Nothing in this title shall be construed to authorize any Federal officer or employee to exercise any supervision or control over the practice of medicine or the manner in which medical services are provided, or over the selection, tenure, or compensation of any officer or employee of any institution, agency, or person providing health services; or to exercise any supervision or control over the administra­tion or operation of any such institution, agency, or person.

—Social Security Act, Sec. 1801, Title XVIII[1]

There is good news and bad news on Medicare reimbursement. The good news is that Members of Congress are unhappy with the Medicare physician payment program that they created. It is a complex system of administered pricing and price controls, governed by elaborate statutory formulas and charac­terized by mind-numbing regulatory micromanage­ment. In sharp contrast to reimbursement for professional services in other economic sectors, Medicare providers are not paid according to their skill levels, their innovative treatments, the quality of the care delivered to individual Medicare patients, or the specific benefits provided to patients. Moreover, under current government formulas, they can look forward to future reductions in Medicare reimburse­ment even though they are expected to treat a dramat­ically larger Medicare population.

Needless to say, most physicians are unhappy with Medicare’s payment system—a view increasingly shared by senior Members of Congress. House Ways and Means Committee Chairman Bill Tho­mas (R–CA) and Health Subcommittee Chairman Nancy Johnson (R–CT) have said, “It is time to change this irrational system.”[2]

The bad news is that, instead of enacting real reform, Congress is preparing not only to keep Medicare’s rigid system of price controls and cen­tral planning, but also to add another layer of reg­ulatory control over physician behavior. Senate Finance Committee Chairman Charles Grassley (R–IA) and Ranking Member Max Baucus (D–MT) are sponsoring the Medicare Value Purchasing Act of 2005 (S. 1356),[3] which would implement “pay for performance” in the Medicare program by tying physician payment to compliance with govern­ment-defined medical guidelines. Representative Johnson has introduced a similar bill in the House. The approach is well intentioned, but more central planning will only intensify the Medicare reim­bursement problem, not ameliorate it.

A Misguided Approach

The concept of “pay for performance” in Medi­care is unquestionably attractive to federal policy­makers and suggests—correctly in our view—that Medicare patients and the taxpayers are not get­ting the best value for their money. Using the rhet­oric of “best practices” and “evidence-based medicine” to describe this approach, proponents are creating the false impression that new govern­ment guidelines would promote market-like com­petition, control costs, and improve the quality of health care delivered within Medicare. They believe that adopting this approach would simul­taneously control the growth of Medicare costs in a more rational fashion and close the gap “between the health care we now have and the health care we could have.”[4]

Before succumbing to the latest health care pol­icy fad, Members of Congress should carefully con­sider two things: the likely impact of government incentives designed to secure physician adherence to centrally determined standards and whether or not those standards can indeed provide higher quality to patients and better value to taxpayers. Despite the rhetoric to the contrary, this proposal is anything but a free-market approach to physician payment. It is, in fact, a compliance-based system, inherently burdened by serious limitations. For example, such a system would:

• Dump patients into a system of top-down, “cookbook” medicine that is incompatible with high professional standards of patient care;
• Spawn an increasing number of Medicare rules, regulations, and guidelines, further undercut­ting the physician’s professional autonomy and integrity, as well as patient choice and access to care;
• Undermine the more desirable goal of high quality, which requires personalized care;
• Retard medical innovation and introduce unproductive gaming by doctors to secure higher Medicare reimbursement; and
• Further weaken the traditional doctor–patient relationship.

A Better Policy

If Members of Congress had seized the opportu­nity to replace the current Medicare system with a premium support program similar to the Federal Employees Health Benefits Program (FEHBP) dur­ing consideration of the Medicare Modernization Act of 2003, they would not now need to fix the Medicare physician payment system that they cre­ated. The FEHBP is characterized by consumer choice, market competition, and minimal bureau­cracy and regulation. The Office of Personnel Man­agement, which administers the FEHBP, does not prescribe detailed formulas for physician payment for thousands of medical services, enforce price controls, or conduct compliance audits or investi­gations into physician payment.

