New research into controversial pay-for-performance schemes has suggested they may help to save the lives of NHS patients.
A "significant" fall in mortality rates for certain conditions emerged in a study into the use of incentives at hospitals in the North West of England.
Economists and health experts from the Universities of Manchester, Nottingham, Birmingham and Cambridge examined how the introduction of a scheme that paid bonuses to hospitals based on measures of quality affected the delivery of care.
The initiative was found to be associated with a saving of almost 900 lives over 18 months.
Lead author Matt Sutton, Professor of Health Economics in The University of Manchester's Institute for Population Health, said: "Researchers have generally concluded that paying bonuses to hospitals for improving quality of care does not affect patient health.
"We examined a unique initiative in which a bonus system from the US was adopted only in North West of England. We found that while research has shown the US scheme had no effect on patient health, the same scheme in the NHS did and resulted in 890 lives being saved during the 18 month scheme."
The research team suggests a number of reasons for this:
The bonuses in the UK were larger than in the US and there was a greater probability of earning a bonus.
Despite the competitive nature of the programme, staff met regularly within the region to share problems and ideas of best practice.
In the US, these schemes are voluntary and only 5% of hospitals take part, whereas all hospitals in the North West of England took part in the scheme.
The research focused on Advancing Quality, a scheme introduced in 2008 at all 24 NHS hospitals providing emergency care in the North West.
The first of its kind in England, the initiative required each hospital to submit data on 28 quality measures concerning five clinical conditions. The researchers examined mortality rates for three of these five specified clinical conditions - pneumonia, heart failure and myocardial infarction.
They compared the figures for in-hospital deaths within 30 days of admission in the 18 months before and after the scheme's introduction.
The combined decrease for all three conditions was 1.3%, the equivalent of a 6% relative reduction - or some 890 lives. The study concludes that the possibility of incentives having a "substantial" effect on reducing deaths in NHS hospitals cannot be ruled out.
Co-author Ruth McDonald, Professor of Health Innovation and Learning at Nottingham University Business School, claimed the findings could have major policy implications.
She said: "Pay-for-performance schemes are being widely adopted, yet until now there's been little evidence that they improve patient outcomes.
"Our findings suggest they can make a positive and significant difference but that, whether they do so, depends very much on how they're designed and implemented."
Performance-related bonuses totalling -3.2m were paid out at the end of the first year, with a further -1.6m following six months later. It was agreed from the outset that the money would be allocated to top-performing clinical teams to invest in further improvements in care.
In total, information for nearly a million patients - including more than 134,000 at the hospitals that took part in the scheme - was examined.
A nationwide pay-for-performance system based on withholding payments rather than paying bonuses now operates at all NHS hospitals.
Professor McDonald added: "These schemes can seem very simple on paper, but in practice they can be very difficult to implement successfully."