The overall quality of health care delivered to Americans is worse than it should be. While many quality improvement efforts have been undertaken, their success has been limited by current payment systems. The existing systems do not reflect the relative value of health care services in important aspects of quality, such as clinical quality, patient-centeredness, and efficiency. Nor do current payment systems recognize or reward care coordination, an omission reflected in such shortcomings as the limited focus on prevention and the treatment of chronic conditions as patients move across various care settings. Fundamental changes in approaches to health care payment are necessary to remove impediments to and create incentives for significant quality improvement.
The Institute of Medicine (IOM) report Crossing the Quality Chasm: A New Health System for the 21st Century made the case for changes in the health care system, including restructuring of payment methods, to close the quality gap. Five years later, however, the concerns raised in that report persist. The report identified six aims for health care that should guide quality improvement efforts—safety, effectiveness, patient-centeredness, timeliness, efficiency, and equity—and noted that payment systems supporting the organization and delivery of the nation’s health care services do not align incentives to support the realization of these aims. Instead, current payment policies reinforce the existing organizational structure and delivery processes of the American health care system by paying according to the number and complexity of services by setting rather than recognizing the relative value of those services. New payment incentives must be created to encourage the redesign of structures and processes of care to promote higher
value. Although the magnitude of incentives necessary to achieve significant and sustainable change while avoiding adverse consequences is uncertain, steps can be taken now to begin to address the deficiencies of current payment systems and encourage progress toward significant quality improvement.
STUDY CHARGE AND SCOPE
This study is the third in the IOM’s Pathways to Quality Health Care series, which offers tools for implementing the vision of improved health care delineated in the Quality Chasm report. The first report in the Pathways series, Performance Measurement: Accelerating Improvement, recommended a strategy for developing and implementing a comprehensive performance measurement system, including creation of a national board to coordinate that effort. The second report, Medicare’s Quality Improvement Organization Program: Maximizing Potential, recommended an emphasis on technical assistance to providers for quality improvement. The present report builds on those studies and offers an operational plan for introducing into Medicare payment incentives that would encourage and reward high-quality care. While alignment can occur in many areas, this report is limited to examining the link between payment incentives and provider performance.
In the context of current efforts to test pay for performance in both the public and private sectors, the U.S. Congress, as part of the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 (Public Law 108-173, Section 238), directed the IOM to identify and prioritize options for aligning performance with payment in the Medicare program under Title XVIII of the Social Security Act (42 U.S.C. 1395 et seq.). The congressional mandate identified three topics for the study to address:
The performance measure set to be used and how that set should be updated.
The payment policy that should be used to reward performance.
The key implementation issues involved, such as data and information technology requirements.
In response to this mandate, the IOM Committee on Redesigning Health Insurance Performance Measures, Payment, and Performance Improvement Programs explored the design and implementation of payment rewards for performance in Medicare. It considered both specific topics involved in introducing pay for performance within Medicare and the implications of using this payment approach as part of a long-term multipayer effort to better align the health care system with a vision of quality.
PAY FOR PERFORMANCE: AN IMMEDIATE OPPORTUNITY
The objective of aligning incentives through pay for performance—paying providers for higher-quality care as measured by selected standards and procedures—is to create payment incentives that will:
Encourage the most rapid feasible performance improvement by all providers.
Support innovation and constructive change throughout the health care system.
Promote better outcomes of care, especially through coordination of care across provider settings and time.
Pay for performance is not simply a mechanism to reward those who perform well or to reduce costs. Its purpose is to align payment incentives to encourage ongoing improvement in a way that will ensure high-quality care for all. Pay for performance will not necessarily reduce the cost of care, but it will help ensure that what is paid for will be more beneficial to patients. In theory, payment incentives induce certain predictable responses or behaviors. The notion that paying more for some attribute of a good or service will stimulate further production of that attribute is fundamental in most sectors of the economy, but the explicit linkage of incentives to quality and performance in health care markets is a relatively new concept. Therefore, introducing payment incentives to reward high quality in a national health care program requires attention to effects on providers, purchasers, health plans, and consumers.