However, short of serious and comprehensive Medicare reform, Congress should go back to the drawing board and design a new reimbursement system for Medicare doctors, combined with reli­able market-based updates for physicians’ services. Meanwhile, Congress should:

• Abolish the current fee schedules and the update formulas;
• Eliminate Medicare restrictions on balance billing (effectively a price control system) and allow doctors to charge either more or less than the Medicare fixed price for medical ser­vices; and
• Require physicians, as a condition of participat­ing in Medicare, to disclose the prices that they charge for Medicare services.

As a national market develops, private-sector organizations (e.g., consumer, professional, and seniors’ groups) could generate information on the quality of health care services, meeting the market demand for quality information.

The Current Medicare Reimbursement System

Medicare’s current reimbursement policy is a complex, formula-driven system of administrative pricing, central planning, and price controls.[5] It has three main features: the fee schedule, updates and controls, and balance billing restrictions.

The RBRVS Fee Schedule.[6]Medicare uses the resource-based relative value scale (RBRVS) to pay for physician services. Under this formula, Medicare offi­cials compute the “objective value” of an estimated 7,000 procedures.[7] Each component of a medical service is assigned a weighted value that is calculated by using social science measurements of the time, energy, and effort required to perform a given proce­dure, including resource inputs such as medical equipment, malpractice insurance, and administra­tive costs. These weighted “values” are then con­verted into dollar amounts and used to determine the fees that Medicare pays to physicians for those ser­vices.[8] The diagnosis related group (DRG) system reimburses hospitals using a similar strategy.[9]

Updates and Controls. Attempting to limit Medi­care physician costs, Congress also created volume controls, based on an official projection of the “appropriate” growth rate of Medicare physician ser­vices. Since 1997, these volume controls have been tied to the Sustainable Growth Rate (SGR), the con­gressionally created formula for determining annual updates in physician reimbursement rates under the Medicare fee schedules. Under the SGR, the annual update in reimbursement is linked to the aggregate level of Medicare spending for physician services. If spending exceeds the government target, which is based on growth in the national economy, a statutory algorithm reduces the increases in the reimburse­ment rate.

The SGR system has been ineffective in control­ling volume—the volume of physician services per beneficiary rose by almost 22 percent between 1999 and 2003[10]—while creating new problems for physician reimbursement. The Centers for Medicare and Medicaid Services (CMS) recently reported that Medicare’s sharp 15.2 percent increase in spending for physician services in 2004 was due almost entirely to volume growth. The result was a scheduled 4.3 percent cut in physician reimbursement rates, beginning January 1, 2006.[11] CMS actuaries are projecting similar negative pay­ment updates of 5 percent annually for the next seven years, which means that physician payments would decrease by more than 31 percent from 2005 to 2012. During the same period, physician practice costs would go up by 19 percent.[12]

As doctors find it financially burdensome to treat Medicare patients, they will stop accepting new ones. According to a recent American Medical Association survey, 38 percent of physicians will reduce the number of new Medicare patients that they see as a result of the impending 2006 cuts.[13] More than one-fifth of Medicare enrollees already have trouble finding a primary care physician, and 27 percent report delays in getting an appointment, according to a recent Medicare Payment Advisory Commission (MedPAC) study.[14]

There are several reasons for the SGR’s inability to control volume. For one, the growth rate of the national economy (as measured by GDP) has very little to do with the growth rate of services, making it a poor benchmark for a spending target. For example, during a recession, the spending target is pushed downward, punishing physicians, even though practice costs and demand for services do not drop proportionally. In fact, the growth in Medicare spending is driven largely by new tech­nology, patient needs, and public policies related to the provision of services. These factors are outside the direct control of physicians, making them use­less targets for volume control incentives.

