Impact of two different health systems on the burden of type 2 diabetes

Abstract

Objectives

Health policy directs the management of patients with chronic disease in a country, but evaluating nationwide policies is difficult, not least because of the absence of suitable comparators. This paper examines the management of patients with type 2 diabetes in two demographically comparable populations with different health care systems to see if this represents a viable approach to evaluation.

Methods

A secondary analysis of centralized prescribing databases for 2010 was undertaken to compare the levels and costs of care of patients with type 2 diabetes in Northern Ireland’s National Health Service (NHS) (NI, n = 1.8 million) which has structured care, financial incentives related to diabetes care and an emphasis on generic prescribing, with that of the Republic of Ireland (ROI, n = 4.3 million) where management of diabetes care is guided solely by clinical and other guidelines.

Results

The prevalence of treated type 2 diabetes was 3.59% in NI and 3.09% in ROI, but there were similar and high levels of prescribing of secondary cardiovascular medications. Medication costs per person for anti-diabetic, anti-obesity and cardiovascular medication were 46% higher in ROI than NI, due to differences in levels of generic prescribing.

Conclusions

These different health care systems appear to be producing similar levels of care for patients with type 2 diabetes, although at different levels of cost. The findings question the need for financial incentives in NI and highlight the large cost savings potentially accruing from a greater shift to generic prescribing in ROI. Cross-country comparison, though not without difficulties, may prove a useful adjunct to within-country analysis of policy impact.

Keywords

diabetes health policy evaluation health system comparisons

Introduction

The organization of health care, and in particular primary care services, independently contributes to the health outcomes of a population.¹ This is particularly true for the management of patients with diabetes in the UK which has over the past 30 years seen a shift in responsibility for routine diabetes care from secondary (specialist) to primary (generalist) care.^2,3

Structured diabetes care, which comprises a number of intervention strategies at the level of organizations, professionals and patients, has been associated with improved quality of care for patients.^4, ⁵ It was introduced in the UK National Health Service (NHS) in 2001 as part of a wider focus on chronic diseases and included National Service Frameworks (NSF), comprising national guidelines and plans for shared services. This was followed in 2004 with a new General Practitioner (GP) contract with financial incentives through the Quality and Outcomes Framework (QOF) (Table 1),^6, ⁷ where GPs attract additional income for targets relating to regular recall and review, tight glycaemic control and management of co-morbid conditions including cardiovascular disease (CVD). Research has shown that prevalence of a condition can influence the incentive reward to a GP.⁸ The introduction of the QOF saw the average GP wage rise of 58% in three years, and the overall cost currently stands at £1 billion annually.⁹

Table 1.

List of QOF diabetes indicators related to glycaemic control, blood pressure, cholesterol and prescribing of ACE inhibitors, 2009/2010.

Indicator name	Indicator wording	Available points	Threshold (%)
DM 5	The percentage of patients with diabetes who have a record of HbA1c or equivalent in the previous 15 months	3	40–90
DM 11	The percentage of patients with diabetes who have a record of blood pressure in the past 15 months	3	40–90
DM 12	The percentage of patients with diabetes in whom the last blood pressure is 145/85 mmHg or less	18	40–60
DM 15	The percentage of patients with diabetes with a diagnosis of proteinuria or micro-albuminuria who are treated with ACE inhibitors (or ARBs)	3	40–80
DM 16	The percentage of patients with diabetes who have a record of total cholesterol in the previous 15 months	3	40–90
DM 17	The percentage of patients with diabetes whose last measured total cholesterol within the previous 15 months is 5 mmol/l or less	6	40–70
DM 23	The percentage of patients with diabetes in whom the last HbA1c is 7 or less (or equivalent test/reference range depending on local laboratory) in the previous 15 months	17	40–50
DM 24	The percentage of patients with diabetes in whom the last HbA1c is 8 or less (or equivalent test/reference range depending on local laboratory) in the previous 15 months	8	40–70
DM 25	The percentage of patients with diabetes in whom the last HbA1c is 9 or less (or equivalent test/reference range depending on local laboratory) in the previous 15 months	10	40–90

Assessing the effectiveness of these policy changes has been challenging, and researchers have generally used time-series analysis to overcome the absence of comparators that complicates a nationwide initiative.^10, ¹¹ However, such approaches are protracted as the background trends and changes only become apparent after the collation of data over a number of years, and there is often a dearth of sufficient comparable data in earlier years. A comparison of the care of patients with diabetes in countries that are otherwise demographically very similar may provide a more efficient and mutually beneficial appraisal of the different health policies used to influence the management and care of patients with diabetes.