More than 100 reward and incentive payment programs have been launched in the private health care sector. Most of these efforts have not yet been fully evaluated. The Centers for Medicare and Medicaid Services (CMS) in the U.S. Department of Health and Human Services (DHHS) has also initiated a series of demonstration projects to explore the potential of pay for performance to achieve quality improvement in Medicare. Some of these programs have begun to show that providers respond positively to payment incentives that promote and reward quality improvement practices, but it remains unknown whether the improvements seen will be significant and sustained. The literature evaluating the effectiveness of pay for performance consists of fewer than 20 studies, yielding mixed conclusions on overall impact. Some studies have shown a positive effect on the quality of care, but others have not demonstrated this relationship. In general, the effect of most of these programs has not been examined sufficiently. Despite these uncertainties, however, Medicare payment systems, if left unchanged, will pose a barrier to improved health care quality.
MEDICARE AND PAYMENT INCENTIVES
Medicare, the government health program for the elderly (ages 65 and over) and qualified disabled populations, covered nearly 42 million Americans in 2004. Medicare is the nation’s largest single payer for health care services, with total expenditures of $309 billion in 2004, and this amount is estimated to grow rapidly. Although Medicare is federally administered and largely federally financed, its beneficiaries are served almost entirely by private providers. For services provided to the 88 percent (approximately 37 million) of beneficiaries enrolled in the traditional fee-for-service option, Medicare pays providers amounts that are set prospectively on the basis of resource cost and complexity of services delivered. For the remaining 12 percent of beneficiaries who have opted to receive their Medicare services from private plans participating in the Medicare Advantage program, Medicare pays a fixed, risk-adjusted monthly amount per beneficiary to the plans, which in turn pay providers via diverse methods.
The current Medicare fee-for-service payment system is unlikely to promote quality improvement because it tends to reward excessive use of services; high-cost, complex procedures; and lower-quality care. Through bundled and prospective payment arrangements for institutions, Medicare has attempted to create incentives for efficiencies, but significant price and payment distortions persist.
Services that contribute greatly to high-quality care that are labor- or time-intensive and rely less on technical resources, such as patient education in self-management of chronic conditions and care coordination, tend to be undervalued and are not adequately reflected in current payment arrangements. Little emphasis is placed on efficiency (achieving high clinical quality with a given amount of resources). The lack of incentives for comprehensive, coordinated care discourages services targeting early intervention and prevention that can ultimately reduce the use of expensive services, such as avoidable hospitalizations. Providers often miss opportunities for collaboration since the payment system rewards neither team management nor the integration of services across care settings. Medicare’s fee-for-service payments, the relative profitability per service, and most private purchasers’ payment mechanisms create incentives for providers to specialize in fields that are more resource-intensive at the expense of primary care, which has not fared well under the current Medicare reimbursement systems.
Aligning payment incentives with quality improvement goals represents a promising opportunity to encourage higher levels of quality and provide better value for all Americans. However, pay for performance needs to be closely monitored because it could have unintended adverse consequences, such as decreased access to care, increased disparities in care, or impedi-
ments to innovation. Careful monitoring of implementation should minimize any such adverse consequences. The committee thus reached two key conclusions regarding pay for performance as a new payment strategy for Medicare:
The systematic and deliberate use of payment incentives that recognize and reward high levels of quality and quality improvement can serve as a powerful stimulus to drive institutional and provider behavior toward better quality.
The incentives introduced by pay for performance, by themselves, will not be sufficient to achieve the broad institutional and behavioral changes needed unless certain operating conditions are met, such as the use of electronic health records, public reporting, beneficiary incentives, and education of boards of directors, which could lead to significant and synergistic gains in quality improvement.
INITIAL IMPLEMENTATION: A PHASED APPROACH
While an evidence base is not yet available for determining with certainty what type of payment incentive strategy would best advance the quality improvement agenda, experiences with pay for performance to date have been promising. Consequently, the committee favors an approach that would capture key lessons from these early experiences and maintain the flexibility to make subsequent changes where necessary. Specifically, the committee concludes pay for performance should be introduced through a phased approach: rewarding performance in selected settings, with a small level of funding, on specific measures, and moving eventually to include all provider settings, with a larger level of funding, on more measures. Such a phased approach requires attention to the timing and pace of implementation for specific settings, reward amounts, and measures. A phased approach also provides an opportunity to examine other long-term approaches.
