Abstract

Background The incentivization of UK primary care through the Quality and Outcomes Framework (QOF) has released an unprecedented supply of data that in theory could aid health equity audit and reduce health inequalities. The current system allows for ‘exception reporting’ whereby patients can be excluded from calculation of payment for reasons such as failure to attend review. We speculated that such exclusions could be linked to socioeconomic deprivation.

Methods We assessed ‘exception reporting’ rates for 15 diabetes indicators using 2004/05 QOF data for 49 general practitioner (GP) practices in Brighton and Hove and related it to a deprivation ranking for each practice.

Results The standardized diabetes prevalence was 26% higher (P < 0.001) in the highest compared to the lowest quintile of deprivation. Correlations between ‘exception reporting’ and deprivation were seen for 10 of the 15 diabetes indicators (r = 0.20–0.41, P < 0.05). Practices with a more deprived patient population were more likely to report ‘exceptions’ for QOF indicators, although there was no such relationship with the achievement of QOF targets.

Conclusions Strategies to reduce health inequalities need to take into account that high levels of exception reporting, particularly in practices with deprived populations, may be disguising unmet need in those populations.

Introduction

If social and other inequalities in health are to be tackled, then appropriate audit tools need to be developed so as to inform the planning and commissioning of health service delivery. As part of the new General Medical Services (GMS) contract in the UK, general practitioners (GPs) are encouraged to use evidence-based interventions, particularly in the management of chronic diseases such as diabetes.1 An integral part of this contract is the Quality and Outcomes Framework (QOF) system. The QOF is in essence a payment system for GPs but is already showing potential as a rich new resource of primary care data, particularly for those interested in chronic diseases. One attraction of this database is the ability to link health information to other sources of information about the contributing practices, such as the local socioeconomic environment.2

The clinical domain of the QOF identifies 11 areas of chronic disease and subdivides them into 146 clinical indicators. A computer programme [Quality Management Analysis System (QMAS)] is used to extract data from patient record files. The QMAS calculates the number of patients diagnosed with a specific chronic condition (such as diabetes) and thus eligible for clinical indicator criteria (such as measurement of blood pressure or plasma cholesterol levels). The programme then awards ‘points’ by calculating the proportion of patients with the specified condition who meet these criteria. Final payment is adjusted for total number of patients registered by the practice as well as the prevalence of these diseases within the practice population.

There may be a valid reason why a patient does not meet the expected clinical indicator criterion, such as intolerance of or allergy to medication, dissent or not attending reviews after invitation. To ensure that in such cases the practices are not unfairly penalized, such patients may be excluded from the payment calculations—the process of ‘exception reporting’. There is some concern3 that this system will focus care on those patients who figure in targets whilst reducing the standard of care given to those ‘‘exception reported’ patients’, leading to an increase in health inequalities.

Diabetes mellitus affects more than 1.8 million people in the UK4 with an estimated further 25% still undiagnosed.5 Type 2 diabetes mellitus has been linked to socioeconomic deprivation: diabetic patients with a lower level of education are less likely to follow advice on lifestyle and medicines6 and less likely to attend their GP for a review of their condition.7,8 If this is so, then screening and disease management programmes should be targeted at such individuals. In the present study, we have combined QOF reporting data with socioeconomic deprivation indices to look at ‘exception reporting’ of patients from diabetes indicators in Brighton and Hove general practices.

Methods

Quality and Outcomes Framework (QOF) data

We examined QOF diabetes indicators using data from the first full year of collection (2004–05) from the 49 Brighton and Hove City Primary Care Trust (BHCPCT) general practices. These data were extracted from QMAS on 19 April 2005. Three practices had to be excluded from our study: two because of lack of electronic records and one because of an atypical population of exclusively homeless people. The total number of patients with diabetes mellitus is recorded as indicator 1 (DM1). For all subsequent indicators, the number of patients who achieved the indicator was recorded in the numerator and the number of patients who consented and were considered suitable for the treatment was recorded in the denominator. Table 1 lists a set of reasons for excluding patients from this calculation (exception reporting).

