P4P--Measuring the Right Outcomes?

Pay for performance is the highly touted concept of providing financial rewards to hospitals and physicians who can achieve better outcomes for the treatment of various disease entities or lesions. On first inspection, it seems like a great idea. If one physician has lower than average 30-day mortality and morbidity rates for repairing an abdominal aortic aneurysm (AAA) or a lower major adverse event rate for treating carotid bifurcation disease, that physician may seem deserving of a higher fee than are surgeons with higher mortality and morbidity rates. It makes sense, or does it?

Improved outcomes may be a reflection of better care, or they may be the result of treating easier cases. Conversely, worse outcomes may result from inferior care or from the treatment of less favorable, higher-risk patients. Clearly, under pay-for-performance (P4P) systems, there is a financial incentive to treat easier cases, some of which may be better served by receiving noninterventional medical treatment. This is certainly the case with asymptomatic patients who have a moderate- or low-grade carotid stenosis or an AAA less than 4.5 cm in diameter. Invasive treatment of such patients, although financially rewarding to surgeons and hospitals, may not benefit the patients. Such treatment also contributes substantially to rising health care costs. Worse still for patients, P4P provides a financial disincentive to treat invasively the symptomatic carotid stenosis patient with multiple risk factors and unfavorable carotid and arch anatomy. These are precisely the patients who are most likely to be helped by such aggressive treatment and most likely to be harmed if not so treated.

The incentive not to treat the patients who need treatment the most is made even worse if surgeon outcomes are made public, as many have advocated. Surgeons will shy away from taking on high-risk, difficult cases, even though they might be quite capable of doing them successfully most of the time.

This was certainly true with cardiac surgeons in New York state after the mortality rates of these surgeons were published. Few surgeons would operate on high-risk patients for fear of having their practice unfairly harmed by their published higher-than-average mortality rates. Even though risk stratification can help prevent such unfairness, most systems for stratifying risk are imperfect. Moreover, patients who look at mortality rates usually do not read the fine-print qualifiers.

It certainly is appropriate to reward institutions and surgeons who can achieve better outcomes for therapeutic procedures. Pay-for-performance systems can do that, but only if comparisons are carried out in rigidly defined similar patient groups with similar lesions--a difficult task to accomplish and one that may be impossible to police.

Moreover, whatever system is introduced to do so must recognize the very real potentials for unfair comparisons of surgeon-providers, for possible negative effects on patient care and for increased costs by promoting procedures that are unnecessary.

It is important that any system that is adopted must include steps to avoid these detrimental effects.

As with many concepts in life in general and health care in particular, good ideas can create more problems than they solve. This is certainly true with P4P systems.