Pay for performance in Medicare can help address current problems and stimulate complementary quality improvement strategies. Indeed, the long-term potential of pay for performance may lie in its ability to encourage the use of mutually reinforcing quality improvement strategies, such as technical assistance, use of information technology, professional certification, public reporting, and provider and consumer education. Pay for performance cannot significantly improve quality and reduce costs in isolation from other changes in the Medicare system, and could in fact pose a barrier to achieving the transformational changes required to improve care if implemented in ways that would reinforce the current fragmented delivery system. The hope is that payment incentives can offer a stimulus to move health care practices overall from the status quo toward new organizational and individual behaviors that will result in better quality of care.
Recommendation 1: The Secretary of the Department of Health and Human Services (DHHS) should implement pay for performance in Medicare using a phased approach as a stimulus to foster comprehensive and systemwide improvements in the quality of health care.
Achieving the promise of pay for performance to recognize and reward quality in Medicare requires answer to questions about several key design features, including:
The sources of revenue for rewards.
The types of performance measures that should receive preferential treatment in the early stages of implementation.
The appropriate design of the reward system.
The performance measurement framework will have to be sophisticated and nuanced to account fairly for complex clinical situations, such as the treatment of patients with multiple chronic diseases, in which the accepted care for one condition may be in conflict with that for another. Complex measures to address concerns about efficiency and patient-centeredness will require attention. Providers in different institutional settings (e.g., hospitals or skilled nursing facilities), diverse organizational environments (e.g., managed care or solo practices), and different specialty fields will need new capabilities (e.g., databases, information tools, and technical assistance) to comply with new reporting and payment procedures. If payment incentives are not carefully aligned with desired outcomes or if adequate resources or risk adjustments are not readily available, some providers may avoid accepting patients whose conditions could jeopardize their performance rating.
Recognizing the complexities of current circumstances, as well as the demand for action, the committee emphasizes the importance of introducing pay for performance not only through a gradual and phased approach that varies by setting, amount of reward, and measure, but also within a learning system that can evaluate experience with early efforts. Caution must be exercised to ensure that the proposed phased approach does not widen current gaps in performance among providers and domains of care. A learning system depends on monitoring and evaluation and collaboration between the private and public sectors that enables all stakeholders to learn from experience. Ultimately, major restructuring of basic Medicare payment systems beyond the incentives discussed in this report will be necessary. Such restructuring, which could require a transformation away from fee-for-service payments, could include elevating the value of integrated care management, relying more heavily on electronic health records, and facilitating payment that rewards high performance and coordination of services across care settings. Because measures and information systems needed to
monitor both meaningful outcomes of care and the health status of patients across different care settings are not yet available, this shift would have to occur in the future. Further research will be necessary to develop benchmarks that can guide the process of phased implementation and restructuring of payment arrangements.
FUNDING OF PAY FOR PERFORMANCE
There are three potential sources of funding for a pay-for-performance program in Medicare: (1) existing funds, (2) generated savings, and (3) new money. Combinations of these sources are used in current pay-for-performance experiments, though new money is rarely agreeable to payers. Existing funds represent monies that are already projected to be part of the payment system. Payment incentives could be financed by reducing the base payments of all providers or by reducing scheduled payment increases. These funds would then be awarded to high-quality providers. Alternatively, a portion of each payment could be withheld, with the balance returned to those providers who achieved quality goals. However, rewards would initially be small; for example, modeling reductions in all payments by 2 percent for three clinical conditions resulted in physician rewards averaging $88 per physician per year (assuming that half of treating physicians would qualify for rewards). The generated-savings model creates a reward pool through cost-reducing reforms and efficiencies; however, these efficiencies have not yet been adequately demonstrated in pay-for-performance efforts. The new-money model taps the Medicare Trust Funds or calls for a separate appropriation of general revenues that would be awarded as bonuses to high-quality providers in addition to the scheduled base payments and updates all providers receive.
