We appreciate Dr. Metoyer's thoughtful comments, many of which the American College of Physicians Ethics, Professionalism, and Human Rights Committee agree with in the College's pay-for-performance position paper. We do not agree, however, with her conclusion that, in some regards, pay-for-performance programs are inherently unethical. Every payment system creates incentives and potential conflicts of interest, such as the incentives in fee-for-service payment to increase care or the incentives under capitation to do less rather than more. The College believes pay-for-performance programs have promise if they can be focused on patient perspectives on care and professionalism, including the duty to ensure medically appropriate care before financial and other considerations. As pointed out in the ACP Ethics Manual (1), "medical practice" does not stand still. Clinicians must be prepared to deal with changes and reaffirm what is most important. The pay-for-performance ethics paper attempts to lay out principles that may guide clinicians in dealing with pay-for-performance programs, as well as provide ethical guidance to those who would design such systems. The problem of current payment systems rewarding substandard care and then paying again to improve that care goes beyond pay-for-performance programs.

Dr. Lowther suggests that pay-for-performance is an untested strategy to improve quality of care and that the concept should be tested before being instituted in widespread fashion. We noted the lack of evidence of effectiveness in our paper. Nongovernmental payers are looking at that now in our market-based system, and questions have begun to appear regarding the effectiveness of pay-for-performance strategies in this setting (2). Dr. Lowther also implies that pay-for-performance is primarily a government function, but that is not currently the case. In fact, the Physician Quality Reporting Initiative program instituted by the Centers for Medicare & Medicaid Services thus far is not a pay-for-performance program but rather a pay-for-reporting program with modest monetary incentives. Nonetheless, we do agree that the risk for deselection of patients is serious and caution that this risk may increase if pay-for-performance programs grow in the current payment environment.

Dr. Brody raises the specter that pay-for-performance may threaten medical professionalism even more than depicted in our paper. The College's hope is to redirect the evolution of incentives for quality by insisting on a focus that puts the needs of the patient first. Fundamental reform of the payment system to encourage comprehensive, coordinated care of the patient would achieve much more than widespread deployment of pay-for-performance. If that were to happen, pay-for-performance programs could play a small but significant role in encouraging certain outcomes, as long as professionalism principles and safeguards against unwanted outcomes were built in.