Pay-for-performance model sputters over long term

A pay-for-performance model in England failed to show much sustained improvement in 30-day mortality for heart failure (HF) and acute MI (AMI), in spite of incentives and penalties, according to a study published online Aug. 7 in the New England Journal of Medicine.

The study also reported that over time, the effects of the pay-for-performance model seemed to decrease compared to control hospitals, even though improvements continued to be seen compared to the participant hospitals’ own starting rates.

Pay-for-performance models have become a global go-to in the pursuit of improving hospital care in industrialized nations. But, were early successes sustainable in the long term?

This was the question posed by the research team led by Søren Rud Kristensen, PhD, of the Manchester Centre for Health Economics at the University of Manchester. They reviewed the performance of hospitals in the northwest region participating in the model and compared them to controls in other areas of England as well as to the performances of these northwestern hospitals from before the program inception to 24 months following. Their analysis included pneumonia as well as HF and AMI.

They found initial short-term improvement in 30-day mortality across all conditions being surveyed of around 1.7 percent within the region and 0.9 percent compared with other regions. Yet this change could only be sustained so far. Long-term improvement to 30-day mortality was seen only in the combined endpoint (3.3 percent reduction within region and 0.1 percent compared to other regions from beginning to 24 months following the programs’ start).

AMI had short-term improvement of 30-day mortality of around 1.1 percent within the northwest region and 0.1 percent when compared with other regions. Beyond the first 18 months, AMI saw a decline in improvement against other regions of 0.2 percent, showing improvement only within region by 2.2 percent. Notably, AMI 30-day mortality rates for the control hospitals improved independently (2.3 percent improvement from start to 24 months) during the same period, driving the apparently lackluster comparison with hospitals that were part of the pay-for-performance group.

After an initial improvement of 1.1 percent compared to start within region and 0.9 percent compared with other regions, HF showed no long-term improvement with outside the region. However, it did improve by 2.7 percent from start to 24 months within the region.

"These results suggest that the benefits of initiatives such as paying for performance may be temporary,” Kristensen said in a press release. “Our findings could also be explained by the decision taken midway through the program to change the incentives from bonuses for good performance to fines for failing to achieve targets.

"But we also found evidence to suggest that unintended but desirable spill-over effects may have occurred. These include improvements in the quality of care provided to both patients treated in hospitals in other regions, as well as patients admitted in the North West for conditions not covered by the incentive program."

Kristensen et al noted that the effects of spillover on other hospitals, both negative and positive, were a promising area of future study.