OUP user menu

The novel use of commonly captured data to assess a district's diabetes service that encompasses both primary and secondary care

E.C. Russell-Jones , A. Gough , S. Lawrence , I.N. Scobie
DOI: http://dx.doi.org/10.1093/qjmed/hct097 737-745 First published online: 26 April 2013


Aim: To identify commonly captured data in the UK to look at the performance of a district’s diabetes care that encompasses both primary and secondary care.

Methods: Primary care quality outcomes framework (QOF) measures for diabetes, referral rates for first appointment for specialist secondary care and emergency admission rates for diabetes (Dr Foster/HES) were used to produce a performance index scoring system. Illustrative measures from QOF were total diabetes points, DM23 attainment of HbA1c <7% (53 mmol/mol) and its exemption rate (number of patients excluded from analysis).

The performance index was used to study the effectiveness of the Medway district diabetes service and this was compared to another district (Guildford) within the same Strategic Health Authority and nationally.

Results: Medway has the highest prevalence of Diabetes (6.1%) of the 8 Primary Care Trusts examined, the lowest achievement of diabetes QOF points (96.1%) and the lowest achievement of an HbA1c level <7% (53 mmol/mol) (54.3%). Exemption reporting was the 3rd highest. SAR for first diabetes out-patient appointment to the hospital was low at 281 (predicted 576) 48% of expected. The emergency admission rate was high at 225 (predicted 168) 133% of expected. Thus primary care diabetes needs to raise performance and implement a lower threshold for OPD referral to prevent emergency admissions.

Conclusion: It is possible to produce an assessment of diabetes care that transcends primary/secondary care that gives a true reflection of a district’s performance which will be useful to plan future health service provision.


There is much interest and concern about the rising prevalence of diabetes worldwide and the resultant chronic disease burden and economic consequences. It is estimated that there will be 300 million people worldwide with diabetes by the year 2025.1 These figures have recently been revised upwards with recent data showing that in 2008 there were already 347 million people with diabetes.2 The vast majority of patients have uncomplicated type 2 diabetes and primary care is being encouraged to look after patients with this diagnosis.

The Quality Outcomes Framework (QOF) was introduced in 2004 which gave financial incentives for the achievement of certain standards of diabetes care.3 There has been a range of studies that have shown benefit from this approach,4 with improvements in glycaemic control and cholesterol management and a reduction in variation and inequality of diabetes care across the UK.5 Secondary care physicians have been worried that with the emphasis on primary care, coordination of services has declined.6 Some have also suggested that the association between QOF scores and emergency admissions and mortality were small and inconsistent, while the impact of socio-economic deprivation was much higher.7

In today’s society health care funding is under considerable pressure. There is thus a need for quality of treatment and care to be constantly evaluated on a beneficence/cost-effective scale. There is a need to assess performance of diabetes care that encompasses both primary and secondary care. None of the quoted schemes do this.8,9 QOF looks at primary care while national hospital diabetes audits and DUK in-patient audits look at secondary care. Data are however being routinely collected and are freely available. This would allow a new index to be created that looks at both primary and secondary care within a district or region. The routinely collected data include QOF data for diabetes, out-patient referral rates for first appointment or new diabetes appointment to secondary care, district diabetes prevalence, district deprivation scores and emergency admission rates for diabetes.

The aim of this study was to use these freely available data to look at the performance of a district’s diabetes care that encompasses primary and elements of secondary care to provide guidance to commissioners and healthcare professionals to target precious resources.


The district selected was the Medway towns and Swale located in Kent, South-East England ∼30 miles from London. Medway PCT and Swale has a population of 360 000 and encompasses both rural and urban communities.

In order to determine the quality of diabetes care in the Medway and Swale district a list of all the General Practices was made. General practices included the Isle of Sheppey, Hoo Peninsula and Rochester as well as those in the local vicinity to the Medway hospital. This helped established the boundaries of the Medway and Swale cluster of diabetes care.

We identified freely accessible data that were being collected in primary or secondary care. Data included primary care QOFmeasures for diabetes, referral rates for first appointment for specialist secondary care and emergency admission rates for diabetes from Hospital Episode Statistics (HES) and Dr Foster data. These data were used to create a new performance index scoring system.