The island of Ireland presents an interesting exemplar. It comprises two separate jurisdictions, Northern Ireland (NI) and the Republic of Ireland (ROI). They have similar demographic profiles, birth and mortality rates, quality of life, health service utilization and obesity levels (Table 2). However, they have distinctly different health care systems.¹² NI forms part of the UK, and the NHS provides health care free at the point of delivery. ROI has a two-tier health care system, funded through a combination of taxation and fee for service dependent on income, age, illness or disability. This ‘one island: two systems’ approach therefore offers the unique potential of a natural experiment for comparing the effects of two distinct health care systems.

Table 2.

Demographic, socioeconomic and health characteristics of the NI and ROI populations.

Characteristics	ROI	NI
Demography
Age structure
0–15	21%	21%
16–64	67%	62%
65+	12%	17%
Crude birth rate (per 1000)	16.1	14.1
Life expectancy at birth
Men	77	78^a
Women	82	82^a
Adult mortality rate (per 1000, 15–59)	77	77^a
Under-5 mortality rate (per 1000 live births)	4	5^a
Ethnicity White	95%	99%
Prevalence of risk factors
Obesity	24%	25%
Overweight	39%	36%
Physical activity
High levels	24%	30%
Sedentary	29%	24%
Smoking (aged 15+, 2006)
Men	34%	26%
Women	28%	24%
Health
Quality of life
Very good/good	90%	86%
Poor	5%	5%
Diagnosed with a chronic illness at some time	10%	13%
Visited a GP in the last year	74%	73%
Hospitalized in the last year	10%	12%
Cervical Smear Test (in past year)	16%	30%
Smoking Cessation (advice in past year)	34%	58%

UK estimates.

Diabetes care in ROI is shaped by recommendations from an Expert Advisory Group, aligned with guidelines published by professional bodies such as the Irish College of General Practitioners (ICGP). Most diabetes care in primary care is non-structured and does not benefit from standardized patient registers or review^13–15; GPs do not receive financial incentives for the type or quality of care provided. Approximately 70% of the population pay a fee to attend their GP and to attend secondary care without a GP referral. There is a recognized need for closer integration between primary and specialist care,¹⁶ although some isolated areas adopt a structured or shared care programme and have been shown to have levels of process measures of performance, such as recording of HbA1c comparable to those in the UK.^14, ¹⁵ However, whether the structured and incentivized care in NI produces higher quality of care than that in ROI, which is not incentivized and guided solely by professional and other guidelines, is not known. According to the World Health Organization ranking, the health care systems in the UK and ROI are ranked 15th and 16th in Europe and 18th and 19th in the world, respectively.

The cross-country comparison also offers the opportunity to compare the effectiveness of cost containment relating to prescription medication in the two health care systems, and this is important as the costs associated with type 2 diabetes are high and disproportionate to the overall prevalence of the disease. Type 2 diabetes care in NI costs £400 m annually, representing 10% of the overall NHS budget and between 55 and 85% of this cost goes towards treating the complications of diabetes.¹⁷ In ROI, the cost of treating type 2 diabetes alone was estimated at £303 m per year in 2006.¹⁸ Drug costs in the UK are estimated to account for less than 10% of the overall cost of diabetes, but 25% in the ROI.^18, ¹⁹ Generic prescribing, although lower than in other parts of the UK, is a priority for the Health Service in NI and levels have increased as GPs have gained control of their practices’ prescribing budget.²⁰ Generic prescribing in ROI has not been encouraged to the same extent, and the growth of ‘branded generics’, generic drugs with alternative proprietary names that have come off patent, has also reduced the potential savings that may come from generic prescribing.