Table 1

Reasons for QOF ‘exception reporting’ of patients

Patients recorded as refusing to attend review who have been invited on at least three different occasions during the last 12 months 
Patients for whom it is not appropriate to review the chronic disease parameters because of particular circumstances, for example terminal illness and extreme frailty 
Patients newly diagnosed within the practice or who have recently registered with the practice, who should have measurements made within 3 months and delivery of clinical standards within 9 months 
Patients who are on maximum tolerated doses of medication whose levels remain sub-optimal 
Patients for whom prescribing a medication is not clinically appropriate, for example those who have an allergy 
Where a patient has not tolerated medication 
Where a patient does not agree to investigation or treatment and this has been recorded in their medical records 
Where the patient has a supervening condition which makes treatment of their condition inappropriate 
Where an investigative service or secondary care service is unavailable 
Patients recorded as refusing to attend review who have been invited on at least three different occasions during the last 12 months 
Patients for whom it is not appropriate to review the chronic disease parameters because of particular circumstances, for example terminal illness and extreme frailty 
Patients newly diagnosed within the practice or who have recently registered with the practice, who should have measurements made within 3 months and delivery of clinical standards within 9 months 
Patients who are on maximum tolerated doses of medication whose levels remain sub-optimal 
Patients for whom prescribing a medication is not clinically appropriate, for example those who have an allergy 
Where a patient has not tolerated medication 
Where a patient does not agree to investigation or treatment and this has been recorded in their medical records 
Where the patient has a supervening condition which makes treatment of their condition inappropriate 
Where an investigative service or secondary care service is unavailable 

A proxy deprivation ranking was developed for each practice, using the revised 2004 super output area level Index of Multiple Deprivation (IMD).9 IMD ratings for practice populations were calculated proportionately according to number of patients from their list represented in each super output area. Practices were ranked according to their population IMD score ranging from 14 (least deprived) to 52 (most deprived).

Because QOF data do not distinguish between diabetes mellitus types 1 and 2, we calculated the age- and sex-standardized prevalence of the combined conditions between practices using the indirect method. The expected prevalence of types 1 and 2 diabetes for different sex and age bands was then calculated using Diabetes UK data,10 and the practice age distribution was obtained from the Exeter system. Finally, the standardized diabetes prevalence was divided into quintiles according to the practice population deprivation score.

Two of the 17 performance indicators (DM4 ‘smoking cessation advice’ and DM15 ‘appropriate use of angiotensinogen-converting enzyme inhibitors’) were excluded from our analysis as they refer to specific subset populations of the disease register, which were not defined elsewhere, making it impossible to determine ‘exception reporting’ rates. For the remaining 15 indicators, ‘exception reporting’ was calculated as a percentage based on the difference between the number of patients considered eligible for the indicator (given as the denominator for each indicator) and the number of diabetics registered within the practice (DM1). Target achievement was calculated as percentage of diabetes patients deemed eligible who had reached the indicator criteria.

Statistical calculations were performed using SPSS version 13, except for the test of outliers, which was performed using Grubb’s test (GraphPad Software online). Data on achievement and ‘exception reporting’ did not conform to a normal distribution and so were analysed using non-parametric tests (Spearman’s correlation coefficient). The age- and sex-standardized diabetes prevalence in the most deprived compared to the least deprived areas was compared using the Kruskal–Wallis test.

Results

Of the 283 437 patients registered within the 49 BHCPCT practices, 7157 patients had been diagnosed with either type 1 or type 2 diabetes mellitus, giving a pooled prevalence of 2.5%. The median practice list size was 5602 (range 14 194, n = 46) and disease register size 141 (range 419, n = 46). There were no significant differences in practice list or disease register size between quintiles of socioeconomic deprivation. The mean age- and sex-standardized diabetes prevalence was 95% (95% confidence interval 87.3–102.0). The standardized diabetes prevalence was significantly different between different quintiles of deprivation, with a 26% higher prevalence of diabetes in the most deprived compared to the least deprived practice populations (P < 0.001).

The extent of ‘exception reporting’ correlated (P < 0.05) with deprivation in 10 of the 15 indicators we were able to analyse (Table 2), with Spearman’s correlation coefficients ranging from 0.20 to 0.41. Correlations between deprivation and the other five indicators were weaker but in the same direction (Spearman’s correlation coefficients ranging from 0.19 to 0.28). Deprivation accounted for 9–16% of ‘exception reporting’. In contrast, there was no significant correlation between achievement for any of the 17 individual diabetes indicators and deprivation (Spearman’s correlation coefficients ranging from –0.27 to 0.17, P = 0.07–0.96).

