Project objectives

1. Generate disease denominators by identifying eligible patients for each QOF condition. 
2. Map patient-level performance on all clinical indicators (89 as of 2009/10).
3. Estimate the effect of indicator removal from the QOF.
4. Estimate the ‘met’ and ‘unmet’ exception reporting levels across the all QOF domains.
5. Examine the profiles of excluded and non-excluded patients across all QOF domains (especially in terms of multimorbidity) and use multilevel logistic regression analysis to identify differences between the two groups in terms of age, sex, co-morbidities and service utilisation.
6. Examine the timing of exceptions in each financial year and assess whether they have been used appropriately.
7. Compare overall levels of care between newly diagnosed patients and patients diagnosed in previous years, before and after the introduction of the scheme.
8. Investigate the effect of blood-pressure, HbA1c and total cholesterol levels on all cause mortality, diabetes related mortality and diabetes-related complications adjusting for patient characteristics and treatments.
9. Replicate the UKPDS results of interest and perform sub-analyses on patients aged 65 or over.
10. Investigate the effect of the QOF targets on patient outcomes.
11. Estimate how the gains from diabetes related QOF intervention vary across patient groups.

Brief summary

Aim 1: Withdrawing incentives

Objectives

To investigate the effect of withdrawing incentives on recorded quality of care, in the context of the UK Quality and Outcomes Framework pay-for-performance scheme.

Design

Retrospective longitudinal study.

Setting

Data for 644 general practices, from 2004/5 to 2011/12, extracted from the Clinical Practice Research Datalink.

Participants

All patients registered with any of the practices over the study period, 13,772,992 in total.

Intervention

The removal of financial incentives for aspects of care for patients with asthma, coronary heart disease, diabetes, stroke and psychosis.

Main outcome measures

Performance on eight clinical quality indicators withdrawn from a national incentive scheme: influenza immunisation (asthma) and lithium therapy monitoring (psychosis), removed in April 2006; blood pressure monitoring (coronary heart disease, diabetes, stroke), cholesterol level monitoring (coronary heart disease, diabetes) and blood glucose monitoring (diabetes), removed in April 2011. Multilevel mixed effects multiple linear regression models were used to quantify the effect of incentive withdrawal.

Results

Mean levels of performance were generally stable after the removal of the incentives, both short- and long-term. For the two indicators removed in April 2006, levels in 2011/12 were very close to 2005/6 levels, although a small but statistically significant drop was estimated for influenza immunisation. For five of the six indicators withdrawn from April 2011, there was no significant impact on performance following removal and differences between predicted and observed scores were small. Performance on related outcome indicators retained in the scheme (e.g. blood pressure control) was generally unaffected.

Conclusions

Following the removal of incentives, levels of performance across a range of clinical activities generally remained stable. This indicates that health benefits from incentive schemes can potentially be increased by periodically replacing existing indicators with new indicators relating to alternative aspects of care. However, all aspects of care we investigated remained indirectly or partly incentivised in other indicators and further work is required to assess the generalisability of the findings when incentives are fully withdrawn.

Aim 2: Clinical outcome targets for diabetes

Objectives

Diabetes is estimated to affect nearly 300 million people globally and is a contributory factor in over 10% of adult deaths. However, consensus has not been reached on optimal levels of control for biological parameters. We aimed to describe the shape of observed relationships between risk factor levels and clinically important outcomes in type 2 diabetes, after adjusting for multiple confounders.

Design

Retrospective cohort study.

Setting

600 practices contributing to the Clinical Practice Research Datalink between 2006 and 2012.

Participants:

246,544 adult patients with type 2 diabetes.

Main outcome measures

Proportional hazards regression models quantified the risks of mortality, microvascular or macrovascular events associated with four modifiable risk factors: glycated haemoglobin (HbA1c), systolic and diastolic blood pressure and total cholesterol, while controlling for numerous patient and practice covariates.