The performance index

Primary care QOF measures for diabetes

QOF data from all the GP surgeries were obtained. QOF is a voluntary incentive scheme for general practitioners in the UK to reward them for how well they care for patients. QOF data for diabetes consists of a series of indicators.13 For the performance index illustrative measures from QOF were total diabetes points and DM23 attainment of HbA1c <7% (53 mmol/mol). Practices are also allowed to exempt particular patients if they fail to engage to improve their condition despite help and medical input, or patients who are unable to reach the required target due to other acceptable medical reasons. It was therefore also important to look at the exemption rate (number of patients excluded from analysis) as well as the achievement rate for any of the named indicators.

The out-patient referral rate for ‘new’ or first time visit to diabetes secondary care and the emergency admissions data for diabetes

Standardized Admission Rate (SAR) for new or first time visit to specialist diabetes services and SAR for emergency diabetes admissions were used. The data were obtained from HES and Dr Foster data and are available at practice level.

HES are a collection of data that details all admissions and out-patient appointments in the UK.10 Many have concerns about the accuracy and validity of the HES hospital activity data due to the poor quality of the information provided and mis-coding.

Dr Foster data comprise information gained for hospital trusts and Primary Care Trusts (PCTs) and are based around activity that was invoiced and paid for by PCTs.11


It was important to gage an idea of deprivation as this can influence the quality of diabetes care and ease of access to care. The Diabetes Health Intelligence, a branch of the National NHS Diabetes information service was used to explore deprivation.12

The use of the performance index scoring system in a different district

The performance index was piloted in a district diabetes service at Medway and Swale and then compared to other PCTs within the Strategic Health Authority.

The Guildford area in Surrey was also chosen as a comparative district as it also resides within the same SHA as Medway. The prevalence of diabetes is much lower and the area more affluent. A number of initiatives have taken place over the last 10 years to improve care in primary care, with a regular shared care educational meeting (cited by DOH as an example of best practice), over 200 General practitioners and practice nurses having completed the Warwick diabetes diploma and a number of incentive schemes run by the Primary Care Groups to get patients out of hospital clinics.


Medway prevalence and deprivation

There is a much higher prevalence of diabetes within the Medway and Swale area relative to both other areas of the Strategic Health Authority and nationally (Figure 1). Medway had the highest prevalence within the 8 different PCTs that make up the Strategic Health Authority and had the 27th highest prevalence out of 152 PCTs in the country (Table 1). The higher prevalence is likely to be due to the demographic make-up of the population with higher deprivation and pockets. The Diabetes Health Intelligence has highlighted that deprivation scores are higher in Medway than those you would expect using the country as a whole (Figure 2). Thus roughly 45% of the population at Medway has a deprivation score of 2 or below. Looking into demographics of Kent and their surroundings also shows that there is a high level of unemployment and the average wage is significantly lower than in other parts of the SHA.

Figure 1.

A graph to show the diabetes prevalence for the Medway cluster in relation to the SHA average and national average.

Figure 2.

A graph to show the deprivation in Medway relative to the rest of England [12].

View this table:
Table 1

Diabetes QOF data from 2009 to 2010 for Medway cluster in comparison to PCT practice average within the SHA (KENT) and the national average

Medway QOF

Figure 3 and Table 1 show that the performance of the Medway practices is slightly below the SHA average but comparable to the national average and this trend appears in both years studied. Medway came 5th out of the 8 different PCTs within the SHA and was 66th out of 152 nationally.

Figure 3.

Total QOF points for diabetes in the Medway area in comparison to the SHA and national averages.

The most important individual QOF data were attainment of good HbA1c values and complication surveillance of important core complications (DM 5, 6, 7, 9, 10, 13, 16) as examples. (Figures 4 and 5; Tables 2 and 3)

Figure 4.

The performance of glycaemic control based on QOF indicators for the Medway cluster relative to the SHA and national average.

Figure 5.

The exception values for the performance of glycaemic control based on QOF indicators.

View this table:
Table 2

The performance of gylcaemic control based on QOF indicators for the Medway cluster relative to the SHA and national average

View this table:
Table 3

The exception values for the performance of gylcaemic control based on QOF indicators

Although Medway was only just behind the SHA average and 4th out of 8 at achieving the most difficult HbA1c threshold of <7% (53 mmol/mol), the exception rates were considerably higher (Table 3). Medway is ranked low in comparison to the National PCT ranking (138th out of 152 for measuring HbA1c) and is 7th or 8th out of 8 in the PCT ranking within the SHA (Table 2). This indicates a poor performance considering Medway PCT has the highest prevalence of diabetes.