Moreover, even if physicians did exert a large degree of control over Medicare spending, individ­ual doctors cannot be expected to respond to col­lective incentives. An aggregate spending target and universally applied rate update will not decrease (and may even increase) the short-run incentives for individual physicians to increase vol­ume, because they know that any personal effort to reduce services would not result in a proportional increase in payments.[15]

Balance-Billing Restrictions. The third feature of Medicare’s current reimbursement policy is the restriction on balance billing, the amount charged by a provider to a patient above what Medicare is will­ing to reimburse. With the enactment of the Bal­anced Budget Act of 1997, Congress imposed new legal obstacles to private contracts for services per­formed on Medicare recipients, cutting off an escape route for doctors and patients who might want to enter into a private payment system outside the Medicare program.[16] Meanwhile, providers must incur losses because of inadequate reimbursement for Medicare services,[17] and they do not have the flexibility to adjust prices to attract new business or react to market challenges and opportunities. By removing any remnant of a price mechanism, bal­ance-billing restrictions compound the inefficiencies of the Medicare physician payment system and stifle improvements in quality and value.

In summary, Medicare pays doctors according to a resource-based formula that embodies an “objective” theory of value that is utterly inconsistent with mod­ern economics, combined with inefficient price reg­ulation and an illogical reimbursement update formula. The current Medicare fee schedule does not and cannot account for differences in physicians’ skills, quality of service, and benefit to the patient any more than a physician can account for the state of the national economy. While common sense would dictate abolishing this outdated approach in favor of a rational system of market pricing, Con­gress is instead preparing to impose another layer of regulatory compliance on physicians.

The Genesis of Pay-for-Performance Reimbursement

Health policy experts who advocate pay for per­formance in provider reimbursement invariably rely on the concept of evidence-based medicine (EBM). Originally developed in the 1970s and 1980s by clinical epidemiologists at McMaster Uni­versity in Canada, EBM is an attempt to apply epi­demiological principles to clinical care and promote reliance on research data, particularly ran­domized controlled trials (RCTs), in the practice of medicine.[18]

In a performance-based reimbursement system, EBM is used to develop “clinical practice guide­lines” and compensate health care providers according to their compliance with the “best prac­tices” dictated by a third party’s interpretation of RCTs. Health maintenance organizations (HMOs) seized upon this concept in the 1990s, employing strict practice rules to override the clinical judg­ment of the treating physician in order to control utilization and limit costs. In practice, they used guidelines to rule instead of to guide, stressing evi­dence-based guidelines to the exclusion of clinical judgment.

Despite provoking a backlash from physicians and patients during the mid to late 1990s, practice guidelines continued to be developed for physician practice at a rapid rate. According to one estimate, more than 1,000 guidelines are being developed annually by quality-of-care organizations, medical associations, and health insurance plans.[19]

Those who advocate tying clinical guidelines to financial compensation through pay-for-perfor­mance reimbursement claim that it will narrow the gap between “ideal,” cost-effective care and actual care observed in clinical settings. In their view, financial teeth are needed to motivate physicians to standardize their treatment decisions. They believe that this would reduce medical errors, optimize quality of care, and control escalating health care costs by controlling price and utilization more directly.[20]

Enter Medicare. The Centers for Medicare and Medicaid Services obtained the authority to experi­ment with pay for performance in Medicare from the Medicare Modernization Act of 2003 and the Medicare, Medicaid, and SCHIP Benefits Improve­ment and Protection Act of 2000.[21] The agency has established eight demonstration initiatives: Two address the quality of clinical care in hospitals, three deal with physician offices and integrated health systems, and three test specific models of chronic care improvement and disease management.

The Premier Hospital Quality Incentive (PHQI) demonstration, the most advanced initiative, is designed to track and reward performance in treat­ing five chronic health conditions at 270 hospitals around the country.[22] Provider compliance with a uniform set of 34 quality indicators yields signifi­cant financial rewards for participating providers, and failure to follow them adequately leads to decreased compensation. While it will be several years before conclusions can be drawn from even this first demonstration, CMS did release a round of tentative first-year results in May, showing an aver­age increase in the composite quality score of par­ticipating hospitals from 79 percent to 86 percent.[23]

While seemingly impressive, however, these pre­liminary results do not constitute evidence of improved quality, but rather of increased compli­ance. The composite quality score is merely a mea­sure of the percentage of the time that hospitals followed treatment instructions in pursuit of a financial bonus.