The committee used four criteria to assess the appropriateness of these three possible funding sources: adequacy, stability, fairness, and impact. In addition, the committee gave overall priority to funding approaches that would be budget conscious (or, preferably, budget neutral), ensuring that budget concerns would be explicitly recognized and addressed.
Recommendation 2: Congress should derive initial funding (over the next 3–5 years) for a pay-for-performance program in Medicare largely from existing funds.
Congress should create provider-specific pools from a reduction in the base Medicare payments for each class of providers (hospitals, skilled nursing facilities, Medicare Advantage plans, dialysis facilities, home health agencies, and physicians).
Congress should ensure that these pools are large enough to create adequate motivation for improved performance on selected measures. Because of unique challenges of physician payment relating to the sustainable growth rate (SGR), investment dollars may be necessary to create adequate resources to effect change.
Initial funding should be budget conscious in taking into account the resources needed for both funding the pools and implementing the program.
Because the proposed pools would be created by reducing base payments for all Medicare providers in each setting, all should have the opportunity to participate in the performance reward program. New money may initially be necessary in some provider settings to create adequate resources to influence change. The feasibility of using other funding sources, particularly the generated-savings model, should be tested and evaluated over the next 3–5 years to assess the likely impacts and consequences.
One of the primary goals of new payment incentives should be to stimulate collaboration and shared accountability among providers across care settings for better patient-centered health outcomes. Although the implementation of pay for performance will most likely begin with pools created by setting, CMS should build toward an ultimate vision of aggregating funds for rewards into one integrated pool that would accommodate shared accountability and encourage coordination of care.
Recommendation 3: Congress should give the Secretary of DHHS the authority to aggregate the pools for different care settings into one consolidated pool from which all providers would be rewarded when the development of new performance measures allows for shared accountability and more coordinated care across provider settings.
STRUCTURE OF REWARDS
CMS will have many issues to consider in the distribution of rewards, such as what measures to use in assessing performance and how performance should translate into rewards for individual providers. The magnitude and relative distribution of rewards should depend on program priorities; little evidence exists to guide the distribution of rewards. Continuous monitoring and adjustment will be necessary to ensure that providers are appropriately rewarded for the care they deliver. In the absence of evidence, the committee provides recommendations in two key areas: rewards for specific domains of quality and performance objectives.
Rewards for Specific Domains of Quality
The ultimate goal of pay for performance is to improve quality and patient outcomes. However, current capabilities focus largely on measuring processes of care. Many providers are skeptical that reliable and valid performance measures can be introduced for complex clinical processes. They are also doubtful that incentives can be instituted to reward performance in areas that truly matter—those necessary to improve the health and care of their patients. A major challenge confronting the introduction of pay for performance, therefore, is overcoming the fear that efforts to improve upon one domain of performance may lead to reductions in quality in other domains. Under a new payment mechanism, for example, improved efficiency may greatly benefit the overall quality of the system, yet more important, it may also compromise clinical quality or patient-centered care. Similarly, many purchasers and public officials are concerned that focusing on enhancing clinical quality or patient-centered care will not adequately address concerns about the growing costs of health services or reduce current waste and inefficiencies.
To create new payment incentives that can foster overall quality improvement and better patient outcomes, the committee consolidated the six quality aims of the Quality Chasm report into three domains—clinical quality, patient-centered care, and efficiency. Eventually, if pay for performance is found to have positive effects, other aspects of care should also be measured and rewarded.
Recommendation 4: In designing a pay-for-performance program, the Secretary of DHHS should initially reward health care that is of high clinical quality, patient-centered, and efficient.
Two categories of performance benchmarks deserve consideration in designing a payment incentive program: (1) improvement—rewarding all providers who demonstrate significant improvement, and (2) excellence—rewarding those providers who meet or exceed a recognized threshold of desired quality. Current private-sector pay-for-performance programs have reward structures that utilize one or both of these categories.
Recommendation 5: The Secretary of DHHS should design a pay-for-performance program that initially rewards both providers who improve performance significantly and those who achieve high performance.
