We've seen this before. The government is unable (or unwilling) to enact changes to a health-care system badly in need of reform, so the market implements practices intended to reform the system. The most obvious example was the rise of market-driven managed care following the failed attempt by then President Clinton to enact health-care reform legislation. Now, in the absence of any legislative change, a marketplace trend called "pay-for-performance" is fast gaining momentum.
Pay-for-performance is described as an effort, initiated by payers, to realign incentives in health-care services delivery so as to provide incentives for quality improvement. The major focus of the roughly 50 current employer-initiated pay-for-performance programs is to attempt to determine how well health professionals care for their patients and to reward the ones who do best, usually with bonus payments at the end of the year.
Following the lead of the private market, the Centers for Medicare and Medicaid Services (CMS) is experimenting with pay-for-performance programs. In January, CMS initiated the Physician Group Practice Demonstration involving 10 large physician groups across the country. In announcing the demonstration, CMS administrator Mark McClellan, MD, PhD, said, "It is time that we pay for the quality of the health care provided to our beneficiaries, not simply the amount. We are working to apply this in every setting in which Medicare and Medicaid pays for care." CMS has also implemented some pay-for-performance pilot projects in hospitals.
While this emerging focus on health-care quality may, at first blush, be much preferred to the excessive preoccupation with limiting care and containing costs over the last decade, what constitutes a pay-for-performance program is far from standardized, or even clear. To begin with, there is considerable confusion as to whether the incentives should be provided in response to performance-based measures or clinical outcome measures. Some programs include measures of preventative services, such as cancer screening, mammography and immunizations, as well as measures of care for such chronic diseases as asthma, diabetes and cardiovascular disease. Many programs also use client satisfaction measures. Some incorporate utilization measures, such as percentage of generic drugs prescribed or number of medically unnecessary tests ordered. And some programs reward physicians for using various forms of information technology.
In addition to different programs using different types of measures, some programs attempt to incent performance and good quality, while other programs incent quality improvement. Paradoxically, if a program does target quality improvement, it may actually work to disincent performance that has already achieved good quality care, since there may be little room for that care to improve and no possibility of incentive pay. Leaving aside the lack of consistency among pay-for-performance initiatives, it remains an open question whether these initiatives really do improve the quality of care delivered. To the extent that the incented performance is made more difficult by sicker or less adherent patients, preselection of only healthier and compliant patients may result. In addition, most analysts agree that for pay-for-performance programs to succeed, they require a level of information technology (IT) support (e.g., interoperable electronic health records) beyond the current state of IT in the health-care system.
Where will the resources come from to build the necessary IT infrastructure? And, even more basic, where will the resources come from to pay the incentive bonuses from typically finite budgets, particularly if large numbers of health professionals meet the performance goals? (Medicare's pilot pay-for-performance programs have been described as "budget-neutral," meaning no additional funds would be appropriated.) Theoretically, better performance and improved quality could reduce the cost of health care and free up resources for incentive payments, but that remains to be proven.
Although providing incentives for health professionals' efforts to improve the quality of care being delivered seems a good idea in concept, implementation will make all the difference. Specifically, depending on what performance is incented, a program can facilitate the delivery of quality care or actually impede it. If, for example, payers incent those who provide the least amount of service, pay-for-performance will be nothing more than the latest version of managed care. It is important that health professionals delivering the care have considerable input into determining what measures are used as a basis for pay-for-performance programs. To that end, the APA Practice Directorate is working with the New Jersey Psychological Association on a demonstration project to determine how practitioners can best collect outcomes data in a routine, efficient manner to demonstrate the effectiveness of what they do. Other efforts to assure adequate input will also be required. If we are to be paid more for performance, we need to be included in determining just what is good performance.
Letters to the Editor
- Send us a letter