Triumph of Process. Although the score does contain two components, a process score and an outcome score, it is heavily weighted toward rec­ommended processes, not outcomes. Of the 34 quality indicators in the hospital demonstration, 27 measure compliance with dictated processes, while only seven measure outcomes like mortality or readmission rates.[24]

The Physician Group Practice Demonstration, which began in April of this year and is the first major Medicare pay-for-performance initiative to concentrate on physicians, similarly focuses on pro­cess over outcome. According to the demonstration design report, “The major focus of the demonstra­tion will be on measuring process indicators of quality. They are the indicators most easily mea­sured.”[25] They may be the most easily measured, but are they the most indicative of quality care?

Thus, despite the relative lack of objectivity and flexibility, Medicare’s demonstrations rely primarily on process mandates rather than outcome measures to implement pay for performance. David M. Eddy, Senior Advisor for Health Policy and Management at Kaiser Permanente, acknowledges that “a process measure, by its very nature, micromanages. Instead of leaving plans free to set their own priorities for improving health outcomes, a process measure tells plans precisely what their priorities should be.”[26] Even if outcomes were used exclusively, the process of choosing which outcomes to include and decid­ing the relative importance of each outcome would necessitate deference to the values, perspectives, and agenda of the policymaker.

Congressional Action. Bipartisan congressional efforts are now underway to move Medicare beyond demonstrations into a full-scale pay-for-performance reimbursement system. CMS Admin­istrator Mark McClellan appears determined to move ahead with a Medicare pay-for-performance expansion under the regulatory prerogatives of his agency. He is making the effort one of his top prior­ities and recently speculated that within the next five to 10 years, performance-based compensation could comprise up to 30 percent of the govern­ment’s payments to providers.[27]

Any national Medicare pay-for-performance payment system will likely be similar to the CMS demonstration initiatives, emphasizing process over outcome and rewarding compliance with centrally defined practice guidelines. Preliminary reports indi­cate that the proposed legislation would pay for compliance-based bonuses by withholding up to 2 percent of regular reimbursements from all physi­cians.[28] In other words, most doctors would receive less so some could receive more. This is not quality-based compensation; it is redistribution of income toward those providers who subordinate their judgment and creativity to the mandated pro­tocols most successfully.

Because of Medicare’s vast size, the change would reverberate throughout the private payer system, compounding its effect. With almost 42 million enrollees and $290 billion in annual expen­ditures, which will dramatically increase in 2006 when the Medicare Part D drug entitlement takes effect, Medicare is the largest purchaser of health services in the United States.[29] In an open letter published in Health Affairs, Medicare pay-for-per­formance advocates point out that “a major initia­tive by Medicare to pay for performance can be expected to stimulate similar efforts by private pay­ers, just as Medicare’s adoption of prospective pay­ment for hospitals did two decades ago.”[30]

Problems of the Medicare Pay-for-Performance Project

Before launching Medicare into a pay-for-perfor­mance program, Congress should consider the problems that will necessarily arise out of a Medi­care payment system that requires adherence to centrally defined protocols. Congress cannot safely ignore these difficulties.

Problem #1: Limitations of Evidence-Based Medicine

The gold standard of evidence in evidence-based medicine is a combination of double-blind, ran­domized, controlled trials and a systematic review of medical studies called meta-analysis. Although it is conceptually attractive because of its appeal to statistics, no evidence supports overriding the treating physician’s medical decisions with RCT-based guidelines issued by a third party who has never even seen the patient. In 2004, the Agency for Healthcare Research and Quality (AHRQ) in the U.S. Department of Health and Human Services reviewed the literature related to the efficacy of EBM-based compensation, which it referred to as “quality-based purchasing.” The review found “only nine randomized controlled trials” and con­cluded that “little unequivocal data” supported this approach.[31]

