SAL Farooqiq7 PD 3-Col

State of the art lecture

Primary care: the custodian of diabetes care?

Azhar Farooqi Abstract General Practitioner and Honorary Professor, Type 2 diabetes is the disease of our times. With some 3 million cases in the UK, primary care University of Leicester, East Leicester Medical Practice, has a vital role in coordinating and delivering evidence-based care which includes the Leicester, UK prevention, detection and management of diabetes. However, primary care has not always been able to deliver such a role; up to the end of the 1980s, despite the drawbacks of busy hospital outpatient clinics, primary care could rarely Correspondence to: offer the systematic care and skills that people with diabetes require. Professor Azhar Farooqi OBE, East Leicester Medical Quality improvement and audit in the 1990s heralded the increased adoption of Practice, 131 Uppingham Rd, Leicester LE5 4BP, UK; email: [email protected] evidence-based practice in primary care. Many GP practices significantly improved the organisation and quality of care for diabetes as a result. The widespread adoption of IT Received: 30 July 2012 systems and the emergence of a more robust evidence base for care (for example, UKPDS) Accepted: 1 August 2012 accelerated this process. More lately, investment in general practice through the Quality and Outcomes Framework This paper was presented as the 2012 Mary and practice education programmes have helped deliver significant improvements in the Mackinnon lecture at the 2012 Diabetes UK Annual quality of primary care diabetes. Professional Conference held in Glasgow However, there is still much to do, with variation in care and health inequalities persisting. The development of clinical commissioning offers further opportunities to make the best use of available resources and target investment where it is most likely to benefit patients. A health care system where primary care in collaboration with other stakeholders coordinates the care of people with diabetes offers the best hope in addressing this modern epidemic that we face. Copyright © 2012 John Wiley & Sons. Practical Diabetes 2012; 29(7): 286–291 Key words primary care diabetes; quality improvement; clinical commissioning; education and diabetes

Introduction Background: the burden and One of the cornerstones of the challenge of diabetes National Health Service in the Type 2 diabetes is the disease of our United Kingdom is a primary care times. The UK prevalence of this system where all citizens are able to chronic condition has increased register with a doctor-led service. from an estimated 1.4 million in This service not only holds the 1998 to almost 2.6 million in 2010; central patient record, but also this is predicted to rise to 4 million offers health care from conception by 2020.3 The numbers of people to the grave. It is ideally placed with impaired glucose tolerance and to coordinate and deliver the at risk of future diabetes is estimated life-long, holistic care required by at double those with confirmed dis- people with long-term conditions ease. The financial costs of diabetes, such as diabetes. which already consume >10% of Mary Mackinnon (whose consid- the health care budget,4 are set to erable contributions to diabetes are increase with the rising numbers of well described in Richard Holt’s patients, and with the increase in Mary Mackinnon lecture of 2011)1 expensive complications such as recognised this potential over three renal failure and cardiovascular decades ago. Mary spent much of disease. The human costs of this her early career improving commu- huge burden of poor health and lost nity services for people with dia- years of life are incalculable. betes; she subsequently went on The positive side of this dim to develop and deliver pioneering picture is that, unlike 30 years ago, diabetes education for primary care we now have a robust evidence base professionals.2 It is therefore fitting on prevention, early detection (in that in this lecture I consider the the case of type 2 disease) and progress of diabetes services in management of established dis- primary care, and whether primary ease.5,6 Appropriate interventions care is indeed achieving its potential can prevent or delay the complica- to be a fit custodian for the care of tions of diabetes, and we now have people with diabetes. technologies which allow many

