabetes Care

2006; 29(1): 32-7.

30. Scholes, S, Bajekal, M, Love, H, et al. Persistent socioeconomic inequalities

in cardiovascular risk factors in England over 1994-2008: a time-trend analysis

of repeated cross-sectional data. BMC Public Health 2012; 12: 129.

31. Saydah, S, Imperatore, G, Beckles, GL. Socioeconomic status and mortality

- contribution of health care access and psychological distress among US adults

with diagnosed diabetes. Diabetes Care 2013; 36: 49-55.

32. Rawshani, A, Svensson, AM, Zethelius, B, Eliasson, B, Rosengren, A,

Gudbjornsdottir, S. Association Between Socioeconomic Status and Mortality,

Cardiovascular Disease, and Cancer in Patients With Type 2 Diabetes. JAMA

Intern Med 2016; 176(8): 1146-54.

21

33. Bertoluci, MC, Rocha, VZ. Cardiovascular risk assessment in patients with

diabetes. Diabetol Metab Syndr 2017; 9: 25.

34. American Diabetes Association: Standards of Medical Care in Diabetes

2017. Diabetes Care 2017; 40: S1 - S135.

35. Allan, GM, Nouri, F, Korownyk, C, Kolber, MR, Vandermeer, B,

McCormack, J. Agreement among cardiovascular disease risk calculators.

Circulation 2013; 127(19): 1948-56.

36. Mackenbach, JP. Can we reduce health inequalities? An analysis of the

English strategy (1997-2010). J Epidemiol Community Health 2011; 65(7): 568-75.

37. Department of Health: Reducing Health Inequalities: An Action Report.

Department of Health, 1999. Available at:

http://webarchive.nationalarchives.gov.uk/+/http://www.dh.gov.uk/en/

Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/

DH_4006054 (Last accessed: 4 June 2018).

38. Capewell, S, Graham, H. Will cardiovascular disease prevention widen

health inequalities? PLoS Med 2010; 7(8): e1000320.

39. Burns, EM, Rigby, E, Mamidanna, R, et al. Systematic review of discharge

coding accuracy. J Public Health (Oxf) 2012; 34(1): 138-48.

40. Robinson, WS. Ecological correlations and the behavior of individuals. Int

J Epidemiol 2009; 38(2): 337-41.

41. Ashworth, M, Lloyd, D, Smith, RS, Wagner, A, Rowlands, G. Social

deprivation and statin prescribing: a cross-sectional analysis using data from the

new UK general practitioner 'Quality and Outcomes Framework'. J Public

Health (Oxf) 2007; 29(1): 40-7.

22

Table 1. Characteristics of people with diabetes admitted to hospital for major cardiovascular events and procedures between 2004-2005 and 2014-2015 in England

