‘Pay-for-performance’ programs may need a reboot


June 17, 2016 — A new study suggests that so-called “pay-for-performance” programs—in which hospitals are financially rewarded for better patient outcomes and penalized for worse outcomes—may not be working. Ashish Jha, senior author of the study, discusses the implications.

Pay-for-performance programs have been touted as an important way to improve hospital care, but your study suggested otherwise. Did this surprise you?

We were surprised. We looked at the Hospital Value-Based Purchasing (HVBP) program, a national program that Medicare introduced in 2011, the largest of its kind in the world, under which hospital performance in a number of areas—such as patient outcomes, hospital efficiency, and patient experience of care—is tied to Medicare reimbursements. Our study focused on patient outcomes, specifically mortality. While improving outcomes is hard, we had hoped that hospitals would have taken the pay-for-performance program as a signal to improve their care. But we found that the program had no impact on mortality rates within 30 days of patients’ hospitalizations for heart attack, heart failure, or pneumonia—three conditions that are specifically incentivized under the program.

Even when we examined just the worst performing hospitals—the ones that started with poor outcomes at the beginning of the study—we found that the program had no effect on helping them improve.

In retrospect, one could argue that the results were predictable. Two other big hospital-based pay-for-performance programs, one in the U.S. and one in the U.K., have also failed to make a big impact on patient outcomes.

You and your co-authors suggest that policymakers should consider changing the structure and size of the bonuses and penalties that are offered to hospitals. What might these new incentives look like?

Currently under HVBP, hospitals receive either higher-than-usual Medicare payments or lower-than-usual Medicare payments depending on how hospitals perform on a range of measures. Right now, the incentives—both the bonuses and the penalties—are small and it is clear that hospitals are not making big changes as a response to those incentives. We should at least experiment with putting more money on the table. Right now, a hospital can get a bonus (or receive a penalty) that is up to 2% of its total Medicare payments. What if we experimented with 5% or 10% of Medicare payments as potential bonus or penalty? Something like that should motivate organizations to invest in improvement.

Another issue is the complexity of the incentives. The bonuses and penalties are based on a variety of factors, not just how well you perform on various measures. The formula also includes how much a hospital improves, whether it hits certain national targets, and how it compares to other hospitals. The incentive scheme is so complicated that most hospitals have no idea if they are doing well or badly. That needs to change.

Finally, Medicare should focus incentive programs on a small number of higher value measures, such as mortality and infection rates, instead of a broad swath of both meaningful and non-meaningful measures. For example, some hospitals are rewarded for how well they document certain tests. That’s not useful information for consumers and doesn’t motivate doctors to get better.

Are there other approaches besides pay-for-performance that hold promise?

The U.S. is beginning to experiment with new models of paying for hospital care. The HVBP program is still relatively new and there are other efforts, such as penalizing hospitals for high readmission rates. We don’t know what approach is best for improving hospital care at this point. But the bottom line is that improving hospital care remains enormously important and only through experimentation and careful evaluations can we ensure that all Americans receive high-quality hospital care.

Karen Feldscher

Learn more

Are Hospital Pay-for-Performance Programs Failing? (Harvard Magazine)