complications, if detected early (for 11 12 example, diabetic retinopathy) to Hayes TM, Harries J. Cardiff, UK Singh B, et al. Wolverhampton, UK be treated effectively. GP patients vs OPD Structured GP mini-clinics vs OPD These developments, which apply to a large patient group with 200 patients, 100 discharged to GP Matched pair design, n=4222 patients lifelong disease, pose a real challenge on how best to organise 5-year follow up services for diabetes. A conservative estimate based on just four reviews HbA1c 9.5% (hospital) vs 10.5% (GP) 9.6% (hospital) vs 9.7% (GP) OHA or follow ups a year, means that in the UK a minimum of 12 million Higher mortality in primary care 10.6% (hospital) vs 10.8% (GP) BD insulin consultations per year are required for people with proven diabetes. Only 13% in GP seen annually Only small % of all practices ran mini-clinics While some of these patients will have complex disease or complica- Hospital care superior No difference in HbA1c after 2 years tions which require hospital-based between hospital and GP patients specialist care, it is clear that for the vast majority of patients care can GP = general practice; OPD = outpatient diabetes clinic; OHA = oral hypoglycaemic agents; only realistically be met in primary BD = twice-daily care, where general practices already deliver a doctor-led service Table 1. Should type 2 diabetes be managed in hospital or general practice? Summary of 2 papers providing 330 million face-to-face published in the British Medical Journal in 198411,12 patient consultations per year.7 The challenges for primary care, primary care. Eric Wilkes10 explored the hospital service had better gly- however, are considerable. These some of these issues further when a caemic control and lower mortality include ensuring staff have the large cohort of patients was dis- compared to those followed up in appropriate skills, support and time charged from the Sheffield hospital general practice. The study showed to deliver consistent and high service in the late 1970s. Wilkes, in a that only 13% of patients in the gen- quality diabetes services. These prospective study, demonstrated that eral practice arm attended regular challenges have increased over the many practices did not have the follow up. In contrast, a study by last three decades as diabetes has infrastructure to adequately follow Singh, Holland and Thorn,12 pub- become more common, and the up such patients. In his perceptive lished in the same journal, demon- evidence base for treatment more summary of this study, Wilkes articu- strated that patients discharged to clear. I will explore how primary lates the dilemma patients faced: selected practices which provided care has evolved to try and ensure it ‘...deaf, anxious partially sighted organised diabetes care (diabetes can deliver the services people with patients not accustomed to rapid learning mini-clinics) had outcomes which diabetes deserve. in strange hospital environments absorb were as good as hospital-based little of what they are told, but busy follow up. Interestingly, these The 1980s, the decade of GPs attuned to contractual requirements practices included those which realisation and herculean tasks in the 5-minute Thorn had worked with in the 1970s The debate as to whether diabetes consultation may not seek out extra to develop mini-clinics.9 (in particular type 2 disease) should work; even if they do, they will have to Both the Hayes and the Singh be managed in hospital by specialists pay for the detection and pursuit of studies show that the glycaemic con- or by GPs in primary care is not a non-attenders themselves.’ trol for hospital and general practice new one. Even in the 1970s when the In the 1980s, a common view patients was by today’s standard prevalence of diabetes was much (particularly among specialists) was poor, with the average HbA1c lower, questions were being asked as that general practice simply could achieved >10% (86mmol/mol). to where care was best undertaken. not be trusted to look after diabetes These and other studies provide Pioneers such as Professor John and, however imperfect, large-scale objective evidence that most general Malins recognised as far back as the hospital outpatient care was the practices at this time had neither the 1960s that crowded hospital outpa- only solution. structure, nor the organisation, to tient clinics were not best placed to In 1984 two papers, published in look after patients with diabetes provide care for many patients. the British Medical Journal, con- effectively. However, well-organised Malins in Birmingham and Thorn in tributed further to this debate practices (although still a minority) Wolverhampton did some of the (Table 1). Hayes and Harries11 pub- could deliver care equivalent to that early work to enhance the capability lished the results of a study in which provided in hospital diabetes clinics. and expertise of general practice, patients attending a hospital dia- This conclusion is further con- including assisting practices to set betes clinic were randomly allocated firmed in a meta-analysis by Griffin et up structured mini-clinics.8,9 either to continue attending hospi- al. (Figure 1).13 The difficulties of oversubscribed tal, or to receive follow-up care in The time up to and including the diabetes clinics prompted some hos- general practice. Five-year follow-up 1970s were the decades of darkness pitals to discharge patients back to data showed that patients attending as far as primary care diabetes was