Table 2

Relationship between ‘exception reporting’ for selected QOF diabetes indicators and practice population deprivation

QOF indicator details  ‘Exception reporting’ (%) according to quintile of deprivation (mean ± SEM)
 
 Correlation between ‘exception reporting’ and deprivation
 
 
  Least deprived Most deprived r P 
Percentage of all patients with diabetes mellitus (DM1)...      
DM2 ...whose notes record BMI in the last 15 months 1.8 ± 0.7 4.1 ± 1.0 0.26 0.080 
DM3 ...in whom there is a record of smoking status in previous 15 months 1.0 ± 0.4 2.3 ± 0.7 0.28 0.064 
DM5 ...who have a record of HbA1c or equivalent in previous 15 months 2.2 ± 0.8 5.6 ± 1.4 0.28 0.056 
DM6 ...in whom the last HbA1c was 7.4 or less in previous 15 months 17.5 ± 3.1 9.7 ± 2.7 0.32 0.031 
DM7 ...in whom the last HbA1c was 10 or less in previous 15 months 11.2 ± 1.5 5.1 ± 1.3 0.35 0.018 
DM8 ...who have a record of retinal screening in previous 15 months 2.7 ± 0.6 6.6 ± 1.5 0.41 0.005 
DM9 ...with a record of the presence or absence of peripheral pulses in previous 15 months 3.2 ± 0.8 9.0 ± 2.4 0.35 0.019 
DM10 ...with a record of neuropathy testing in previous 15 months 3.7 ± 1.2 9.1 ± 2.2 0.35 0.018 
DM11 ...who have a record of blood pressure in previous 15 months 0.8 ± 0.4 2.4 ± 0.6 0.31 0.037 
DM12 ...in whom the last blood pressure was 145/85 or less in previous 15 months 9.1 ± 1.8 5.4 ± 1.0 0.20 0.177 
DM13 ...who have a record of micro-albuminuria testing in previous 15 months 7.1 ± 1.9 11.3 ± 3.2 0.19 0.210 
DM14 ...who have a record of serum-creatinine testing in previous 15 months 2.0 ± 0.8 5.8 ± 1.7 0.30 0.045 
DM16 ...who have a record of total cholesterol in previous 15 months 2.1 ± 0.8 5.6 ± 1.3 0.30 0.045 
DM17 ...whose last measured total cholesterol was 5 or less in previous 15 months 15.0 ± 2.8 7.6 ± 1.9 0.35 0.016 
DM18 ...who have had influenza immunization in the preceding 1 September–31 March 10.3 ± 2.7 16.8 ± 3.0 0.35 0.017 
QOF indicator details  ‘Exception reporting’ (%) according to quintile of deprivation (mean ± SEM)
 
 Correlation between ‘exception reporting’ and deprivation
 
 
  Least deprived Most deprived r P 
Percentage of all patients with diabetes mellitus (DM1)...      
DM2 ...whose notes record BMI in the last 15 months 1.8 ± 0.7 4.1 ± 1.0 0.26 0.080 
DM3 ...in whom there is a record of smoking status in previous 15 months 1.0 ± 0.4 2.3 ± 0.7 0.28 0.064 
DM5 ...who have a record of HbA1c or equivalent in previous 15 months 2.2 ± 0.8 5.6 ± 1.4 0.28 0.056 
DM6 ...in whom the last HbA1c was 7.4 or less in previous 15 months 17.5 ± 3.1 9.7 ± 2.7 0.32 0.031 
DM7 ...in whom the last HbA1c was 10 or less in previous 15 months 11.2 ± 1.5 5.1 ± 1.3 0.35 0.018 
DM8 ...who have a record of retinal screening in previous 15 months 2.7 ± 0.6 6.6 ± 1.5 0.41 0.005 
DM9 ...with a record of the presence or absence of peripheral pulses in previous 15 months 3.2 ± 0.8 9.0 ± 2.4 0.35 0.019 
DM10 ...with a record of neuropathy testing in previous 15 months 3.7 ± 1.2 9.1 ± 2.2 0.35 0.018 
DM11 ...who have a record of blood pressure in previous 15 months 0.8 ± 0.4 2.4 ± 0.6 0.31 0.037 
DM12 ...in whom the last blood pressure was 145/85 or less in previous 15 months 9.1 ± 1.8 5.4 ± 1.0 0.20 0.177 
DM13 ...who have a record of micro-albuminuria testing in previous 15 months 7.1 ± 1.9 11.3 ± 3.2 0.19 0.210 
DM14 ...who have a record of serum-creatinine testing in previous 15 months 2.0 ± 0.8 5.8 ± 1.7 0.30 0.045 
DM16 ...who have a record of total cholesterol in previous 15 months 2.1 ± 0.8 5.6 ± 1.3 0.30 0.045 
DM17 ...whose last measured total cholesterol was 5 or less in previous 15 months 15.0 ± 2.8 7.6 ± 1.9 0.35 0.016 
DM18 ...who have had influenza immunization in the preceding 1 September–31 March 10.3 ± 2.7 16.8 ± 3.0 0.35 0.017 