We postulate that the existence of a structured diabetes programme with financial incentives in NI will result in increased case ascertainment of type 2 diabetes in primary care due to the increased incentivization for good management of cases, resulting in:

An increased prevalence of treated type 2 diabetes;

A greater proportion of patients on co-morbidity medication, including combination insulin therapy, preventative CVD medication and anti-obesity medication;

A greater proportion of patients on generic medication;

And a lower overall medication cost per patient.

We use individual level prescription data to investigate these four objectives.

Northern Ireland data sources

The Enhanced Prescribing Database (EPD) was initiated in NI in May 2008 with the information on every prescription presenting to a pharmacy relayed electronically to a central body. This includes patient demographic data, including a unique patient identifier, prescribing practice code and information pertaining to the type, strength and amount of dispensed medication. The population denominators used in this study were the official population statistics from the NI Statistics and Research Agency for individuals aged 16 and over, and were n = 1,406,769 and n = 1,417,370 for 2009 and 2010, respectively. Ethical approval to use these data was granted by the Office for Research Ethics Committees Northern Ireland (Reference: 10/NIR02/20).

Republic of Ireland data sources

The ROI data were obtained from the NHS Executive-primary care reimbursement services (HSE-PCRS) prescribing databases, which provides free health care to all those with diabetes either through the medical card (GMS) scheme or free medications through the long term illness (LTI) scheme. In addition, the drug payment scheme (DPS) covers patients who are not covered by the GMS or LTI scheme and make a co-payment towards their medications. Data are available for all three community drug schemes, therefore, capturing nearly all prescribing for diabetes in primary care. The HSE-PCRS prescribing database contains demographic information on the patient (age and gender) and information relating to medicines dispensed such as the drug name, strength, quantity, costs and dates, but no information on diagnosis or outcomes. All prescription items are coded using the WHO Anatomical Therapeutic Chemical (ATC) classification system. The population denominators were n = 3,493,974 and n = 3,490,877 for 2009 and 2010, respectively.

We obtained prescription data from EPD and HSE-PCRS on known and pharmacologically treated type 2 diabetes in individuals aged 16 and over, between 1 January 2009 and 31 December 2010. A two-year period was chosen to increase the robustness of the analyses. As neither dataset includes clinical information, type 2 diabetes was defined as diabetes treated with oral anti-diabetic (hypoglycaemic) medication or a combination of oral medication and insulin; non-pharmacologically managed patients were not included. Case ascertainment was based on at least one prescription for anti-diabetic medication in the year in question (in NI, BNF Categories 6.1.1 [excluding 6.1.1.3] and 6.1.2; in ROI, ATC Code A10B). Age-standardized period prevalence was calculated.

The proportion of patients with diabetes on preventive CVD therapies was also assessed. Antiplatelets, beta-blockers, angiotensin-converting-enzyme (ACE) inhibitors and statins along with anti-obesity medication (orlistat) were chosen as they were included in the prescribing guidelines for both countries and were codeable in the respective datasets. Chi-square tests were used to statistically compare the proportions being treated for co-morbidities. The proportion and costs of generic and proprietary drugs prescribed for both oral anti-diabetic medication and statins were calculated. Insulin was excluded from the generic drug analysis, as GPs and pharmacists are advised not to prescribe generics for insulin.²¹ The annual prescription medication cost per person included the cost anti-diabetic medication, anti-obesity medication and preventive CVD therapies. As insulin was not included in the case definition, it was also excluded from the cost analysis. Costs were calculated using a per person average of the ingredient costs multiplied by the number of patients receiving the medication. Costs were calculated in Great British Pounds (GBP), and a conversion rate of 0.824 GBP to 1.00 Euro (EUR) was used.

Analysis was performed using Microsoft Excel, STATA version 11.2 (StataCorp 2009, College Station, TX, USA) and SAS version 9.2 (SAS Institute Inc, Cary, NC, USA).