concerned. Professor John Malins summarised it thus: Favours general practice Favours hospital care ‘Most physicians who run diabetes Trial clinics would be glad to know of any sat- Hayes (1984) (n=164) isfactory method by which patients Hoskins 1 (1992) (n=134) could be returned to the care of their own doctors.’ Hoskins 2 (1992) (n=137) The 1980s was the decade of a Hurwitz (1993) (n=166) realisation that hospitals were not DICE (1994) (n=235) the ideal place for the care of patients, but also that most general Stratified analysis practices were not yet equipped Routine general practice care to offer an alternative service. (2 trials, n=301) Only with this realisation could the Prompted general practice problem begin to be addressed. care (3 trials, n=535) The 1990s, the decade of clinical audit and quality improvement -1.5 -1 -0.5 0 0.5 1 1.5 In the 1990s, important changes Difference were beginning to have an impact in general practice: vocational Weighted difference in mean percentage of glycated haemoglobin between general practice and training was now well established, hospital care. Bars indicate 95% confidence intervals general practices were starting to 13 become computerised and, with Figure 1. Systematic review of GP vs hospital care. (Griffin S, et al. Diabetes care in general the support of the Royal College of practice: meta-analysis of randomised control trials. BMJ 1998; 317[7155]:390–6. Permission to General Practitioners and the publish has been granted from © BMJ Publishing Group Ltd) Department of Health, quality improvement was now on the when many doctors were still not in the incentivised Quality and agenda for general practice. convinced of the benefit of such Outcomes Framework (QOF) for Clinical audit as a tool was increas- activity, and the process of audit was primary care.20 ingly used to promote quality in a time-intensive activity based on primary care, often facilitated by review of data from manual records. The 2000s, the decade of newly established Medical Audit The process of clinical audit frameworks: the NSF and QOF Advisory Groups (MAAGs).14 The highlighted barriers to improved In 2001, the NSF for diabetes was definition of clinical audit, as ‘a patient care and allowed the devel- published; the 12 standards in the quality improvement process that opment of strategies for improve- NSF articulate the ambition that the seeks to improve patient care ment; these included improved NHS has set for the quality of care and outcomes through systematic team working, training and more people with diabetes should receive. review of care against explicit crite- effective use of information technol- These standards encompass the ria and the implementation of ogy for record keeping.18 adoption of evidence-based practice, change’, makes it clear that prac- An important development in the importance of involving patients tices were expected to evaluate the late 1990s was the publication in their own care, and the need to their practice, and make changes of the UK Prospective Diabetes invest in diabetes services in all parts to improve patient care. Study trials,5,6 which provided a of the NHS. The implementation of Diabetes was a popular clinical robust evidence base on the impor- the NSF, although patchy, has had topic for audit, with a nationally tance of effective management of real benefits – examples of which recognised, evidence-based clinical blood glucose, and cardiovascular include the establishment of effec- audit protocol developed in disease risk factors in improving out- tive retinal screening programmes, Leicester.15 The impact of clinical comes in type 2 diabetes. This the development of local clinical audit of diabetes care in over 60 gen- further strengthened the realisation networks for diabetes and the eral practices in Leicester16 demon- that a systematic approach to quality appointment of a national cham- strates that, with the support of improvement in diabetes care pion for diabetes (with Dr Sue external advice, practices were able should be supported, and even Roberts as the very able first to establish accurate disease regis- incentivised, within the NHS. appointment in this post). ters, initiate call and re-call of These developments culminated The NSF contributed to the envi- patients and undertake structured in the emergence of a national ronment that enabled the resourc- care with significant improvement in consensus on the ambitions for dia- ing of the QOF (diabetes standards) performance. Khunti et al.17 demon- betes care with the publication of as a means to achieve improved care. strated how such audits were improv- the National Service Framework The QOF builds on the previous ing care not just in Leicester, but in (NSF) for diabetes,19 and in practi- work in clinical audit and quality many parts of the country. This was a cal changes to improve care improvement, by ensuring diabetes significant achievement at a time such as the diabetes standards registers and audit of performance

