Pay-for-performance programs stir debate

Many health care organizations are instituting pay-for-performance measures, in part because of cost-containment motives. A physician who signs on to such a program may find ethical conflicts between patients' interests and his or her own financial benefit.

Reply:

The American Medical Association's position on pay-for-performance (P4P) programs is defined in two recent reports, one from the AMA Board of Trustees and one from the AMA's Council on Ethical and Judicial Affairs. These reports describe boundaries within which P4P designers must create their plans and spell out the ethical obligations of physicians who participate in them.

The board report stipulates, for example, that such programs must improve quality of care by means of evidence-based measures and must permit variations in individual patient care based on clinical judgment. Valid analytical methods should be used in assessing physician performance, and physicians should be allowed to review, comment upon and appeal results of those analyses.

The CEJA report warns against a variety of conflicts of interest that could undermine patient care and specifies ethical guidelines for physicians in administrative positions and for those in practice. Physician executives who design or implement P4P programs should ensure that incentives are intended and structured primarily to promote quality of care and safety of patients. Cost containment should be neither a primary driver of policy, nor the primary goal of such programs. If too heavily emphasized, cost containment can be inimical to quality care and patient safety. Not only should built-in flexibility allow physicians to accommodate varying needs of individual patients, but performance measures also should be adjusted by risk stratification and case mix to ensure that physicians are not encouraged to avoid high-risk individuals and populations.

Practicing physicians who participate in P4P programs have specific ethical obligations arising from the principles of medical ethics described in the AMA's Code of Medical Ethics. Responsibility to patients must take precedence over all financial considerations, including P4P incentives. The possibility that selective treatment of healthier patients and avoidance of sicker ones could enhance performance outcomes constitutes a conflict of interest that must be resolved in favor of caring for patients along the entire spectrum of illness; cherry-picking is not ethically permissible. Physicians must always provide care in accordance with the needs and preferences of individual patients, even when those needs and preferences conflict with relevant practice guidelines and even though physicians' outcome scores and incentive payments might suffer.

The design of P4P programs undoubtedly will come under politically motivated pressures, as when the push for cost savings threatens the priority of quality care and patient safety. Moreover, if quality incentive programs were to be financed by redistributing existing physician payment funds rather than by adding new money to the system, the quality incentive program would be built on the backs of already overburdened physicians.

Because risk and case-mix adjustments are not easy to make, narrowly conceived outcome measures may find their way into many such programs, increasing the temptation for physicians to avoid sicker patients. Such potential policy flaws are likely to appear in some P4P programs, to the detriment of patients.

Pay-for performance programs that fail to meet the AMA's standards will not be acceptable, and physicians should not participate in unacceptable programs, such as those that compromise their commitment to their patients' best interests.

Financial incentives are powerful motivators, and advocates of P4P programs have proposed to use that power to improve quality of patient care and safety of patients. Yet only programs that meet the standards set forth by the AMA have any possibility of reaching those goals.

Physicians cannot be forced to abide by ethical standards, but there is at least one good practical reason for them to do so: We have the power, if we choose to exercise it, to ensure that no ethically flawed P4P program can be implemented.

That can be done by saying "no" to participation in any unacceptable program. That power is fueled by adherence to the AMA's pay-for-performance guidelines, which are based on the central ethical principle of the Code of Medical Ethics: our paramount responsibility to care first for our patients.

Robert M. Sade, MD, chair, Council on Ethical and Judicial Affairs, American Medical Association; professor of surgery, Medical University of South Carolina, Charleston, S.C.

Reply:

In recent years we have seen the erosion of the ethical and professional standards of the medical profession by a tidal wave of corporatization, bureaucratization and commercialization. Now pay-for-performance, an idea that will contribute to this erosion, is rapidly gaining momentum. P4P, still far from standardized, is nevertheless endorsed by many organizations, public and private.

Its basic concept is to set higher reimbursement rates or bonuses for physicians who comply with specified standards of care delivery. This approach seeks to improve quality but it rests on three flawed premises, and, as a result, its implementation will have major unintended and adverse consequences.

Premise 1: The idea that P4P rests upon a reliable, valid statistical foundation is fallacious.

The statistical methodology of P4P has two major defects. First, it is vulnerable to the weaknesses of all statistical analyses: biased measurements; inadequate sample sizes; multiple confounding factors, many of which cannot be excluded or adjusted for; and manipulation of definitions and data to serve private ends ("gaming the system"). Errors from these statistical weaknesses abound in medical care, and many physicians see themselves as losers in a game they lack the statistical skills to play.

But there is a deeper, structural defect: A statistical measure is valid to the extent that it measures what it purports to measure. But P4P measures fail to evaluate the essential features of a physician's work, i.e., diagnostic skill and clinical judgment -- the cornerstones of medical practice -- and the unmeasurable quality of the relationship between patient and physician upon which diagnosis and treatment depend.

Instead P4P focuses almost exclusively on the statistical goals of population health and practice guidelines, leaving deskilled and dehumanized clinical care in its wake (with consequent increases in malpractice litigation and racial and ethnic disparity in health care).

Thus, P4P measurements are invalid as assessments of clinical performance, and its scores and rankings cannot serve as proxies for the competencies that the public expects and malpractice law demands from a physician.

Premise 2: The idea that a system of rewards and penalties will improve the quality and safety of patient care is another fallacy. To accept this is to ignore or deny the root causes of poor quality care, none of which can be cured by bonuses. One of the most important of these is the time constraints under which physicians function. Pay-for-performance will do nothing to change these increasingly restrictive constraints that form nearly insurmountable barriers to proper practice of our clinical skills.

Premise 3: The idea that P4P is consistent with medical ethics is perhaps the most dangerous of the flawed assumptions upon which the system rests. Indeed, there is an inherent and irresolvable conflict between the conceptual framework of P4P and the ethical standards of the profession. Those standards derive from a core of fiduciary responsibility, in which one person, the patient, depends on the superior knowledge and skills of another, the physician, and places complete confidence in that person in regard to a particular transaction -- in this case, medical care.

The fiduciary is held to a higher standard of legal and moral conduct and trust than a stranger or a business person. Once established, a professional relationship in medicine obligates the physician to do his or her best for the patient regardless of reward. The duty goes beyond the "due care" standard of tort law to a higher level of loyalty and commitment that is not contingent on rewards or penalties.

The incentives of P4P programs presume that the fiduciary relationship is insufficient motivation for physicians to do their best.

To give such an idea official standing through P4P is to erode further our ethical standards and push us farther down the slippery slope to deprofessionalization. Pay-for-performance inevitably will give rise to such unethical practices as "cherry-picking" to avoid difficult cases and the deselection of patients whose care fails to yield sought-after bonuses.

These practices cannot be prevented by dubious statistical procedures such as case-mix adjustment, nor can voluntary participation in P4P long remain a realistic option.

Indeed, under pressure from members of their group practice or other employers or government agencies, physicians may be forced to deselect their difficult patients or be deselected themselves. One does not have to look far to see how such a system will foster practices that violate the timeless ethos at the heart of medical professionalism: compassion (not rejection!) for those who need us most.

Edmund Blum, MD, general internal medicine, Kings County Hospital Center, Brooklyn, N.Y.; clinical assistant professor, Downstate Medical Center, State University of New York, Brooklyn