Results

In 2010, there were 50,991 individuals in NI and 107,937 individuals in ROI aged 16 and over receiving medication for type 2 diabetes. This equated to a crude prevalence of 3.59% in NI and 3.09% in ROI. Prevalence increased from 2009 to 2010 in both jurisdictions, with NI increasing by 13% and ROI increasing by 11%. Prevalence rose in NI in every age group except the youngest and the ‘70–74’age group; the largest differences were in the ‘75+’ age group, with the figure in NI 10% higher over the study time period (Table 3).

Table 3.

Prevalence of treated type 2 diabetes in ROI and NI, 2009 and 2010.

ROI		NI
2009 (%)	2010 (%)	2009 (%)	2010 (%)
Total	2.84	3.09	3.24	3.59
Age group
16–24	0.19	0.17	0.06	0.08
25–34	0.33	0.34	0.34	0.36
35–44	0.93	1.03	1.02	1.11
45–54	2.55	2.76	2.73	2.94
55–64	5.87	6.28	5.85	6.35
65–69	9.06	9.63	8.98	9.64
70–74	11.16	11.67	10.86	11.56
75+	8.90	9.41	9.43	10.98

The proportions of individuals treated on combination insulin therapy were 15% in NI and 13% in ROI, with NI consistently higher across the age groups.

The rates of prescribing of preventive CVD therapies were high in both populations; reaching 95% of patients in NI and 92% in ROI. Similar proportions of patients with type 2 diabetes were prescribed beta-blockers, antiplatelets and ACE inhibitors (Figure 1). Statins were the most commonly prescribed CVD medication at 85% and 75% of patients in NI and ROI, respectively. Only 4% and 3% of the patients in NI and ROI, respectively, received a prescription for anti-obesity medication. The NI proportions were found to be significantly greater for insulin, anti-obesity, total CVD medications and statins, while the ROI proportions were significantly greater in the other three categories. However, although the small differences were all found to be significant at P < 0.001, this was due to the large sample sizes and most of the differences were not clinically significant.

Figure 1.

Proportion of patients with type 2 diabetes co-prescribed medication.

The annual cost of prescription drugs related to the treatment of type 2 diabetes was substantially higher in ROI at £517.37 per person compared to £353.79 in NI (Table 4). The cost of a patient receiving all the listed medication was nearly twice as high in ROI, at £988.50 compared to £562.77 in NI. The total cost per capita of oral anti-diabetic medication alone was more similar, amounting to £127.58 in ROI compared to £118.85 in NI. Most of the difference in costs was due to the cost of preventive CVD medication; the ROI cost of £560.45 per person was nearly 80% greater than the NI figure of £317.32. Preventative CVD medication accounted for 74% of the total prescription cost in ROI and 65% of total costs in NI. Generic prescribing was higher in NI, making up over 50% of all statin and oral anti-diabetic medication, with the comparable ROI figures at 8% and 6%, respectively.

Table 4.

Total cost of diabetes-related medication per capita and proportion of patients with diabetes prescribed medications in NI and ROI, 2010.

Diabetes-related medication	NI				ROI
Diabetes-related medication	Cost per capita (£)	Proportion prescribed (%)	Total Cost (£)	% of total cost	Cost per capita (£)	Proportion prescribed (%)	Total Cost (£)	% of total cost
Type 2 diabetes (oral)	118.85	100.0	5,421,806	34	127.58	100.0	12,650,571	25
CVD total^a	317.32	95.2	13,919,923	65	560.45	92.1	37,756,620	74
Antiplatelets	50.52	64.0	1,474,788	9	17.77	66.3	1,168,459	2
Statins	161.62	85.0	6,266,682	39	302.64	74.3	22,296,532	43
Beta-blockers	34.91	32.4	516,013	3	66.38	34.0	2,237,916	4
ACE inhibitor	70.27	67.9	2,176,692	13	173.66	70.0	12,053,714	23
Obesity	126.60	4.9	282,974	2	300.47	3.0	893,824	2
Total cost (combined)^b	562.77	–	16,138,954	–	988.50	–	51,301,015	–
Total cost (per capita receiving medication)^c	353.79	–	–	–	517.37	–	–	–

Note: Cost per capita is calculated by dividing the total ingredient cost by the number of patients receiving the medication.