became a routine part of general 1999 2009 practice. The improvement of GP Collation of multi-practice audits17 National Diabetes Audit23 computer systems, adoption of evidence-based National Institute for Retinal screen 67.5% 78.9% Health and Clinical Excellence guidance for diabetes (e.g. the guideline Foot check 67.7% 85.2% for type 2 diabetes)21 and support from primary care trusts (e.g. in the HbA1c 72.5% 92.6% adoption of prescribing and practice quality reviews) means that perform- Cholesterol 37.5% 92.4% ance data in most practices are now routinely collected, and outliers can Table 2. Impact of the Quality and Outcomes Framework: uptake of annual process measures17,23 easily be identified and supported to improve. The 2010s, the decade of the Primary Care Diabetes Society, The impact of the QOF and the collaboration and empowerment Department of Health and Diabetes supporting changes on standards of The challenge for this decade is to UK has developed a detailed patient care have been significant. build on the improvements that ‘Diabetes Commissioning Toolkit’, In contrast to the situation that primary care has so far achieved, in which offers clear and systematic prevailed in the 1980s, patients with particular to reduce the variation guidance on the commissioning diabetes are now much more likely and inequalities in care that persist. process.28 Some areas with the to receive regular assessments and Housebound patients26 and those support of PCTs, and with strong standards of care, such as blood from ethnic minorities,27 for exam- clinical leadership, have managed glucose control (HbA1c) and cardio- ple, continue to receive poorer care. to adopt such methods to reshape vascular risk factor management Significant differences in outcomes local services (for example, in having improved significantly. (such as amputation rates) and take Cumbria, Northumberland and Table 2 illustrates the improve- up of key care processes remain Leicester). However, such developments achieved by general practices between different parts of the ments have not been the norm, in England and Wales between 1999 country (with up to a three-fold with the status quo prevailing in (pre-QOF) and 2009.17,22,23 variation between primary care trust many areas. Primary care has also proved [PCT] areas).23 In 2012, the Health and Social adept at detection of diabetes – with While there are many interven- Care Act29 heralded the establish- over 1 million new patients detected tions that can improve services ment of Clinical Commissioning in the last decade,3 the gap between for people with diabetes, I will Groups (CCGs). These clinician-led the actual and expected prevalence focus on three key developments organisations will be fully estab- having narrowed significantly.24 which I believe will help ensure lished by 2014, and have real author- Further evidence on the primary care delivers a diabetes ity to ensure clinical commissioning progress that primary care has made service fit for the 21st century. delivers effective services. in improving the management of Key drivers will be improving type 2 diabetes comes from the Commissioning of effective quality and reducing inequalities of ADDITION study (Table 3).25 This patient pathways service provision. CCGs will also large, multicentre intervention trial The NHS – through the NSF, QOF have a responsibility to provide cost- for cardiovascular risk factors and and national clinical guidelines – effective care – for example, by blood glucose control in newly- has attempted to ensure that moving services from expensive diagnosed type 2 diabetes demon- national standards of care for dia- hospitals (inpatient care currently strates that routine management betes are implemented. While these consumes 66% of the NHS spend on in primary care is comparable to initiatives have enjoyed significant diabetes)4 to more cost-effective and intensive hospital-based care. success, many patients still receive accessible community-based care. The progress I describe is signifi- suboptimal care. One of the key rea- Effective commissioning will also cant and impressive; however, there sons for this has been the failure to enable investment in services which are still difficulties to overcome. reorganise service provision to have poor local provision, such as While Malins would now be reas- address adequately the entire care retinopathy screening, foot care or sured that most patients can now pathway for people with diabetes. dietetic support. indeed be safely discharged to the Such an approach would allow for There is now a growing evidence care of their GP, the National appropriate investment in under- base on providing better services for Diabetes Audit23 demonstrates that funded areas such as prevention and patients from ethnic minorities.30,31 the variation in the quality of care of early detection, as well as ensuring Effective commissioning by estab- patients persists. Wilkes might still services are available for difficult to lishing local needs will help ensure argue that general practitioners reach groups. these are accommodated in local (some at least) still do not have For the past 10 years, PCTs have services. Although CCGs are led by the time and resources to offer the had the ability to commission GPs, the experience to date suggests standard of care that patients diabetes services appropriate for that effective commissioning will should expect. their population. A collaboration of require close collaboration between

Variable Routine care (general practice) Intensive treatment Change from baseline to follow-up β/odds ratio (95% CI)

Baseline Follow up Baseline Follow-up (n=1379) (n=1285) (n=1678) (n=1574)

Total with Value Total with Value Total with Value Total with Value data data data data available available available available (%) (%) (%) (%)

HbA1c (%); mean (SD) 1298 (94.1) 7.0 (1.5) 1226 (95.4) 6.7 (0.95) 1591 (94.8) 7.0 (1.6) 1513 (96.1) 6.6 (0.95) -0.08 (-0.14 to -0.02)

Systolic blood pressure 1346 (97.6) 149.8 (21.3) 1205 (93.8) 138.1 (17.6) 1617 (96.4) 148.5 (22.1) 1517 (96.4) 134.8 (16.8) -2.86 (-4.51 to -1.20) (mmHg); mean (SD)

Diastolic blood pressure 2346 (97.6) 86.5 (11.3) 1203 (93.6) 80.7 (10.8) 1618 (96.4) 86.1 (11.1) 1517 (96.4) 79.5 (10.7) -1.44 (-2.30 to -0.58) (mmHg); mean (SD)

