I’m getting to be a broken record in my skepticism about pay for performance. Here, here, here, here, here, and here just to name a few posts. But there’s another study out in the Annals of Surgery, “Does Pay-for-Performance Improve Surgical Outcomes? An Evaluation of Phase 2 of the Premier Hospital Quality Incentive Demonstration“:
Objective: We sought to determine whether the changes in incentive design in phase 2 of Medicare’s flagship pay-for-performance program, the Premier Hospital Quality Incentive Demonstration (HQID), reduced surgical mortality or complication rates at participating hospitals.
Background: The Premier HQID was initiated in 2003 to reward high-performing hospitals. The program redesigned its incentive structure in 2006 to also reward hospitals that achieved significant improvement. The impact of the change in incentive structure on outcomes in surgical populations is unknown.
Methods: We examined discharge data for patients who underwent coronary artery bypass (CABG), hip replacement, and knee replacement at Premier hospitals and non-Premier hospitals in Hospital Compare from 2003 to 2009 in 12 states (n = 861,411). We assessed the impact of incentive structural changes in 2006 on serious complications and 30-day mortality. In these analyses, we adjusted for patient characteristics using multiple logistic regression models. To account for improvement in outcomes over time, we used difference-in-difference techniques that compare trends in Premier versus non-Premier hospitals. We repeated our analyses after stratifying hospitals into quintiles according to risk-adjusted mortality and serious complication rates.
This study involved nearly a million procedures in 12 states over six years. They tried to reduce severe complications and mortality within 30 days. Did paying for performance work?
While mortality went down over time, it didn’t go down faster in the P4P hospitals than the non P4P hospitals. There were no great improvements in complications, and there were no great improvements in mortality. Even in the worst 20% of hospitals (where there is the most room for improvement), P4P was not associated with great improvements over non P4P.
If improving quality were easy, we’d just do it. P4P implies that if we just incentivize docs and hospitals, they’d do it. I wish it were that simple, but it doesn’t appear to be.
This makes a number of my colleagues upset. A lot of them really believe in the potential off P4P. I remain skeptical.