CVD total = Patient receiving at least one of antiplatelets, statins, beta-blockers and ACE inhibitors.

Total cost (combined) = total additive cost of all medications, i.e. for each patient receiving all medications.

Total cost (per capita receiving medication) = total cost of all medication divided by number of patients receiving that medication, i.e. actual cost per capita.

Discussion

Although NI and ROI have different organizational approaches to diabetes policy and differences in the incentivization of GPs, the treatment of patients with type 2 diabetes in primary care (within the parameters examined) is remarkably similar within the two jurisdictions. There are, however, substantial differences in medication costs, largely related to the substantial variation in rates of generic prescribing.

The variation in prevalence may be due to a number of factors. A true difference in the prevalence of type 2 diabetes between ROI and NI is unlikely, given the populations’ demographic similarities.¹² It is possible that significant differences exist between the jurisdictions in the proportion of patients with type 2 diabetes who are non-pharmacologically managed. However, this population is especially difficult to identify and the use of prescription data automatically excludes them from the study population.

Differences in case-fatality rates are also possible, but could not be checked as the requisite data linkage is not currently available in ROI. However, the consistency of the differences in prevalence across the age-range and equivalent levels of secondary prevention of heart disease in both jurisdictions would argue against this explanation. The most likely explanation is a slight increase in case ascertainment due to the chronic disease management orientation of the NI health system, although specific research into the rate of screening for diabetes in primary care and the characteristics of the incident populations in each jurisdiction would provide greater clarity on this issue. Also, a higher rate of non-pharmacological management of patients with type 2 diabetes in NI, which would be in keeping with the incentives for tighter glycaemic control, cannot be ruled out.

Tight control of blood pressure in patients with type 2 diabetes significantly reduces the risk of complications and death,²² and cholesterol-lowering therapy reduces the rate of major vascular events in patients.²³ Both jurisdictions have guidelines regarding the secondary prevention of CVD and recommend regular blood pressure monitoring of patients with diabetes and prescribing of secondary preventive therapies in individuals presenting with early risk factors.^24, ²⁵ In practice, it is estimated that this would affect more than 80% of all patients with type 2 diabetes.²⁶ Our findings correspond with this estimate, suggesting that the majority of patients at risk are being successfully treated in both populations. It is beyond the remit of this study to identify whether patients are initiated on treatment within primary or secondary care. However, the large proportions indicate that primary care is identifying and treating high risk patients.

Within the NHS QOF, there are five incentivized indicators directly or indirectly related to prescribing of preventive CVD therapies in patients with diabetes; two for the management of cholesterol levels, two for control of hypertension and one for prescribing of ACE inhibitors. There are no equivalent incentives in ROI. Despite this, we found no major differences in levels of prescribing between the populations. Furthermore, prescribing of these therapies is so high that it is likely to have reached a ‘ceiling’, meaning that prescribing levels cannot increase further despite the presence of incentives.²⁷ These findings indicate a need to re-consider the cost-effectiveness of these specific incentives in the NHS.

Large numbers of GPs in ROI do not have any formal shared protocol with secondary care providers, do not maintain a diabetes register and do not engage in any regular recall and review of patients.¹³ In the UK NHS, these aspects of care are all either directly or indirectly incentivized under the QOF and have all been shown to have improved in the UK since the introduction of the QOF in 2004.²⁸ McHugh et al.¹⁴ assessed the performance of three primary care-led initiatives in small geographic areas of ROI, and, although they did not compare the overall quality of diabetes care in ROI, the study highlighted the comparable levels of process measures of performance found in selected areas with a structured system of diabetes care, despite the lack of any financial incentives.