Total cholesterol 1300 (96·3) 5.6 (1.2) 1226 (95.4) 4.4 (0.9) 1593 (94.9) 5.5 (1.1) 1523 (96.8) 4.2 (0.9) -0.27 (-0.34 to -0.19) (mmol/L); mean (SD)

Table 3. Effect of early intensive multifactorial therapy on 5-year cardiovascular outcomes in individuals with type 2 diabetes detected by screening (ADDITION-Europe): a cluster randomised trial.25 (Griffin SJ, et al. Lancet 2011;378(9786):156–67. Permission to publish has been granted from © Elsevier) all stakeholders, including patients, greater levels of skills in primary Conclusion carers and diabetes specialists in care, and has decided to invest One of the most talented academic order to deliver sustainable and £1 million in its 65 member general practitioners of our age, appropriate change. practices to ensure they acquire the Trisha Greenhalgh, commented in skills to deliver diabetes care for 1998:13 Diabetes training for primary care their patients. ‘We still do not know the precise professionals Such an initiative requires collab- competencies required for delivering If primary care is to play a greater oration with specialists who need to different aspects of education, support part in the management of diabetes, lead much of the training activity and surveillance for patients with dia- then primary care professionals will and suggests a model others may betes, and we certainly do not know how need to acquire the necessary skills. wish to follow. to communicate across interprofessional Mary Mackinnon recognised this boundaries in so-called seamless care.’ when she started education pro- Patient empowerment These issues remain a challenge grammes for primary care at Empowerment of patients to self- to this day, but I hope that I have Warwick University in the early manage is key to limiting the devas- demonstrated that primary care has 1990s. Such programmes are now tating impact of the diabetes been on a journey to address these much more available, with other epidemic. Evidence on the effective- challenges and is now in a very dif- centres (such as Leicester and ness of patient education pro- ferent place from even 10 years ago. Bradford) offering postgraduate grammes to enable patient empow- The term ‘custodian’ can refer to courses in diabetes. Professional erment is mixed. the supervision of an act of impris- groups, such as the Primary Care There is good evidence for some onment or to the watchful care of an Diabetes Society and the pharma- programmes such as, for example, individual. While the care of patients ceutical industry, have also invested insulin management for type 1 dia- must always be a collaborative effort resources and time in up-skilling betes (DAFNE),33 and in some set- between health care professionals, I primary care, for example in topics tings intensive programmes for believe that primary care is now in a such as insulin management. prevention of diabetes have been suc- position to discharge its responsibil- While in the 1990s staff with cessful. The evidence for large-scale ity of watchful care for people with enhanced skills in primary care were education programmes for people diabetes; we owe them no less. rare,32 more recent increases in with type 2 diabetes (for example, the 34 uptake of education and training DESMOND programme) and for Declaration of interests mean that many more primary care lifestyle change is less clear. There are no conflicts of interest practitioners now have enhanced Further research in developing declared. skills in diabetes. interventions which are effective in Leicester City Clinical Comm- empowering patients in the primary issioning Group, where the author is care setting is vital, if we are to have References 1. Holt RIG. Undoing Descartes: integrating diabetes based, has recognised that any serv- a real impact in controlling the care for those with mental illness. Pract Diabetes ice reconfiguration will require impact of diabetes in society. 2011:28:270–5.

