Pay-for-performance: clear, simple and wrong – What’s your opinion?


The use of pay-for-performance initiatives in a bid to improve standards of care within the NHS remains a matter of controversy, not least because such schemes threaten to undermine our trust in public servants to do the best for those they serve

A growing body of research suggesting these programmes do not work has not helped their standing. Our evaluations of new incentive structures in contracts for GPs, pharmacists and dentists suggest these have not produced the benefits that policymakers hoped for.

Yet the Department of Health has pressed ahead with pay-for-performance (PfP) schemes aimed at changing the practices of other groups of staff, including hospital doctors and nurses. Since these are public employees, a slightly different approach has been required – one that links performance to organisational rather than personal income.

The Commissioning for Quality and Innovation Payment Framework (CQUIN) makes a proportion of income conditional on attaining certain goals. In the first year this proportion was 0.5%, which sounds meagre until you consider total hospital incomes can amount to hundreds of millions of pounds.

The figure was increased to 1.5% in the second year of the scheme and 2.5% this year. David Nicholson, the NHS’s Chief Executive, has said he wants to see a further rise, to 4% or 5%.

There is some evidence that 1.5% is the figure at which, to put it bluntly, the people who manage organisations really sit up and take it notice. But one of the problems with many PfP initiatives may be that they are overly focused on numbers while paying insufficient attention to the more human elements involved.

This focus on money and a quest to find the right percentage at which PfP schemes ‘work’ is understandable given that much of the thinking that has shaped PfP in the NHS has traditionally been dominated by economists, who can tend to take a somewhat one-dimensional view of human behaviour. But rather than asking whether PfP schemes work, we should be asking why some of them work.

The simple answer is that they are all different. Various health systems around the world are experimenting with programmes that seek to align financial incentives with policy objectives, and what we need to do is to look at the features of each – its design, implementation and context – and gauge the impact to work out how and to what extent it succeeds.

Our recent assessment of the Advancing Quality scheme at 24 hospitals in the North West discovered it saved almost 900 lives over an 18-month period. The bonus payments were small compared to those available through CQUIN, which implies that NHS staff may be perfectly willing and able to improve services but sometimes just need a little help.

In this case that help took a number of forms, including standardised data definitions and bespoke software. Alongside the more technical aspects, however, was the human element.

Collaborative events brought together employees from every participating hospital to share best practice and work through common problems. The development of this Advancing Quality “community” appears to have been vital in providing emotional support for the people involved and, indeed, in making the scheme a success.

Our study clearly demonstrates that PfP can produce the desired results. Doctors and nurses are interested in doing the right thing for their patients, but sometimes they need a little help. What motivated the frontline NHS staff in this case was not some careful calculation of effort versus reward. It was the ability to work together and draw on support that made the difference.

The prevailing fixation with finding the magic percentage that suddenly and miraculously justifies PfP brings to mind HL Menken’s remark that for every complex problem there is an answer that is clear, simple and wrong – as well as the adage that the road to hell is paved with good intentions. It’s time to take a different road with PfP, and understanding that people are more complex than economic textbooks would have us believe would be a good place to start.