The distribution of rewards for both improvement and excellence would allow providers at all performance levels to find at least one of these goals within reach. The fraction of rewards allocated to improve on a given measure set should be reduced over time to only reward care that is truly of high performance. As providers make significant improvements on basic measures, the allocation of rewards for those measures should shift in favor of higher payments for more complex indices of performance. Rewards should also shift to reflect progress in the development of new measure sets and changes in priorities. Even when measure sets or priorities are stable, the focus of rewards should be altered to ensure that providers do not focus their performance improvement efforts too narrowly.
CMS will also need to address procedural and technical issues, including the following:
The procedures by which comparative information on provider performance will be released to the public.
Ways of overcoming the barriers to participation.
The process of improving care coordination among providers serving the same patient.
The role of information technology use in supporting better care delivery and a performance-based payment strategy.
Public Reporting and Transparency
Beyond merely collecting data on provider performance, CMS should make such data publicly available so that consumers will have the opportunity to fully characterize the performance of providers when making health care decisions. Public disclosure of information, with necessary patient protections, can also stimulate higher levels of quality by showing providers how their performance compares with that of their peers. While the evidence remains mixed, peer comparisons may be a more powerful force than monetary incentives in encouraging providers to adopt practices that improve quality of care. However, payment incentives are necessary as a key stimulus to foster widespread public reporting.
Recommendation 6: Because public reporting of performance measures should be an integral component of a pay-for-performance program for Medicare, the Secretary of DHHS should offer incentives to providers for the submission of performance data, and ensure that information pertaining to provider performance is
transparent and made public in ways that are both meaningful and understandable to consumers.
The committee proposes that public reporting requirements precede changes in payment strategies to allow time for providers to give feedback on performance results and comparisons. To advance the pace of adoption, incentives should be offered for the submission of performance measurement data that contributes to public reports. Public reports can inform consumer choices only if they are presented in a manner that is meaningful and easily understandable. Over time, payment incentives for the submission of routine data should be phased out so that this pool of funds can be redirected to the development of measures for areas that are more difficult to assess.
Ways of Overcoming Barriers to Participation
In its deliberations, the committee recognized the importance of establishing the expectation that all Medicare providers would participate in public reporting and pay for performance. However, it also recognized that the pace of implementation, the breadth of measure sets applicable to specific types of providers, and the size and distribution of reward pools would need to vary depending on the availability of measures and the organizational and technological challenges faced by different providers in carrying out performance measurement and reporting.
Many types of Medicare providers, including hospitals, home health agencies, and Medicare Advantage plans, are already submitting performance data for public reporting. For these providers, CMS should begin pay-for-performance programs on existing measures immediately, and move toward comprehensive performance assessment systems and sizable reward pools during the next 3 years.
Although skilled nursing facilities are already publicly reporting data to CMS, the performance measures reflecting their treatment of Medicare beneficiaries are not yet adequate to support pay for performance. There are also currently few, if any, performance measures for other providers, such as clinical laboratories and ambulatory surgical centers. Efforts should begin immediately to develop and test performance measure sets so that these providers can begin to participate in public reporting and pay for performance as soon as possible.
CMS has already begun a voluntary reporting program for physicians on selected measures. CMS should immediately develop and implement a strategy for ensuring that virtually all physicians participate—on at least some measures—as soon as possible. This strategy will need to be sensitive to differences across specialties in the availability of performance measures
and the diversity of information systems and operational supports in various practice settings. Financial incentives adequate to ensure early and broad physician participation in the submission of performance measures and public reporting should be used. Consideration should be given to benefits such as linking accelerated payments or the physician annual payment update to public reporting. Initial measure sets for pay for performance may need to be limited in some physician settings. In establishing the size of the reward pools proposed above, CMS will need to strike a balance between providing financial incentives sizable enough to lead to near-universal participation and recognizing that initial measure sets are narrow, presenting an incomplete picture of a provider’s performance.
The transformational changes in the health care delivery system envisioned in the Pathways series of reports will depend upon the adoption of both longitudinal measures of quality that cut across settings and payment rewards that are substantial. The pay-for-performance strategy should move as soon as practical from a relatively narrow, provider-specific approach to a more comprehensive, longitudinal set of measures and substantial rewards that encompass all Medicare providers.