A 2012 postal survey study found that there remains considerable variation in diabetes care in the UK, from the perspective of health care professionals and patients.²⁹ This corresponds with the National Diabetes Audit in England, which shows that individual quality indicators are generally high. However, the proportion of patients receiving more than six of the nine fundamental elements of diabetes care was under 90%, with the proportion receiving all nine under 55%. Some aspects of care, such as weight management advice and self-management, were found to be lacking, despite health care professionals reporting high rates of discussing these areas. The authors stated that some of these differences may be related to recall bias on the part of the patient as it is easier to remember a distinct examination than advice. This suggests that the indicators that are currently used to define high quality patient care may be inadequate.

This study highlights the large differences in the medication costs for a patient with type 2 diabetes in NI and ROI, especially costs associated with CVD medication, which were 80% greater per capita in ROI. The level of generic prescribing in NI is amongst the lowest in the UK, but at less than 10% the level in ROI is a fraction of this. Increased use of generic medication alternatives is a simple and obvious method of conserving resources. In ROI, use of generic medication is linked to pharmacy contracts; therefore any increase will require policy changes at a national level. However, the introduction of reference pricing and generic substitution in ROI is planned. If generic substitutes were prescribed at similar levels to NI, costs could be reduced to 57% of current costs in ROI (comparing simvastatin and Zocor®). If generic prescribing of statins increased to 80% in both jurisdictions, costs in NI could reduce to 55% of the current level, while costs in ROI could fall to 25%. These are modest estimates. Research has shown that in the UK savings of up to £2bn are still possible, despite the already high levels of generic prescribing.³⁰ Obviously, these costs contribute only a minority of the costs associated with type 2 diabetes; however, we believe that this increases the potential for cost reduction in the future. It must be acknowledged that these cost comparisons are only the tip of the iceberg in terms of the total cost related to diabetes.

This study has several limitations. First, the use of prescription data as a proxy for diagnosis automatically excludes the approximately one-third of patients with type 2 diabetes who are managed without medication.³¹ Second, it was not possible in the current study to examine prognostic markers such as HbA1c or clinical outcomes. A cross-country comparison of mortality rates would be informative and could provide greater evidence to examine the current policies associated with chronic disease management of type 2 diabetes. Finally, neither the NI nor the ROI dataset allowed for adjustment of any confounding variables at the practice level, such as GPs’ propensity to test for diabetes.

Conclusion

This study shows that cross-country comparisons of health care policies can be of mutual benefit and indicate where changes might be made. However, this approach is not without its problems. It is often difficult to garner equivalent sets of relevant data and given the complexities of the organizations of care, it is often difficult to separate out which components of policy and organizational arrangements are effecting change. Notwithstanding these difficulties, cross-country comparisons should be considered alongside traditional intra-country analysis in assessing health care policy appraisal.

Footnotes

Acknowledgements

We thank the HSE-PCRS for supplying the data on which the ROI study was based and the BSO for supplying the data on which the NI study was based.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Starfield

. Primary care: balancing health needs, services, and technology, New York: Oxford University Press, 1998.

Collins

O’Sullivan

Harkins

. Quality of life and quality of care in patients with diabetes experiencing different models of care. Diab Care 2009; 32: 603–605.

Goyder

McNally

Drucquer

. Shifting of care for diabetes from secondary to primary care, 1990-5: review of general practices. BMJ 1998; 316: 1505–1506.

Olivarius

Beck-nielsen

Andreasen

. Randomised controlled trial of structured personal care of type 2 diabetes mellitus. BMJ 2001; 323: 1–9.

Renders

Valk

Griffin

. Interventions to improve the management of diabetes mellitus in primary care, outpatient and community settings. Cochrane Database Syst Rev 2001; 1: 1–142.

Roland

. Linking physicians’ pay to the quality of care – a major experiment in the United Kingdom. N Engl J Med 2004; 351: 1448–1454.

Department of Health, United Kingdom. The National Service Framework for Diabetes: Standards, 2001.

Gravelle H, Sutton M and Ma A. Doctor behaviour under a pay for performance contract: evidence from the quality and outcomes framework, University of York, Centre for Health Economics, working paper, 2007.