Boldface values distinguish between mean and SEM values.

Discussion

Main findings of this study

We found the QOF to be a potentially valuable data source for equity audit. In this preliminary study, QOF data for diabetes were used to calculate the total prevalence of types 1 and 2 diabetes mellitus in Brighton and Hove (2.5%) and to identify a 26% higher prevalence of age- and sex-standardized diabetes in the most deprived compared to the least deprived populations. This estimate of the combined prevalence of types 1 and 2 diabetes in Brighton and Hove is 24% lower than that seen nationally,11 perhaps reflecting a higher proportion of young people in Brighton and Hove and a low representation of people from diabetes-susceptible ethnic groups, such as Afro-Caribbeans and South Asians.10 Data from the National Diabetes Audit5 suggest that 23% of persons with diabetes mellitus are still undiagnosed in the UK, suggesting that diabetes is under-diagnosed in Brighton and Hove and poses a substantial health problem for the city.

Our most striking finding is the correlation between ‘exception reporting’ of diabetic patients and deprivation, with highest rates in the most deprived populations. In contrast, we found no evidence of socioeconomic variation in the achievement of diabetes management targets, on which QOF resource allocation is based. If confirmed, this finding argues for an analysis of ‘exception reporting’, and not simply target achievement, when investigating health inequity.

What is already known on this topic

Although the QOF system did not distinguish between diabetes types 1 and 2, the latter accounts for 87% of all cases10 and so our data are consistent with previous reports of a higher prevalence of type 2 diabetes in more deprived areas.12–14 In addition, deprivation and lower socioeconomic status have been associated with non-attendance at annual reviews, low use of preventive care and screening.7 A previous study looking at the achievement of Coronary Heart Disease (CHD) indicators and socioeconomic deprivation did not demonstrate any inequalities in care2 but failed to take into account exception reporting rates.

What this study adds

Our findings show that people with diabetes living in deprived areas are more likely to be ‘exception reported’ from the QOF clinical indicators. One obvious explanation is that such patients are less likely to attend preventive health care reviews and less likely to comply with health advice. Informal comments from practice nurses do suggest that non-attendance is the main reason for ‘exception reporting’ of diabetes patients, which would be consistent with previous reports. A local diabetes survey found that 53% of practices did not conduct full annual diabetes reviews on patients who were housebound or in residential homes, in some cases ‘exception reporting’ these patients (L. Sigfrid, unpublished results).

Importantly, the level of achievement or targets was similar between practices with populations of different deprivation levels, meaning that they receive equal resource allocation regardless of variations in ‘exception reporting’. This raises the possibility that there is an unmet need in those practices with more deprived populations that may not be captured if resource allocation is based solely on achievement of targets. This highlights the importance of analysing ‘exception reporting’ as well as target achievement when using QOF data to allocate health care resources. Our study suggests that more work is needed to detect diabetes, prevent complications and target hard-to-reach populations. Information on individual exception codes will be made available in 2005/06, enabling detailed analysis of reasons for ‘exception reporting’. PCTs will be able to use this information to identify practices in need of greater assistance and guide the development of local enhancement schemes and outreach initiatives to improve services for hard-to-reach populations.

Limitations of this study

The QOF data collection system was not designed as a research resource. Our inability to obtain information on ethnicity and the nature of diabetes (type 1 versus 2) limited the information we could extract from the data. Furthermore, the data contained in this article are derived from the first full year of QOF data, and there may well be issues of data quality and inadequacy of technology. The QOF is primarily a voluntary payments system, and individual practices may opt into all, some or none of the system. They may also choose to focus on areas not prioritized in the QOF. For example, one local practice deals exclusively with homeless patients whose health care priorities are not recognized as fitting into the national framework. PCTs need to recognize the limitations of the QOF but also the potential benefit in highlighting areas for further enquiry for implementation groups working on chronic disease management.