In the complications panel Medway performed relatively poorly compared to the SHA average being between 5th and 8th out of the 8 PCTs within the SHA apart from checking the peripheral pulses (DM 9) where it appeared to be average coming 4th (Table 4). This shows that Medway is consistently in the bottom half of the PCTs within the SHA and is below average in comparison to that of the PCTS across the UK. In particular NHS Medway appears to be deficient in testing the serum creatinine (DM 22) and calculating the eGFR in patients with diabetes, such that Medway is ranked 135th out of 152 (poor achievement) and has an exception rate of 14th out of 152 (exceptionally high, Table 5).

View this table:
Table 4

The management of potential complications based on QOF indicators for the Medway cluster relative to the SHA and national average

View this table:
Table 5

The exception values for the management of potential complications based on QOF indicators

Medway out-patient referral rates

The hospital performance data are very interesting. It shows that the SAR for new OPD diabetes referrals is a good indicator of how primary care physicians are seeking specialist diabetes input from secondary care for the more complex and problematic patients. The data show that the referral rate from primary care to the hospital specialist services is lower than the national average. This is surprising considering the high prevalence and relatively poor attainment of QOF quality points for diabetes. (Table 6)

View this table:
Table 6

Hospital first or new diabetes out-patient referral rate measured as a SAR

Table 6 shows that the SAR for new/first appointment to the hospital diabetes out-patients is 281. This compares to 576 which is the expected value based on the population, deprivation and prevalence of the area. This means that fewer than 50% of patients that one would expect are actually being referred to secondary care for out-patient hospital specialist opinion. With the high prevalence and the slightly low QOF quality points for diabetes one might have expected a higher referral rate as primary care teams seek specialist input. This however did not appear to be the case.

Medway emergency diabetes admission rate

The Hospital emergency admission data are information on the number of patients with diabetes with and without complications that come to hospital as an emergency for any aspect of their diabetes. (Table 7)

View this table:
Table 7

Hospital emergency admission rate for diabetes measured as a SAR

The SAR data for emergency admission for any aspect associated with diabetes at Medway is considerably higher than that of the national average. The SAR for emergency admissions observed (225) is considerably higher than expected (168) hence the emergency admissions for diabetes and complications associated with diabetes is 133% of what is expected. (Table 7)

Practice-level data

Both QOF and Dr Foster/HES data are available at practice level and there was considerable variation in performance. Thus practices can be identified that are performing less well and targeted for special input.10,13,14

The use of the scoring system in a different district

The performance index was effectively used to indicate diabetes performance in another district. The prevalence of diabetes in Guildford in Surrey is relatively low at 4%. The deprivation scores show that the majority of the population (over 60%) fall into the ‘least deprived’ bracket. Total Diabetes QOF points were 99.3%, (much higher than the SHA average), 61.1% achieved an HbA1c <7% (53 mmol/mol). This was much higher than the SHA average, and a lower exception reporting rate of 9.8% was found compared to the SHA average. The SAR for referral to diabetes out-patients was low at 87.8% (Surrey average 128.9%) and the diabetes emergency admission rate was 85% compared to the Surrey average of 106.7%. Thus this performance index scoring scheme has provided robust data in a different district. It suggests that prevalence is lower than in the Medway area, primary care performance is relatively good, referral to secondary care is low and emergency admission rate is low providing contrast to Medway. New resources can therefore be focussed at other health care problem areas within the Guildford district.


Use of the described performance index makes it possible to produce an assessment of diabetes care that transcends primary/secondary care and gives a true reflection of a district’s clinical and cost-effective performance which will be useful to plan future health service provision.6 This new scoring system makes it is possible to see and construct priorities and strategic goals to improve care.