2. Mary Mackinnon. www.diabetes-stories.com/ 14. Baker R, Fraser RC. MAAGs (Medical Audit Advisory ing (ADDITION-Europe): a cluster-randomised trial. transcript.asp?UID=62 [accessed 13 June 2012]. Groups): the Eli Lilly National Clinical Audit Centre. Lancet 2011;378(9786):156–67. 3. Diabetes in the UK 2010: Key statistics on diabetes. Int J Health Care Qual Assur 1993;6(3):4–8. 26. Farooqi A, Sorrie R. Monitoring of elderly house- Diabetes UK, 2010. www.diabetes.org.uk/. 15. Baker R , et al. Monitoring diabetes. Audit protocol bound and mobile diabetic patients in 31 4. Kanavos P, et al. Diabetes expenditure, burden of CT2. Eli Lilly National Audit Centre, 1993. Leicestershire practices: A comparative study. Pract disease and management in 5 EU countries. London 16. Farooqi A, et al. Does clinical audit improve care? Diabetes Int 1999;16:114–6. School of Economics, Jan 2012. Lessons for clinical governance from a district-wide 27. Millet C, et al. Ethnic disparities in diabetes man- 5. Tight blood pressure control and risk of macrovas- primary care audit diabetes. J Clin Gov 2000; agement and pay-for-performance in the UK: the cular and microvascular complications in type 2 dia- 8:152–6. Wandsworth Prospective Diabetes Study. PLoS Med betes: UKPDS 38. UK Prospective Diabetes Study 17. Khunti K, et al. Quality of care of patients with dia- 2007;4(6):e191. Group. BMJ 1998;317(7160):703–13. betes: collation of data from multi-practice audits 28. Diabetes Commissioning Toolkit. www.dh.gov.uk/ 6. Effect of intensive blood-glucose control with met- of diabetes in primary care. Fam Pract 1999: prod_consum_dh/groups/dh_digitalassets/@ formin on complications in overweight patients 16(1):54–9. dh/@en/documents/digitalasset/dh_4140285.pdf with type 2 diabetes (UKPDS 34). UK Prospective 18. Stevenson K, et al. Features of primary health care [accessed 13 June 2012]. Diabetes Study (UKPDS) Group. Lancet 1998; teams associated with successful quality improve- 29. The Health and Social Care Act 2012. www.dh. 352(9131):854–65. ment of diabetes care: a qualitative study. Fam Pract gov.uk/health/2012/06. 7. Trends in Consultation Rates in General Practice 2001;18(1):21–6. 30. Stone M, et al. Empowering patients with diabetes: 1995 to 2008: Analysis of the QResearch database. 19. The National Service Framework for Diabetes: a qualitative primary care study focusing on South The NHS Information Centre, 2009. Standards. London: Department of Health, 2001. Asians in Leicester, UK. Fam Pract 2005:22:647–52. 8. Malins J, Stuart J. Diabetic clinic in general practice. 20. The New GMS Contract; Investing in general 31. Farooqi A, Bhavsar M. Project Dil: a co-ordinated BMJ 1971;4:161. practice. 2003. www.nhsemployers.org/PayAnd Primary Care and Community Health Promotion 9. Thorn PA, Russel RC. Diabetic clinics today and Contracts/GeneralMedicalServicesContract/QOF/ Programme for reducing risk factors of coronary tomorrow: mini clinics in general practice. BMJ Pages/Rewarding-QOF.aspx. heart disease amongst the South Asian community 1973;2(5865):534–6. 21. National Institute for Health and Clinical Excellence. of Leicester – experiences and evaluation of the 10. Wilkes E. The diabetic, the hospital and primary Management of type 2 diabetes, CG66. NICE, 2008. project. Ethn Health 2001;6(3–4):265–70. care. J Royal Coll Gen Pract 1980;30:199–206. 22. Oluwatowoju I, et al. Improvements in glycaemic 32. Farooqi A, et al. Diabetes service provision in pri- 11. Hayes TM, Harries J. Randomised controlled trial of control and cholesterol concentrations associated mary care: a baseline survey in a city primary care routine hospital clinic care versus routine general with the Quality and Outcomes Framework: a trust (PCT). Pract Diabetes Int 2004;21:13–7. practice care for type II diabetics. BMJ (Clin Res Ed) regional 2-year audit of diabetes care in the UK. 33. Pieber TR, et al. Evaluation of a structured outpa- 1984;289(6447):728–30. Diabet Med 2010;27:354–59. tient group education program for intensive insulin 12. Singh, B et al. Metabolic control of diabetes in gen- 23. National Diabetes Audit. www.ic.nhs.uk/diabetes therapy. Diabetes Care 1995;18:625–30. eral practice clinics: comparison with a hospital audits [accessed 13 June 2012]. 34. Khunti K, et al. Effectiveness of a diabetes educa- clinic. BMJ (Clin Res Ed) 1984;289(6447):726–8. 24. The APHO model for diabetes prevalence. York and tion and self management programme (DESMOND) 13. Griffin S. Diabetes care in the general practice: Humber Public Health Observatory, 2011. for people with newly diagnosed type 2 diabetes meta-analysis of randomised control trials. [With 25. Griffin SJ, et al. Effect of early intensive multifactor- mellitus: three year follow-up of a cluster ran- commentary by T Greenhalgh.] BMJ 1998; ial therapy on 5-year cardiovascular outcomes in domised controlled trial in primary care. BMJ 2012; 317:390–6. individuals with type 2 diabetes detected by screen- 344:e2333.