The premise of RCTs as objective verification of “best practice” encounters several other serious conceptual problems. First, RCTs can conflict with one another. In July 2002, scientists conducting the Women’s Health Initiative found that Preempro, a hormone replacement therapy drug, had risks of heart attacks exceeding its benefits. These results directly contradicted the results of several other previous and ongoing RCTs, which showed a reduced risk of heart disease associated with the drug.[32] According to Dr. R. Brian Haynes, chief of the Health Information Research Unit at McMaster University and one of the originators of EBM, “It is difficult to be smug about the superiority of the research methods advocated by EBM when the results of studies that are similar methodologically not infrequently disagree with one another.”[33]

EBM’s usefulness as the primary benchmark for treatment decisions is also questionable because RCTs can address only limited medical issues. Despite the massive amount of medical literature published every year, legitimate RCTs cover only a small number of conditions and procedures. Almost all are conducted over only a few months or years, leaving the long-term consequences of a therapy undetected.[34] According to Dr. Nuala Kenny, founder of the Department of Bioethics at Dalhousie University in Canada, “Scientific data cannot be expected to guide most medical deci­sions directly. There are not enough randomized trials or epidemiologic studies.”[35] To develop guidelines, developers often must depend upon some studies conducted on relatively small and unrepresentative populations. As a result, explains Dr. Alan M. Garber of Stanford University’s School of Medicine, “Guideline authors nearly always extrapolate to groups that were not adequately rep­resented in the trials.”[36]

Interestingly, early proponents of evidence-based medicine understood its weaknesses and never meant for it to be more than one factor in a “multi-faceted clinical decision-making decision process.”[37] According to Dr. Haynes:

[E]vidence from research can be no more than one component of any clinical decision. Other key components are the circumstances of the patient (as assessed through the expertise of the clinician) and the preferences of the patient. Just how research evidence, clinical circumstances, and patients’ wishes are to be combined to derive an optimal decision has not been clearly stated, except that clinical judgment and expertise are viewed as essential to success.[38]

Thus, research evidence is just one of three deci­sion points in the treatment process. The other two, to which “applied research is a complementary way of knowing,” are clinical circumstances and patient preferences.[39] By using financial incentives to drive compliance with clinical algorithms, a Medicare pay-for-performance scheme would devalue these two aspects of medical decisions in favor of a pre­scribed list of procedures. In other words, it would produce the very “cookbook medicine” that those who conceived of EBM have denounced as a “mis­use of evidence based medicine.”[40]

Problem #2: Dangers of Replacing Patient Choice and Physician Autonomy with Central Planning

As Dr. David M. Eddy has observed, “It is not stretching things too far to say that whoever con­trols practice policies controls medicine.”[41]

Since clinical research produces conflicting, questionable, and limited evidence, the decisions required to reconcile results, assign relative impor­tance, and sift out bad research cannot be made without subjective human judgment. Ultimately, someone’s values are reflected in treatment deci­sions, whether those values are those of the patient, the physician, or a third party. Bias is inherent in such judgments.

Bias is found in the production, interpretation, and application of pay-for-performance quality indicators. Researchers exhibit bias when deciding which areas of research to pursue, which previous research to reference, and how to conduct their experiments. Journal publishers exhibit bias when deciding which research to publish. Published research is then subject to interpretation by guide­line creators, who exhibit bias when choosing which research to incorporate in their guidelines, resolving conflicting results, assessing research flaws, and transforming findings into rules that weight competing priorities.

According to a 2004 Institute of Medicine report, “There are gaps and inconsistencies in the medical literature supporting one practice versus another, as well as biases based on the perspective of the authors, who may be specialists, general practitio­ners, payers, marketers, or public health officials.”[42] Unable to avoid bias, pay for performance cannot live up to the standards of objectivity claimed by those who support imposing their version of “doing the right thing” on the medical community.[43]

Members of Congress need to ask themselves this crucial question: Whose judgment and values do we want to control important decisions about our med­ical care? Pay for performance would give control to third-party insurance or government managers, who have no information about the unique conditions, health history, preferences, and personal values of the individual patient being treated by the individual doctor. The protocols of the distant government agency would in practice overrule the doctor’s medi­cal judgment and the patient’s choices.

