This site is intended for health professionals only

At the heart of general practice since 1960

Should the QOF focus on outcome targets?

A QOF based on process measures has taken general practice as far as it can - it's time to reward GPs for achieving improved outcomes, says Professor Azeem Majeed. But Professor Richard Lilford argues health outcomes may be what matter to patients, but they are far too unreliable to be used as the basis for GP pay

A QOF based on process measures has taken general practice as far as it can - it's time to reward GPs for achieving improved outcomes, says Professor Azeem Majeed. But Professor Richard Lilford argues health outcomes may be what matter to patients, but they are far too unreliable to be used as the basis for GP pay

The QOF is now well established in primary care. With control of the framework now lying with NICE, we can expect a greater focus on the clinical effectiveness and cost effectiveness of quality indicators. In the early years of the QOF, a large proportion of target payments were awarded for achieving process measures, such as measuring blood pressure or HbA1c.

Achievement of such process measures is now very high, with little room for improvement in many practices and limited potential for further health gain.

By contrast, control of key risk factors, such as blood pressure in people with hypertension, or HBA1c in people with diabetes, could still be improved considerably.

But changing patients' behaviour or optimising their management to achieve such improvements is difficult and imposes a considerable workload on primary care. So it is logical to consider increasing the financial incentives for achieving better management of such intermediate clinical outcome measures. It is also appropriate to raise the thresholds for achievement of quality targets to ensure practices that provide the very best care are rewarded appropriately for their work.

At the same time as improving the management of patients with established diseases such as diabetes, hypertension and stroke, we also need to reward practices that are effective in preventing disease and delaying the onset of complications once a disease such as diabetes or hypertension is present. The UK is faced with major public health challenges in areas such as obesity, smoking, alcohol and sexual health.

Some GPs may argue improvements in such areas are outside the scope of primary care and lie with wider societal interventions. But this underestimates the ingenuity of primary care teams.

There are practices that have implemented innovative schemes in targeting public health, with support from their primary care organisations. An example is the QOF Plus scheme in Hammersmith and Fulham PCT, which has incentivised reductions in smoking and alcohol misuse, and areas such as cardiovascular disease prevention.

An example of such an indicator might be the proportion of patients with newly identified impaired fasting glycaemia or impaired glucose tolerance who progress to normoglycaemia after entering a disease prevention programme. Practices that can implement effective interventions to improve management in these areas will have a direct impact on the health of their patients, and should be rewarded.

Outcome-based quality targets, whether for clinical indicators or public health indicators, are more difficult to achieve in some sections of the population. GPs working in inner-city areas may feel they have to work much harder and invest more practice resources to perform well on outcome-based quality targets. This is true - but we can address this by ensuring that, where appropriate, payments for achieving quality targets reflect the population GPs serve, to ensure those working with vulnerable populations are rewarded sufficiently for providing high-quality care and starting to address health inequalities.

We are right to be proud of the NHS. We all want it to provide high-quality care. Despite our best intentions, a considerable gap remains between best achievable practice and the quality of care actually being provided. Focusing the QOF on clinical outcomes will help address this quality gap by rewarding practices that ensure our patients receive the care they need to maintain their health, prevent the onset of disease and optimise their management when patients do develop a chronic disease.

Professor Azeem Majeed is professor of primary care at Imperial College London and a GP in Clapham, south London

Outcome measures are neither a sensitive nor a specific measure of quality and the idea of using them as QOF markers is deeply flawed. There is an association between thequality of care enjoyed by a group of patients and their health outcomes, but it is a statistical association and it's weak. It is weak because it is so noisy - all sorts of factors other than the quality of care influence health outcomes.

The most important influence on outcomes is the severity of the underlying condition, and even with adjustments you will never eliminate this bias. Some attempts at adjustment can even exacerbate bias1. Lots of external factors also influence outcomes, such as the standard of living and lifestyles of a practice's patients.

Outcomes will inevitably be better in the leafy suburbs than in an urban area of high deprivation, but this may have nothing at all to do with quality of care. A QOF based on outcomes is likely to be systematically biased against those practices with the most difficult mix of patients.

Outcome measures are also extremely imprecise, especially in general practice.

An adverse outcome occurs in only a proportion of patients, and even for hospital procedures where adverse outcomes are relatively common, there is so much statistical variation it is difficult to accurately differentiate between clinics.

In general practice, adverse outcomes are rarer - far too rare to accurately differentiate between practices, since the confidence intervals for outcomes are too wide and frequently overlap.

Harold Shipman's practice only just came out as statistically worse than its peers for mortality, even with such an extreme example of poor performance. Finer differences in outcome are impossible to pick up. And a lot of the good things GPs can do to modify patients' health outcomes only do so in the very long term.

We should continue to go down the process route, as we know it works. Instead of looking at how many heart attacks have occurred over a given period, you can look at use of statins in those who have an indication for them.

This takes out some of the bias, since there is likely to be less variation between populations of people with an indication for a statin than between individuals with an illness. Put more technically, the denominator for process errors can be opportunity for error, and this partly corrects for differences in prognosis.

It is also more precise to base QOF indicators on process rather than outcomes, because the number of people who have an indication for a statin is much larger than the number who will suffer a heart attack, so the confidence intervals when comparing practices will be much smaller. And lastly, measuring process is immediate.

So why are so many people pushing for measures of outcome? Well, measuring processes is expensive. Measuring outcomes is the lazy, cheap approach.

If you're serious about quality, you want to improve it across the entire spectrum. It's better to improve the mean than to shift the last 5% of cases, as you may be doing with outcomes targets.

And if you tell a GP their health outcomes are bad, they may not be sure what to do about it. Whereas if you tell them their performance on a process indicator is bad, they know exactly what to do to bring about an improvement.

It is seductive to incentivise GPs based on health outcomes, because outcomes are what matter to patients. But there is a problem. Outcomes do not tell you enough about the quality of healthcare to be used as the basis for GP pay.

Professor Richard Lilford is professor of clinical epidemiology and head of the school of health and population science at the University of Birmingham, and a former GP

Should GP pay be tied to the prevention of complications such as diabetic retinopathy? yes No The shift to outcomes

Outcome measures in the QOF have found favour with a host of politicians, policy bodies and academic researchers. The Department of Health went on record as early as July 2008 with its intention to 'move away from rewards for measuring to rewards for health benefits'. Even if Labour is ousted at the next general election, the Tories said in June 2008 they also wanted GP pay to be tied to outcomes.

Rate this article 

Click to rate

  • 1 star out of 5
  • 2 stars out of 5
  • 3 stars out of 5
  • 4 stars out of 5
  • 5 stars out of 5

0 out of 5 stars

Have your say