Is P4P Really FFS?

It is common for societal trends and preferences to be recycled. Men's neckties have varied from thin to wide and back to thin again. Government policies have favored individual responsibility, then social responsibility, and again individual responsibility. In health care, policy makers have swung between concern about access and concern about cost, attempting to increase or decrease medical care supply in response. Until 1980, there was widespread belief in a fairly severe physician shortage in Canada and the United States. By 1990, it was thought that too many physicians were performing too many services. Now, in the 2000s, physician shortages are again the topic of discussion.

In the 1960s and 1970s, health policy makers were concerned about access and institutionalized a system that expanded insurance for patients (thus separating them from the true costs of care) and organized a reimbursement system based on fee for service (FFS) and costs incurred. As access expanded, so did health care expenditures. By the 1980s and early 1990s, health care costs in Canada and the United States were thought to be too high. The view that supply created its own demand was common and best epitomized by Roemer's law: hospital bed utilization reflects supply.¹ In his article on “protecting the commons,” Hiatt² criticized the FFS model as having significant negative externalities for society, providing too much health care at the expense of other social services. A solution was to pay physicians and hospitals a flat amount to provide care through prospective case payments, capitation, or global budgets rather than per service.

The 1990s were filled with reforms that used 2 mechanisms to correct the problem of rising health care costs and “protecting the commons”: (1) managed care, the insertion of third-party management tools that included incentives for reducing care and (2) competition between health care systems and “plans” for patients on the basis of cost and quality. Together these ideas were known as “managed competition,” an approach originally proposed by Enthoven.³

The objectives of managed competition, capitation, and global budgets (supply-side constraints) were to reduce unnecessary care without reducing necessary or appropriate care. In managed care, health care administrators would be able to control which patients received which types of care to increase the proportion of care that was deemed appropriate. Theoretically, managed competition made sense, but it quickly fell out of favor in many circles, particularly among the public, precisely because it was successful at achieving its goal; ie, it effectively reduced the provision of care and kept health care costs from continuing to escalate. Restricting only unnecessary care proved to be quite difficult. The RAND Health Insurance Experiment⁴ provided an analogous example of the relationship between (consumer) cost sharing (a demand-side intervention) and appropriateness. Increased patient co-payments reduced all health care service use, not just ineffective or unnecessary utilization.⁴ It seems that both supply-side and demand-side attempts to control costs provide financial barriers that limit both necessary and unnecessary treatment.

Now, as the pendulum appears poised to swing away from cost concerns and back toward improving access, the Institute of Medicine's Quality Chasm and Rewarding Performance reports have broadened priorities, also focusing attention on improving quality.^{5 - 6} To improve both quality and access, a new idea has been promoted: pay for performance (P4P).

Pay for performance is a catchphrase for a management tool that establishes incentives for clinicians and institutions (eg, hospitals) that deliver health care services to deliver care that third parties deem is necessary and appropriate to achieve the highest-quality standards and best outcomes. (Who could oppose this?) Pay for performance has many proponents and has rapidly diffused through the United States to individual health plans, associations of health plans, Medicaid, Medicare, and Veterans Administration programs. Pay for performance has also been instituted in the United Kingdom, Australia, and Canada. There is considerable heterogeneity in the focus of P4P initiatives, from primary care to hospitals to regional health care systems, and a wide variety of methods are used to design the incentives.⁷ However, there is some concern that P4P programs have gotten ahead of the evidence, as the research to support these initiatives is currently sparse.^{6 ,8}

The P4P initiative has been principally aimed at rewarding clinicians and health care institutions for process-oriented activities, infrastructure improvements (principally information technology), and even patient outcomes. Considering payment systems used by health care practitioners and organizations in the previous century, P4P is akin to a cost-based payment for information technology and FFS payment for specific process-related clinical activities. As such, P4P appears to have the same limitations and offers little advantage over FFS. It is debatable whether P4P programs will deliver on their promises.

Although P4P aims to increase necessary and appropriate care, it does so in the absence of direct clinician or patient input. Similar to managed competition, a third party determines what services are appropriate and should be incentivized to ensure that clinicians and health care organizations expand delivery of those specific services. Are clinical decisions being taken out of the hands of clinicians and patients? Is this a good idea? Measures derived from clinical practice guidelines are not sensitive to individual patient needs or preferences.⁹ Moreover, good care does not equal rigid adherence to clinical guidelines.¹⁰ Care incentivized by P4P is not a reflection of patient preferences unless those preferences can be expressed by choosing between health plans with and without P4P initiatives. Specific or additional payments for care that physicians want patients to have would provide the “right” incentives, but the right care might be context- or patient-specific and, therefore, difficult to measure.

Moreover, lost in much of the P4P discussion is that these programs do nothing to address misuse or overuse of services that may not be necessary or appropriate. In addition, while P4P provides incentives to increase the delivery of measured services, this will happen only if profits can be earned: marginal revenues obtainable need to be greater than marginal costs for each practice.

When viewed in this light, it may be that the health care delivery system is returning to FFS. But third-party payers are trying to be smarter and more specific this time by encouraging the provision of evidence-based health care and incorporating quality improvements. Pay for performance may be more specific, but is it smarter?