Diabetes is a pandemic and is a major health problem for the 21st century. Diabetes and its complications will take up a large proportion of the NHS budget in the future and the UK government along with all similar governments in the developed world are trying to create successful strategies to cope.1 The UK is the first country that has incentivized primary care physicians to improve diabetes care in the community with the implementation of the QOF indicators.3

To understand how well a district is performing to date, there have not been any satisfactory tools, although regional performance can be gauged by national audits. Despite this, data are routinely collected and are freely available evaluating diabetes care for both primary and secondary care.10,13,14 We have shown how the performance index can be used to produce robust information that can be compared both regionally and nationally and used to focus health service provision. This index is simple, some might argue simplistic but has many advantages as it can be performed simply and quickly by anyone to gauge the cost-effective performance of a district. The data from Medway show that the district has a high prevalence of diabetes, a relatively low attainment of QOF diabetes points and a fairly high exception reporting rate. The referral rate for new diabetes referrals to specialist OPD is low in comparison to the national average and there is a considerably higher emergency admission rate.

The reason for the low OPD referral rate is unknown; however, there was a triage system introduced in Medway because of the large number of referrals to secondary care that were deemed to be inappropriate. This scheme ran from 2003-07 with a GP with special interest. It is unlikely that GPs are over confident in their abilities and are choosing to manage the majority themselves. There is no formal diabetes training programme for primary care teams in the Medway area, although some have completed the Warwick diploma and an ad hoc training programme has been organized by a GP with special interest. In Guildford/Surrey area there was a formal programme that was recommended for practices to be eligible for additional ‘enhanced level’ payment, with over 200 GPs and practice nurses completing the rigorous Warwick diploma.

The prevalence of diabetes within Medway was high and almost double that of other areas within the SHA. This means that diabetes should be even more of a priority in the Medway area, and more financial and manpower resources should be allocated to this chronic condition. Primary care diabetes is struggling to meet quality diabetes care and needs resources and support to raise performance and to identify more complex patients that should be referred to specialist hospital clinics. This in turn should also prevent/minimize emergency admissions.

This same scoring system was used to evaluate the performance of a very different district (Guildford). It performed well and provided very different answers and messages to commissioners.

Measuring the total cost of diabetes care is difficult as there are many direct and indirect costs and prevalence of co-morbidities skews results.15 Thus measuring total costs accurately are extremely difficult within the NHS. Prevalence rates between the two districts were very different making it difficult to compare, on top of this national in-patient audits pick up large numbers of patients with diabetes who are in hospital for other reasons but their numbers are much greater than one would expect from simple population prevalence. Nonetheless using these data sources estimates of cost can be made and both have been estimated to be around 2.8 million per 100 000 population but with very different cost profile (http://webarchive.nationalarchives.gov.uk/+/www.dh.gov.uk/en/Managingyourorganisation/Financeandplanning/Programmebudgeting/DH_075743#_7). Using these estimates costs have also been related to outcomes (HbA1c <7) and compared across the whole of England. Both Medway and Guildford/Surrey on quadrant analysis fall in the low cost with good outcomes quadrant but with Guildford/Surrey having a relative better outcome.12 The total cost was not the prime objective of this current article.

Each study would not be complete without acknowledgement of its limitations. The QOF data used produces some of their performance figures using an average of averages, which unless the practices are all the same size is mathematically flawed.13 In Medway there is almost a 10-fold difference in practice size, however having looked at the individual practice data sets from both areas, the overall message remains the same. Although practice-level data may be available the ability to use such a system in the future to identify individual patients who are admitted frequently or fail to attend OPD appointments may be useful to target intensive help and support.

Understanding how well a district is performing is going to be essential for the future of diabetes care. The PCT or GP commissioners groups will hold the purse strings to commission services. If commissioners do not understand their local diabetes performance they cannot prioritize services effectively. Clinicians who understand the limitations of this simple scoring system will still find it helpful in making informed commissioning decisions.


Combining QOF data, HES and Dr Foster it has been possible to illustrate the performance of a district that transcends primary and secondary care. These data are very easy to access and are freely available. We believe this is the first description of such an approach in the UK and hopefully this may be taken forward as a simple, feasible and practical way for many districts to assess their performance to set priorities for improvement.

Novelty Statement: Commissioners have to constantly improve performance while lowering the cost of health care delivery. This study aims to provide a simple scoring system using freely available and routinely collected data that can be used to assess a district's performance in delivering diabetes care. Combining QOF data, HES and Dr Foster it has been possible to illustrate the performance of a district that transcends primary and secondary care. This simple scoring system is applicable to any region within the UK, and will prove invaluable in identifying priorities for diabetes care, in an environment where financial resources are shrinking.

Conflict of interest: None declared.


View Abstract