Outcome Index of Multiple Deprivation (IMD) quintiles2004-2005

Index of Multiple Deprivation (IMD) quintiles2014-2015

Q1 Q2 Q3 Q4 Q5 P Q1 Q2 Q3 Q4 Q5 PAcute Myocardial InfarctionNumber of events (%) 1,865 (15.3) 2,205 (18.1) 2,499 (20.5) 2,740 (22.5) 2,861 (23.5) 2,556 (15.6) 2,989 (18.3) 3,265 (19.9) 3,509 (21.5) 4,039 (24.7)Age groups, years, N (%) 45-54 116 (6.2) 115 (5.2) 168 (6.7) 224 (8.2) 308 (10.8) 184 (7.2) 235 (7.9) 297 (9.1) 398 (11.3) 526 (13.0) 55-64 268 (14.4) 333 (15.1) 418 (16.7) 458 (16.7) 527 (18.4) 352 (13.8) 487 (16.3) 588 (18.0) 708 (20.2) 941 (23.3) 65-74 538 (28.8) 645 (29.2) 717 (28.7) 837 (30.5) 952 (33.3) 678 (26.5) 796 (26.6) 837 (25.6) 917 (26.1) 1,009 (25.0) ≥75 943 (50.6) 1,112 (50.4) 1,196 (47.9) 1,221 (44.6) 1,074 (37.5) <0.001* 1,342 (52.5) 1,471 (49.2) 1,543 (47.3) 1,486 (42.3) 1,563 (38.7) <0.001*Male, N (%) 1,165 (62.5) 1,372 (62.2) 1,526 (61.1) 1,613 (58.9) 1,687 (59.0) 0.02* 1,628 (63.7) 1,967 (65.8) 2,113 (64.7) 2,203 (62.8) 2,555 (63.3) 0.08*In-patient deaths, N (%) 289 (15.5) 349 (15.8) 387 (15.5) 424 (15.5) 449 (15.7) 0.99* 280 (10.9) 294 (9.8) 338 (10.3) 324 (9.2) 386 (9.6) 0.18*StrokeNumber of events (%) 1,472 (15.5) 1,781 (18.8) 1,849 (19.5) 2,071 (21.8) 2,311 (24.4) 2,774 (16.4) 3,317 (19.6) 3,510 (20.7) 3,624 (21.4) 3,714 (21.9)Age groups, years, N (%) 45-54 56 (3.8) 68 (3.8) 75 (4.1) 123 (5.9) 174 (7.5) 113 (4.1) 140 (4.2) 171 (4.9) 251 (6.9) 358 (9.6) 55-64 144 (9.8) 176 (9.9) 198 (10.7) 299 (14.4) 355 (15.4) 317 (11.4) 351 (10.6) 436 (12.4) 495 (13.7) 607 (16.3) 65-74 357 (24.2) 449 (25.2) 490 (26.5) 627 (30.3) 742 (32.1) 595 (21.4) 782 (23.6) 800 (22.8) 896 (24.7) 1,008 (27.1) ≥75 915 (62.2) 1,088 (61.1) 1,086 (58.7) 1,022 (49.4) 1,040 (45.0) <0.001* 1,749 (63.1) 2,044 (61.6) 2,103 (59.9) 1,982 (54.7) 1,741(46.9) <0.001*Male, N (%) 757 (50.6) 939 (52.7) 931 (50.3) 1,056 (51.0) 1,165 (50.4) 0.58* 1,611 (58.1) 1,825 (55.0) 1,927 (54.9) 1,984 (54.7) 2,006 (54.0) 0.062*In-patient deaths, N (%) 424 (28.8) 533 (29.9) 559 (30.2) 562 (27.1) 581 (25.1) 0.001* 500 (18.0) 573 (17.3) 601 (17.1) 599 (16.5) 547 (14.7) 0.004*CABGNumber of events (%) 696 (15.6) 810 (18.1) 886 (19.8) 952 (21.3) 1,119 (25.1) 861 (16.3) 1,035 (19.6) 1,036 (19.6) 1,176 (22.2) 1,177 (22.3)Age groups, years, N (%) 45-54 54 (7.8) 63 (7.8) 76 (8.6) 108 (11.3) 160 (14.3) 65 (7.5) 78 (7.5) 92 (8.9) 142 (12.1) 157 (13.3) 55-64 191 (27.4) 230 (28.4) 258 (29.1) 299 (31.4) 338 (30.2) 183 (21.2) 261 (25.2) 281 (27.1) 330 (28.1) 372 (31.6) 65-74 303 (43.5) 355 (43.8) 403 (45.5) 417 (43.8) 467 (41.7) 369 (42.9) 390 (37.7) 392 (37.8) 437 (37.2) 407 (34.6) ≥75 148 (21.3) 162 (20.0) 149 (16.8) 128 (13.4) 154 (13.8) <0.001* 244 (28.3) 306 (29.6) 271 (26.2) 267 (22.7) 241 (20.5) <0.001*Male, N (%) 545 (78.3) 639 (78.9) 687 (77.5) 722 (75.8) 818 (73.1) 0.02* 696 (80.8) 835 (80.7) 792 (76.4) 918 (78.1) 910 (77.3) 0.05*

23

In-patient deaths, N (%) 22 (3.1) 40 (4.9) 34 (3.8) 32 (3.4) 23 (2.1) 0.01* 13 (1.5) 16 (1.5) 27 (2.6) 21 (1.8) 18 (1.5) 0.28*