The QOF, designed to evolve with changes in health care priorities, is a positive step forward that provides an opportunity for developing equity audits and strategic service provision. Future data on individual exception codes will enable studies of socioeconomic variations in more detail and enable PCTs to target resources and interventions to improve quality of care and address health inequalities.

Competing interests

All authors declare that they have no competing interests.

Ethical approval

None needed.

Acknowledgements

We thank Matthew Hankins, BSc, FRSS, Surrey and Sussex Integrated Research Networks, Division of Primary Care and Public Health, Brighton, and Sussex Medical School, Brighton, for statistical advice. Chris Dorling (Information Specialist, BHCPCT) kindly provided the practice population deprivation scores. We are grateful to Dr Tom Scanlon and Dr Peter Wilkinson for their comments on the manuscript.

Funding
 None.

References

1
Department of Health.
NHS Patients and Staff Reap Rewards of GP Incentive Scheme as World’s Biggest Health Database Goes Live
 . London: Department of Health,
2005
. http://www.dh.gov.uk/PublicationsAndStatistics/PressReleases/PressReleasesNotices/fs/en?CONTENT_ID=4118533& chk=q8TvGU (5 May 2006, date last accessed).
2
Strong
M
, Maheswaran R, Radford J. Socioeconomic deprivation, coronary heart disease prevalence and quality of care: a practice-level analysis in Rotherham using data from the new UK general practitioner Quality and Outcomes Framework.
J Public Health (Oxf)
 
2006
;
28
:
39
–42.
3
NHS Alliance.
The Future of Access to General Practice-Based Primary Medical Care, Informing the Debate
 . Royal College of General Practitioners,
2004
.
4
Department of Health: Diabetes Team.
Improving Diabetes Services – The NSF Two Years On
 . Department of Health,
2005
. http://www.dh.gov.uk/assetRoot/04/10/67/20/04106720.pdf (5 May 2006, date last accessed).
5
National Diabetes Audit.
Key Findings About the Quality of Care for People with Diabetes in UK
 . NHS Health and Social Care Information Centre,
2005
. http://www.icservices.nhs.uk/ncasp/pages/audit_topics/diabetes/LSD05085NDAReport.pdf (5 May 2006, date last accessed).
6
Bachmann
MO
, Eachus J, Hopper CD et al. Socio-economic inequalities in diabetes complications, control, attitudes and health service use: a cross-sectional study.
Diabet Med
 
2003
;
20
:
921
–9.
7
Goddard
M
, Smith P. Equity of access to health care services: theory and evidence from the UK.
Soc Sci Med
 
2001
;
53
:
1149
–62.
8
Goyder
EC
, McNally PG, Botha JL. Inequalities in access to diabetes care: evidence from a historical cohort study.
Qual Health Care
 
2000
;
9
:
85
–9.
9
Noble
M
, Wright G, Dibben C et al.
The English Indices of Deprivation 2004 (revised)
 . Office of the Deputy Prime Minister,
2005
. http://www.odpm.gov.uk/pub/446/Indicesofdeprivation2004revisedPDF 2198Kb_id1128446.pdf (9 May 2006, date last accessed).
10
Diabetes
UK.
Diabetes in the UK 2004
 . Diabetes UK,
2004
. http://www.freepint.com/issues/160206.htm (5 May 2006, date last accessed).
11
NHS Health and Social Care Information Centre.
National Quality and Outcomes Framework Statistics for England 2004/2005
 . NHS Health and Social Care Information Centre,
2005
. http://www.ic.nhs.uk/services/qof/statisticalbulletin/Bulletin.PDF (5 May 2006, date last accessed).
12
Meadows
P.
Variation of diabetes mellitus prevalence in general practice and its relation to deprivation.
Diabet Med
 
1995
;
12
:
696
–700.
13
Evans
JM
, Newton RW, Ruta DA et al. Socio-economic status, obesity and prevalence of Type 1 and Type 2 diabetes mellitus.
Diabet Med
 
2000
;
17
:
478
–80.
14
Whitford
Griffin SJ
, Prevost AT. Influences on the variation in prevalence of type 2 diabetes between general practices: practice, patient or socioeconomic factors?
Br J General Pract
 
2003
;
53
:
9
–14.