The impact of this control shift would be govern­ment micromanagement of medical care and a cor­responding reduction in physician autonomy and patient choice.[44] Physicians would be compelled either to follow government treatment guidelines or to suffer financial consequences, regardless of whether a particular guideline is in the best inter­ests of a particular patient.

Twila Brase, president of the Citizens Council on Health Care, warns that adopting a pay-for-perfor­mance program “will lead to a limited list of approved health care services—‘best practices’ as determined by the agendas and values of a small cadre of politically motivated, personally-biased individuals sitting around a table somewhere mak­ing treatment decisions far from the patient’s bed­side.”[45] These “elite” decision makers would not even know the name of the patient, much less the patient’s unique circumstances and values.

When made from a distance by budget-focused technocrats, treatment decisions are apt to focus as much on rationing as they do on quality improve­ment. Keith Syrett, professor of law at the Univer­sity of Bristol, observes that “decision making by guideline offer[s] a means of scientifically depoliti­cizing the rationing debate.”[46] Payers are able to create the impression that there is scientific legiti­macy behind cost-based decisions to restrict patient access to medically necessary treatments.

The American Medical Association expressed concern over such potential EBM-based rationing at its national meeting in June 2005. “There is a potential concern when there is another intent behind pay for performance,” said AMA Trustee John Armstrong, M.D. “Some so-called pay-for-performance programs are a lose/lose proposition for patients and their physicians, with the only benefit accruing to health insurers.”[47] As with any health care system that employs top-down plan­ning to ration health care, the ultimate effect will be to limit access to appropriate health care services according to priorities imposed by a centralized bureaucracy.

This should be considered in the context of med­ical malpractice. It is bad policy to empower a cen­tralized bureaucracy to construct and determine which clinical algorithms are to be used and which practices are “best” while also excusing it from any “responsibility for the clinical consequences.”[48] In reality, the ultimate responsibility remains with the treating physician, whose best medical judgment may dictate proceeding in one direction while the third-party algorithm forces the physician to go in another. This could produce adverse results for the patient, and it would further complicate the medi­cal malpractice crisis that is deepening in many states of the union.

Problem #3: Undermining of Personalized Care by Population-Based Medicine

Health care providers treat individual patients, not statistically significant groups. What may be the best treatment for the group on average might not necessarily be the appropriate treatment for an individual patient. Enforcing uniform clinical guidelines on patients whose conditions and values are anything but uniform is like trying to dress everyone in average-sized clothes regardless of their particular sizes and preferences.

Medical treatment decisions depend on a combi­nation of factors—such as age, ethnicity, genetic background, severity of disease, comorbitidies,[49] and patient values—which physicians must incorporate into their evaluation of a patient’s treatment options. The Medicare population is especially heterogeneous because of the prevalence of multiple illnesses in the elderly. Twenty percent of Medicare beneficiaries have five or more chronic conditions, and 50 percent are receiving five or more medications.[50]

Accordingly, Congress should question the appro­priateness of using financial incentives to impose population-based clinical results on individual patient care. One study of 1,755 Type 2 diabetics found a 24 percent rate of noncompliance with clin­ical guidelines by internists. Far from finding a defi­ciency in quality, the researchers found instead “a deficiency in the definition of what constitutes best practices.”[51] Among the clinical guidelines were periodic retinal eye exams and urine protein screens for microalbuminuria. Many physicians questioned the usefulness of the urine screen for patients already receiving ACE inhibitors, the indicated therapy for microalbuminuria. In another example of “noncom­pliance,” some patients did not receive eye exams because they were blind. The authors concluded: “Our data suggest that failure to follow guidelines is not necessarily explained by ‘bad doctors’ or forget­fulness; rather, noncompliance may reflect valid questions about the usefulness and applicability of a best practice to an individual patient.”[52]

In addition to limiting a doctor’s ability to act in the best medical interest of an individual patient, one-size-fits-all clinical guidelines ignore the role of patient preferences and values in health care deci­sions. This is especially relevant to the Medicare population. Evidence shows that elderly patients with multiple conditions vary widely in their prefer­ences regarding longer survival, disease prevention, quality of mental and physical functioning, level of inconvenience and pain, and risk of complication.[53]