PCINumber of events (%) 1,046 (14.9) 1,281 (18.3) 1,305 (18.6) 1,526 (21.8) 1,841 (26.3) 2,668 (16.9) 2,974 (18.8) 3,248 (20.5) 3,242 (20.5) 3,682 (23.3)Age groups, years, N (%) 45-54 115 (11.0) 152 (11.9) 162 (12.4) 246 (16.1) 370 (20.1) 277 (10.4) 345 (11.6) 400 (12.3) 523 (16.1) 651 (17.7) 55-64 285 (27.2) 420 (32.8) 412 (31.6) 522 (34.2) 542 (29.4) 591 (22.1) 705 (23.7) 889 (27.4) 964 (29.7) 1,089 (29.6) 65-74 437 (41.8) 490 (38.2) 509 (39.0) 523 (34.3) 656 (35.6) 497(35.5) 1,011 (34.0) 1,058 (32.6) 982 (30.3) 1,034 (28.1) ≥75 209 (20.0) 219 (17.1) 222 (17.0) 235 (15.4) 273 (14.8) <0.001* 853 (32.0) 913 (30.7) 901 (27.7) 773 (23.8) 908 (24.7) <0.001*Male, N (%) 736 (70.4) 925 (72.2) 919 (70.4) 1,029 (67.4) 1,174 (63.7) <0.001* 1,977 (74.1) 2,160 (72.6) 2,378 (73.2) 2,308 (71.2) 2,512 (68.2) <0.001*In-patient deaths, N (%) 19 (1.8) 10 (0.8) 25(1.9) 18 (1.2) 31 (1.7) 0.08* 76 (2.8) 54 (1.8) 75(2.3) 73 (2.2) 112 (3.0) 0.01*

Note: CABG: Coronary artery bypass grafting; PCI: Percutaneous coronary intervention.*Chi-squared test

24

Table 2. Risk of admissions for cardiovascular disease causes amongst people with diabetes over the age of 45 years in England between 2004-2005 and 2014-2015 by Index of Multiple Deprivation 2010 (IMD) quintiles

Model 1* Model 2**

Outcome Quintile of IMD (Q)

IRR 95% CIIRR for interaction between year & IMD

95% CI

Acute Myocardial Infarction

Q1 least deprived Reference ReferenceQ2 1.19 1.10 – 1.29 1.01 0.99 – 1.04Q3 1.41 1.30 – 1.52 1.01 0.98 - 1.04Q4 1.64 1.52 – 1.78 1.01 0.98 – 1.03Q5 most deprived 1.94 1.79 – 2.10 1.01 0.99 – 1.04

StrokeQ1 least deprived Reference ReferenceQ2 1.20 1.12 – 1.30 1.00 0.98 – 1.02Q3 1.36 1.26 – 1.46 1.00 0.98 – 1.03Q4 1.60 1.49 – 1.73 0.99 0.97 – 1.02Q5 most deprived 1.92 1.78 – 2.07 0.99 0.96 – 1.01

CABGQ1 least deprived Reference ReferenceQ2 1.17 1.09 – 1.26 1.00 0.98 – 1.02Q3 1.25 1.16 – 1.34 1.00 0.98 – 1.02Q4 1.38 1.29 – 1.49 0.99 0.97 – 1.02Q5 most deprived 1.66 1.50 – 1.74 0.98 0.96 – 1.01

PCIQ1 least deprived Reference ReferenceQ2 1.15 1.07 – 1.23 1.00 0.98 – 1.03Q3 1.24 1.15 – 1.33 1.01 0.98 – 1.03Q4 1.46 1.36 – 1.57 0.99 0.96 – 1.01Q5 most deprived 1.76 1.64 – 1.89 0.98 0.96 – 1.00

Note: CABG: Coronary artery bypass grafting; PCI: Percutaneous coronary intervention; IRR: Incidence Rate Ratio; CI: Confidence Interval

* Model 1: Negative binomial regression models adjusted for age, sex, study year, and quadratic term for year (year squared)