A monitoring system should be incorporated into the implementation process to inform future decisions about the pace of expansion of performance measure sets and make it possible to determine whether the voluntary approach initially recommended for physicians is achieving the goal of near-universal participation.
Recommendation 7: The Secretary of DHHS should develop and implement a strategy for ensuring that virtually all Medicare providers submit performance measures for public reporting and participate in pay for performance as soon as possible. Initially, measure sets may need to be narrow, but they should evolve over time to provide more comprehensive and longitudinal assessments of provider and system performance. For many institutional providers, participation in public reporting and pay for performance can and should begin immediately. For physicians, a voluntary approach should be pursued initially, relying on financial incentives sufficient to ensure broad participation and recognizing that the initial set of measures and the pace of expansion of measure sets will need to be sensitive to the operational challenges faced by providers in small practice settings. Three years after the release of this report, the Secretary of DHHS should determine whether progress toward universal participation is sufficient and whether stronger actions—such as mandating provider participation—are required.
Rewarding providers on the basis of performance will require that Medicare know which providers delivered care to specific patients. Patients frequently interact with more than one provider, and treatment of complex conditions often requires consultation with multiple specialists. On average, Medicare beneficiaries are treated annually by 5 physicians; beneficiaries with the chronic conditions of chronic heart failure, coronary artery disease, and diabetes see an average of 13 physicians annually. Enhancing care coordination is essential to improving quality.
Recommendation 8: The Centers for Medicare and Medicaid Services (CMS) should design the Medicare pay-for-performance program to include components that promote, recognize, and reward improved coordination of care across providers and through entire episodes of illness. Thus, CMS should (1) encourage beneficiaries and providers to identify providers who would be considered their principal responsible source of care, and (2) pay for and reward successful care coordination that meets specified standards for providers who take on that role.
Not all providers treating Medicare beneficiaries would be willing or able to provide this coordinating function; thus CMS should design a strategy to reward those who are capable of and willing to assume this role. Beneficiaries should be encouraged to designate their responsible sources of care through incentives such as reductions in their Medicare Part B premiums. All such activities should protect patient confidentiality and be carried out in compliance with regulations of the Health Insurance Portability and Accountability Act.
Information technology has enormous potential to be used as a transformative tool in systems change toward improving the quality of health care. Pay for performance can influence the rate of information technology adoption, but information technologies are not a necessary component of pay for performance. While promising, the infrastructure required to automate patient-specific clinical information has not yet fully been embraced. Without clear standards, experimentation will likely continue slowly, in a piecemeal fashion.
Recommendation 9: Because electronic health information technology will increase the probability of a successful pay-for-performance program, the Secretary of DHHS should explore a variety of approaches for assisting providers in the implementation of electronic data collection and reporting systems to strengthen the use of consistent performance measures.
MONITORING, EVALUATION, AND RESEARCH
Monitoring, evaluation, and research should be integral components of any pay-for-performance program. Issues to be addressed include use of current data to evaluate impact; processes for developing robust performance measures; and development of real-time monitoring systems to identify unintended adverse consequences. A successful pay-for-performance program must also encompass the elements of a true learning system, including having strong leadership, a shared vision, and an environment that allows for action in response to observations.
Recommendation 10: The Secretary of DHHS should implement a monitoring and evaluation system for the Medicare pay-for-performance program in order to:
Assess early experiences with implementation so timely corrective action can be taken.
Evaluate the overall impact of pay for performance on clinical quality, patient-centeredness, and efficiency.
Identify the best practices of high-performing delivery settings that should be shared with others to improve care throughout the nation.
This active learning system should be complemented by a research agenda identified through consensus among the major stakeholders to create the context for future investigations as actual experience raises new questions. Research should also be aimed at building an evidence base to guide the design of future pay-for-performance programs.
Collaboration between the public and private efforts is critical. While multiple stakeholder groups are now developing reliable, valid, and accurate performance metrics in the area of clinical quality, these efforts are not coordinated and often produce competing and inconsistent measures that are burdensome to providers.
CMS should conduct demonstration projects to evaluate different options that are theoretically sound but untested. Such projects could limit risks and accelerate progress in payment realignment by confirming benefits and minimizing the risk of undue hardship for beneficiaries or providers.