Pay-for-Performance Doesn’t Lower Mortality Rates

Pay-for-performance programs, which increase or decrease reimbursement to providers based on either direct health outcomes or on process measures of care, have now been in effect for a long enough time to analyze whether they have a sustained effect on those outcomes or process measures of care.  A study in the New England Journal of Medicine examines the results of a hospital pay-for-performance program begun in England in 2008.   (NEJM Article)   The basic design of the program was to incent 24 hospitals in the northwest region of England to reduce 30-day mortality for five conditions:  heart attacks; pneumonia, heart failure, hip or knee surgery and coronary artery bypass surgery.  The hospitals were measured on a number of care processes and the top performers received bonuses.  There was no penalty for poorer performance at the start of the program, but there were later on.  A previous analysis had indicated that the hospitals had achieved mortality reductions after 18 months, but the current study, which extended to 42 months after the start of the program and included hundreds of thousands of admissions, found that this effect was not sustained.  The hospitals did show improvement in performance on the quality measures over the study period, with the fastest gains coming early on and with performance plateauing near the end, but at a high level.  But over the entire study period, while mortality was reduced in this region of England for the study conditions, it was reduced even more in hospitals in other regions which were not part of the program.  There were some potential “spillover” effects in reduction of mortality for conditions not covered by the incentives.

The results of this program are similar to those of other hospital pay-for-performance programs.  It is unclear whether the incentives really have no long-term effect or whether the heightened visibility of quality improvement as a priority in health care has led to such general increases in care quality that specific pay-for-performance program effects are hard to tease out.  The results should lead policymakers to reconsider whether the costs incurred for these initiatives are worth it or whether pay-for-performance is actually the best approach to improving care quality.