**Model 2: Negative binomial regression models adjusted for age, sex, study year, and quadratic term for year (year squared) and an interaction between year and index of multiple deprivation quintiles

25

Table 3. Rate ratios for in-patient case-fatality in people with diabetes over the age of 45 years in England between 2004-2005 and 2014-2015 among deprivation quintile groups relative to the least deprived quintile

Model 1* Model 2 **

Outcome Quintile of IMD (Q)

IRR for IMD

95% CIIRR for interaction between year & IMD

95% CI

Acute Myocardial Infarction

Q1 least deprived Reference ReferenceQ2 0.99 0.95 – 1.05 0.99 0.91 – 1.09Q3 1.01 0.96 – 1.05 0.99 0.91 – 1.08Q4 0.99 0.94 – 1.04 0.98 0.90 – 1.07Q5 most deprived 1.04 0.99 – 1.09 1.01 0.93 – 1.11

STROKEQ1 least deprived Reference ReferenceQ2 1.00 0.96 – 1.04 1.00 0.99 – 1.01Q3 1.00 0.96 – 1.04 1.00 0.99 – 1.02Q4 0.97 0.93 – 1.00 0.99 0.99 – 1.01Q5 most deprived 0.94 0.91 – 0.98 1.00 0.99 – 1.02

CABGQ1 least deprived Reference ReferenceQ2 1.07 0.90 – 1.26 0.96 0.90 – 1.01Q3 1.14 0.97 – 1.35 1.02 0.97 – 1.08Q4 1.22 1.04 – 1.43 1.01 0.96 – 1.07Q5 most deprived 1.16 0.99 – 1.37 0.99 0.93 – 1.04

PCIQ1 least deprived Reference ReferenceQ2 0.95 0.83 – 1.08 1.00 0.95 – 1.05Q3 1.07 0.94 – 1.21 0.97 0.93 – 1.02Q4 1.03 0.91 – 1.17 0.99 0.95 – 1.03Q5 most deprived 1.24 1.10 – 1.39 0.99 0.96 – 1.04

Note: CABG: Coronary artery bypass grafting; PCI: Percutaneous coronary intervention, IRR: Incidence Rate Ratio; CI: Confidence Interval

* Model 1: Negative binomial regression models adjusted for age, sex, study year, and quadratic term for year (year squared)

**Model 2: Negative binomial regression models adjusted for age, sex, study year, and quadratic term for year (year squared) and an interaction between year and index of multiple deprivation quintile

26

Figure 1. Locally weighted scatterplot smoothing curves showing age- and sex-standardised admissions rates for CVD causes in people with diabetes over the age of 45 years by deprivation quintiles between 2004-2005 and 2014-2015 in England. A) Acute myocardial infarction; B) Stroke; C) Coronary Artery Bypass Graft; D) Percutaneous Coronary Intervention. Rates are expressed as 100,000 people with diabetes over the age of 45 years.

100

120

140

160

180

Rat

e pe

r 100

,000

peo

ple

with

dia

bete

s

2004 2006 2008 2010 2012 2014Year

Quintile 1 (least deprived) Quintile 2 Quintile 3

Quintile 4 Quintile 5 (most deprived)

Acute Myocardial InfarctionA.

100

120

140

160

Rat

e pe

r 100

,000

peo

ple

with

dia

bete

s2004 2006 2008 2010 2012 2014

Year

Quintile 1 (least deprived) Quintile 2 Quintile 3

Quintile 4 Quintile 5 (most deprived)

StrokeB.

3040

5060

70R

ate

per 1

00,0

00 p

eopl

e w

ith d

iabe

tes

2004 2006 2008 2010 2012 2014Year

Quintile 1 (least deprived) Quintile 2 Quintile 3

Quintile 4 Quintile 5 (most deprived)

Coronary Artery Bypass GraftC.

6080

100

120

140

Rat

e pe

r 100

,000

peo

ple

with

dia

bete

s

2004 2006 2008 2010 2012 2014Year

Quintile 1 (least deprived) Quintile 2 Quintile 3

Quintile 4 Quintile 5 (most deprived)

Percutaneous Coronary InterventionD.

27