A system of standardized treatment decisions is simply incompatible with the variability found in medicine. Even if it were possible to create evi­dence-based rules for every possible variation of patient characteristics and conditions, the sheer number of guidelines would be overwhelming and impossible to implement. Medical studies provide useful information about treatment options that may or may not work in a given situation, depend­ing on the unique combination of circumstances involving the individual patient. However, the very nature of the statistical process, which minimizes bias and seeks mean tendencies, makes it insuffi­cient to make the final decision for every patient. Mathematical models or algorithms cannot capture that precise moment of human choice that is the essence of medical judgment.

Physicians must be free to use their best medical judgment to make clinical decisions that incorpo­rate all relevant factors and available evidence. Appropriate medical decisions cannot be made using an assembly-line mentality that treats every patient like a generic commodity traveling down a conveyer belt.

As expressed by Dan Mendelson, president of the Health Strategies Consultancy in Washington, D.C., “Patients expect their doctor to tailor care to their individual condition, incorporating their medical history and preferences, the doctor’s expe­rience with similar patients, the most current research, and alternative therapies.”[54] To do other­wise not only does a disservice to the patient, but also can ultimately increase morbidity and mortal­ity, which in turn can increase the cost of Medicare.

Problem #4: Deterioration of Clinical Judgment and Medical Innovation

The fourth problem with a pay-for-compliance health care system is the deterioration of creativity, innovative ability, and medical judgment that will occur in an environment that devalues such quali­ties. With compensation hinging on adherence to guidelines, providers will become highly skilled at adhering to guidelines. In anticipation of a Medi­care pay-for-performance shift, the growing “Medi­care industrial complex” of lobbyists, lawyers, consultants, and professional “experts” who make a living deciphering and explaining the Medicare bureaucratese have already swung into action. Companies are already advertising “Pay-for-Perfor­mance Prep Guides,” containing 400 pages of strat­egies to “ensure [that] your practice/organization succeeds with P4P [pay for performance].”[55]

As doctors treat the practice of medicine as if it were an SAT exam, with right and wrong answers and grades handed out by the government, their ability to be flexible, innovative, and discerning in patient care will suffer. Focused on the specific tasks that are linked to financial rewards, automatic practitioners of government-prescribed behaviors will replace doctors who are skilled in combining multiple sources of knowledge with their best med­ical judgment in providing patient care.

Medical students, interns, and residents will become trained in applying the third-party rules that govern their clinical decisions instead of devel­oping keen clinical judgment and learning to con­stantly seek better ways to treat patients.[56] Proponents claim that this will “structure the envi­ronment in which care is delivered so that ‘doing the right thing’ becomes automatic.”[57]

However, the “right thing” can differ from patient to patient and often changes over time. A 2001 study published in the Journal of the American Medical Association found that of 17 clinical prac­tice guidelines published by the Agency for Healthcare Research and Quality and still in circu­lation at the time, seven were in need of a major update, six required a minor update, and only three were determined to be valid. No conclusion was reached for the remaining guideline. Using survival analysis, the researchers found that about half of the guidelines were outdated in 5.8 years.[58] A health care system that ties reimbursement to sometimes outdated or low-quality guidelines would, at least occasionally, force providers to choose between financial compensation and their ethical duty to provide high-quality care. Even worse, it could produce physicians who do not know the difference.

Doctors reimbursed according to compliance will also lose the incentive and ability to innovate that has produced so many important medical advances. In the late 1960s and 1970s, Dr. Charles Kelman, an innovative ophthalmologist, chal­lenged the entrenched “best practices” of his day and pioneered groundbreaking new methods of cataract surgery despite severe derision from colleagues. Dr. Kelman’s innovative techniques revolutionized the field of cataract removal and ultimately became the standard by which all cataract surgeries are per­formed today. In fact, many consider him to be one of the greatest medical innovators because of the mira­cle that he wrought with cataract surgery and the mil­lions of patients who have benefited from it.[59]

If government-instituted compliance mechanisms had been in place in the 1960s and 1970s, however, Dr. Kelman would have dared such innovation only at his own financial peril, because any deviation from the guidelines would have lowered his pay-for-performance scores. Unless we believe that today’s medical practice has reached a state of perfection and is unlikely to be improved, it seems shortsighted to discourage similar advances in the future.