Results

We observed U-shaped relationships between all-cause mortality and levels of the four risk factors. The lowest risks were associated with HbA1c 7.25-7.75%; total cholesterol 3.5-4.5 mmol/L; systolic blood pressure 135-145 mmHg; and diastolic blood pressure 82.5-87.5 mmHg. Coronary and stroke mortality related to the four risk factors in a positive, curvilinear way, with the exception of blood pressure which related to coronary deaths in a U-shape. Macrovascular events showed a positive and curvilinear relationship with HbA1c but U-shaped for total cholesterol and systolic blood pressure. Microvascular related to the four risk factors in a curvilinear way: positive for HbA1c and systolic blood pressure but negative for cholesterol and diastolic blood pressure.

Conclusions

We identified several relationships which, when considered in the light of relevant trial data, support a call for major changes to clinical practice. Most importantly, our results support trial data indicating that normalisation of glucose and blood pressure can lead to poorer outcomes, which makes a strong case for target ranges for these risk factors rather than target levels.

Aim 3: Patients excepted from quality standards

Work in progress

Plain English summary

In the UK, general practitioners are paid for treating patients with certain conditions according to clinical evidence in a large and expensive performance-related pay scheme. This scheme is adjusted each year. Using a large dataset on the care received by individual patients, we propose to examine three important issues: what happens when payment for a particular activity is stopped; should some patients be excluded from the scheme; can payments be more closely linked to health gains?

Indicators are used to measure their performance and they express the percentage of the patients for which the appropriate treatment, test, examination etc was performed (e.g. percentage of patients whose blood pressure has been measured).

In order to maximise the benefit from the scheme, indicators need to be routinely replaced. However, we do not know what the effect of removal will be – will quality revert to its original levels, remain stable or deteriorate slightly? Three indicators were removed in the third year of the scheme and we will investigate the impact on their performance.

The scheme allows for doctors to exclude patients from the payment calculations for a variety of reasons. This is in place to protect patients from being discriminated against. For example a patient may refuse treatment and if exclusion was not an option, she would cost the doctor money.

However, the true levels of this provision are unknown. We will estimate its actual levels, investigate the profile of excluded patients and use the timing of exceptions to assess whether they have been used appropriately.

Diabetes is one of the conditions for which general practitioners are rewarded. Some of the indicators are based on evidence of impact on health, though the original study included only patients aged 25-65. We will investigate the effect of all of the indicators on survival and complications for the entire population and examine whether payments to practices should vary with patient characteristics if they are to reflect health gain.

Dissemination

• Kontopantelis E, Springate DA, Reeves D, Ashcroft D, Valderas JM and Doran T. Withdrawing Performance Indicators: Retrospective Analysis of General Practice Performance Under the UKs Quality and Outcomes Framework. British Medical Journal, 2014;348:g330. doi: http://dx.doi.org/10.1136/bmj.g330

Public involvement

Not applicable since analysis of a secondary database with specific policy relevant questions.

Impact

The project investigated three important aspects of incentivisation in UK primary care:

1. It found that partially withdrawing incentives had little or no effect and hence reassured policy makers that re-organising the Quality and Outcomes Framework (QOF) is a valid practice to maximise benefits from the scheme (published). The information has already been used by NICE, as far as we are aware, and will be further used in re-organising the QOF (which is reviewed at least once every two years).
2. It found that relationships between clinical targets (e.g. blood pressure control) and outcomes in the general population are not linear and the risk is as high if not higher for very low levels. Therefore we argued that an upper threshold on which to base quality of care is only half the story and ranges are more appropriate and recommended (under review). We expect this work, when published, will lead to changes in quality indicators in the QOF, to account for our findings.
3. Since quality of care in the UK is largely driven by the Quality and Outcomes Framework it is important to identify if the indicators are widely applicable to the population or they are ‘excepted’ from the quality standards for various reasons. We examined the characteristics of the patients who are exception reported for all conditions incentivised under the scheme and the relationships between types of exception reporting and mortality (not submitted yet). We expect this work to lead to a re-assessment of the use of the ‘exception reporting’ practice, under which GPs are allowed to exclude patients from quality standards included in the QOF.

This project was funded by the National Institute for Health Research School for Primary Care Research (project number 141)

Department of Health Disclaimer

The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the NIHR School for Primary Care Research, NIHR, NHS or the Department of Health.