Cashin C. United Kingdom: Quality and Outcomes Framework (QOF). Major Developments in Results-Based Financing (RBF) in OECD Countries: Country Summaries and Mapping of RBF Programs. The World Bank, 2011.

10.

Campbell

Reeves

Kontopantelis

. Effects of pay for performance on the quality of primary care in England. N Engl J Med 2009; 361: 368–378.

11.

Khunti

Gadsby

Millett

. Quality of diabetes care in the UK: comparison of published quality-of-care reports with results of the Quality and Outcomes Framework for Diabetes. Diabet Med 2007; 24: 1436–1441.

12.

Ward M, McGee H, Morgan K, et al. SLÁN 2007: Survey of lifestyle, attitudes and nutrition in Ireland. “One Island – One Lifestyle?” Health and lifestyles in the Republic of Ireland and Northern Ireland: comparing the population surveys SLÁN 2007 and NIHSWS 2005. Dublin: Stationery Office, Department of Health and Children, 2009.

13.

Mc Hugh

O’Keeffe

Fitzpatrick

. Diabetes care in Ireland: a survey of general practitioners. Prim Care Diab 2009; 3: 225–231.

14.

Mc Hugh

Marsden

Brennan

. Counting on commitment; the quality of primary care-led diabetes management in a system with minimal incentives. BMC Health Serv Res 2011; 11: 348–348.

15.

Brennan

Harkins

Perry

. Management of diabetes in primary care: a structured-care approach. Eur J Gen Pract 2008; 14: 117–122.

16.

Smith

. Primary care diabetes in the Republic of Ireland. Prim Care Diabetes 2007; 1: 207–208.

17.

Gale

. Diabetes in the UK: time for a reality check? Diabet Med 2010; 27: 973–976.

18.

Nolan

O’Halloran

MCKenna

. The cost of treating type 2 diabetes (CODEIRE). Ir Med J 2006; 99: 307–310.

19.

Diabetes UK. Diabetes in the UK 2010: key statistics on diabetes. 2010.

20.

Rafferty

Wilson-Davis

McGavock

. How has fundholding in Northern Ireland affected prescribing patterns? A longitudinal study. BMJ 1997; 315: 166–170.

21.

DHSSPSNI. Pharmaceutical clinical effectiveness programme, http://www.dhsspsni.gov.uk/pas-psip (2008, accessed 7 August 2013).

22.

Stratton

Adler

Neil

HAW

. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. BMJ 2000; 321: 405–412.

23.

Collins

Armitage

Parish

. MRC/BHF Heart Protection Study of cholesterol-lowering with simvastatin in 5963 people with diabetes: a randomised placebo-controlled trial. Lancet 2003; 361: 2005–2016.

24.

The National Collaborating Centre for Chronic Conditions. Type 2 diabetes: national clinical guideline for management in primary and secondary care (update), National Institute for Health and Clinical Excellence: Guidance. London: Royal College of Physicians, 2008.

25.

Mcgee H. Department of Health and Children. Changing cardiovascular health. National Cardiovascular Health Policy 2010–2019. Dublin: Government Publications, 2010.

26.

Bennett

Williams

Feely

. Under-prescribing of cardiovascular therapies for diabetes in primary care. Eur J Clin Pharmacol 2003; 58: 835–841.

27.

Campbell

Reeves

Kontopantelis

. Effects of pay for performance on the quality of primary care in England. N Engl J Med 2009; 361: 368–378.

28.

Sutton

Elder

Guthrie

. Record rewards: the effects of targeted quality incentives on the recording of risk factors by primary care providers. Health Econ 2010; 19: 1–13.

29.

Hawthorne

Hrisos

Stamp

. Diabetes care provision in UK primary care practices. PLoS One 2012; 7: 1–7.

30.

Moon

Bogle

. Switching statins. BMJ 2006; 332: 1344–1345.

31.

Hippisley-Cox

Pringle

. Prevalence, care, and outcomes for patients with diet-controlled diabetes in general practice: cross sectional survey. Lancet 2004; 364: 423–428.