The opportunity costs of pursuing P4P initiatives are the return on investment that could be achieved if the resources were used to provide services that are not part of P4P. Is the return on investment for target conditions the highest possible? What else contributes to desired goals? The “elephant in the room” is everything else not being measured. Is a reasonable assumption that what is not being measured is being done well? Doubtful. Will measured care crowd out unmeasured care that is also appropriate and of high quality? Possibly. Will excessively remunerated care, even if unmeasured, ever decrease? Highly unlikely.

It is unclear exactly what P4P is trying to accomplish. The mantra that seems to justify P4P is “value for money.” The money part is easy to quantify, but what about value? To know if P4P delivers value, specific, reliable, and ongoing measurement is essential. Even if payers participating in P4P see changes in clinicians' and health care institutions' measured performance, this does not necessarily equate outcomes or value.^{11 - 12} Even if some guideline-based services are known to improve patient outcomes, P4P can address only one barrier to the adoption of optimal care: financial incentives. But other factors, such as clinician acceptability, patient preferences, and social or ethnic contexts, are not accounted for by this policy instrument. Moreover, the mechanisms that are required to track the process and outcome improvements are quite expensive, perhaps exacerbating the problem of excessive “administrative costs” of health care. An additional concern is that guidelines may be the product of authors who have a potential bias to deliver more expensive care because of their prior financial relationships with those who have much to gain from those guidelines—specifically, pharmaceutical companies.¹³

It is easy to see why for-profit insurance companies (principally in the United States) would embrace managed competition, because the objective of those management tools (reducing the amount of care delivered to patients) align with a profit-making objective. It is more difficult to imagine for-profit insurance companies truly embracing a management tool aimed at increasing care (albeit appropriate care) for its enrollees unless it was also expected to decrease costs. This may underlie the limited success of P4P programs observed in the United States thus far.¹⁴ For P4P to be successful in a for-profit marketplace like the United States, successful implementation would need to lead to a substantial increase in market share, particularly a market-share increase among the “right” patients (ie, those whose expected costs are less than the charged premiums).

Physicians and other clinicians, objectives of care, performance measures, and target levels clearly differ across health care interventions; therefore, it is sensible that the evidence regarding P4P is context- or situation-specific. Concluding that “it works” or that “it does not work” should be rephrased as “it works/does not work here.” Important contextual factors include existing accountability relationships, existing forms of payment, and the clinician's direct control over the performance being rewarded. Because P4P worked in primary care for smoking cessation in the United Kingdom¹⁵ does not mean that it should have worked for acute myocardial infarction in US hospitals.¹⁶

There is variation in P4P implementation, from individual-level incentives to institutional-level or regional-level incentives.⁷ The most effective incentives to improve care practices are most likely delivered at the level of the individual clinician and patient rather than at the level of the hospital or health plan. Clinicians and patients are most proximal to the decision and therefore are likely to achieve the highest impact. Pay for performance must become more transparent and proximal to patient care if it is going to have a beneficial effect.

To improve clinicians' performance in the use of appropriate medications after acute myocardial infarction or the provision of smoking cessation counseling, one approach is simply to use FFS and provide fee codes at the individual health care clinician level to induce that behavior. Rosenthal and Dudley⁷ describe such an approach implemented by the Hudson Health Plan for improving the quality of diabetes care. For example, for smoking cessation, there could be a fee code for the “provision of an effective smoking cessation program,” with an initial payment of $30. The full payment for service cannot be received until outcome ascertainment—in this case, a urine test for biological verification of tobacco abstinence at 1 year; if there is no evidence of nicotine metabolites, there is an additional payment of $120. In another example, for post–acute myocardial infarction care, P4P could use a fee code for “prescribing β-blockers, angiotensin-converting enzyme inhibitors, antiplatelet agents, and cholesterol-reducing medications and referral to a cardiac rehabilitation program for patients following myocardial infarction,” with an initial payment of $50. The outcomes followed up may be reinfarction, rehospitalization, or mortality—whichever are considered most important. Follow-up could be at 30 days or 1 year. If the patient is alive at 1 year, the additional payment might be $150 (the “capitation payment”). In other words, if the goal is to improve health care and health, then outcomes need to be measured.

However, some quality improvements require changes in systems (eg, information technology or organizational changes); these objectives require incentives that are best targeted to entire organizations or health systems.

The adoption of P4P as a health policy initiative may have been premature and does not appear to have been based on clear evidence. This commentary highlights that the pendulum continues to swing and that P4P is really modified FFS payment and is unlikely to achieve a radical improvement in the efficiency of an entire health care system. Nonetheless, FFS may be the best kind of incentive system to make the majority of clinicians and health care institutions deliver desirable care.¹⁷ However, P4P will expand access in areas where the new incentives produce profit for health care clinicians and organizations, perhaps at the expense of other services. Health policy analysts who were critical of FFS in the 1970s because it induced expansion of services may find the same fault with P4P in the 2010s. Those who did not like FFS then are unlikely to like P4P now.