In the long run, automatic adherence to proto­cols is counterproductive. Physicians need to respond effectively to a changing medical world and unique patient challenges, but compliance-based payment systems would deprive them of the very ability to judge appropriate care and adapt with innovative methods of treating illness. Ulti­mately, the political negotiations of bureaucrats and statisticians would replace the medical judgment of individual doctors and remove their incentive to do anything more than what is expressly required to earn their reward.

Problem #5: Poor Quality Because of Unproductive Gaming Behavior

In a letter published in the New England Journal of Medicine, Roy B. Verdery, Ph.D., M.D., claimed:

Economic incentives are always subject to “gaming,” inappropriate manipulation of data, and “cherry-picking” of patients.… Most physicians (and other professionals) work for rewards that are more important than money, including the respect of their patients and peers and the personal satisfaction of a job well done.[60]

By diverting the focus of doctors and other med­ical professionals from appropriate patient-cen­tered medical care to superficial financial rewards, pay for performance will likely create incentives to game the system in several detrimental ways that may cause real quality to decline even while mea­sured indicators are improving.

First, basing financial compensation on specific indicators leads to adverse selection. In other words, providers will tend to select relatively healthier patients who have a higher probability of complying with physician orders, achieving better outcomes, and thus improving the provider’s bottom line.

Two recent studies on cardiologist report cards in New York illustrate this concern.[61] The first study looked at more than 80,000 patients from New York and Michigan and found that doctors in Mich­igan, which does not issue report cards, were more likely to perform angioplasties on very sick patients. The second, published in Archives of Inter­nal Medicine, found that approximately 80 percent of New York cardiologists said that the system made them less likely to treat severely ill patients.[62] If selection is such a significant concern when information about physician performance is merely reported, one can imagine the impact when it is directly tied to compensation.

Compliance-based compensation could also encourage providers to falsify records to circum­vent the system and provide needed care. One study found that 39 percent of physicians already falsify insurance records to secure needed services for patients.[63] Government-endorsed, standard­ized medicine would magnify this problem.

Finally, if Congress ties money to specific medi­cal interventions, doctors and other medical pro­fessionals will be pressured to focus on those interventions to the detriment of other important areas of medicine. Pay for performance’s very premise is that financial incentives alter behavior. While very few doctors will allow them to com­pletely consume their behavior, the nature of finan­cial incentives will push them in certain directions. As a result, some conditions and some procedures will receive less than adequate attention.

“Inevitably…the dimensions of care that will receive the most attention will be those that are most easily measured and not necessarily those that are most valued,” according to a recent study in Health Affairs on provider incentives.[64] Mitigating this problem would require increasing the number of required measurements, which would soon become overwhelming and counterproductive, even if there existed the possibility of doing so adequately. Thus, while adherence to measured indicators might show improvement, overall quality might become worse.

Instead of adjusting their behavior to a set of standard rules and guidelines, health care provid­ers should be encouraged to meet their patients’ needs and preferences in a comprehensive way. This cannot be accomplished in an arbitrary system of compliance-based incentives that encourages providers to manipulate the “game” to their finan­cial advantage.

Problem #6: Further Weakening of the Doctor–Patient Relationship

“One major barrier to the adoption of EBM,” according to analysts writing in Health Affairs, “is the overwhelming support for preserving the phy­sician–patient relationship.”[65] A Medicare pay-for-performance system would lead to a decline in this relationship. When patients understand that their physicians are being pressured to meet standard­ized treatment directives rather than to provide them with customized care based on their unique conditions and preferences, trust in their physi­cians will be compromised—and for good reason.