A National Strategy To Put Accountable Care Into Practice
- 1Mark McClellan (mmcclellan{at}brookings.edu) is director of the Engelberg Center for Health Care Reform at the Brookings Institution in Washington, D.C. He is also senior fellow, economic studies; and holds the Leonard D. Schaeffer Chair in Health Policy Studies at Brookings.
- 2Aaron N. McKethan is research director of the Engelberg Center.
- 3Julie L. Lewis is assistant director for health policy, Center for Health Policy Research, at the Dartmouth Institute for Health Policy and Clinical Practice in Hanover, New Hampshire.
- 4Joachim Roski is managing director of the High-Value Health Care Project at the Engelberg Center.
- 5Elliott S. Fisher is director for Population Health and Policy and director at the Dartmouth Institute for Health Policy and Clinical Practice, and a professor of medicine and of community and family medicine at Dartmouth Medical School.
- ↵*Corresponding author
Abstract
The concept of accountable care organizations (ACOs) has been set forth in recently enacted national health reform legislation as a strategy to address current shortcomings in the U.S. health care system. This paper focuses on implementation issues related to these organizations, building on some initial examples. We seek to clarify definitions and key principles, provide an update on implementation in the context of other reforms, and address emerging issues that will affect the organizations’ success. Finally, building on the initial experience of several organizations that are implementing accountable care and complementary reforms, we propose a national strategy to identify and expand successful approaches to accountable care implementation.
The historic enactment of national health reform legislation occurred amid consensus that current health care payment systems are neither effective nor sustainable. The current system, based on volume and intensity, does not disincentivize, but rather pays more for, overuse and fragmentation. Providers note that current payment systems undermine efforts to invest money and effort in delivery-system improvements that can sustainably reduce costs. Payers are concerned that simply adding new payments for additional services, such as coordinating care or using health information technology (IT), might not necessarily reduce costs and might even add to them. Yet finding better and more feasible alternatives for organizing or paying for care has been difficult.
The concept of accountable care organizations (ACOs) as an alternative, however, now has the official imprimatur of health reform.1 The secretary of health and human services has been granted authority to launch pilot projects to test the concept and, should it prove successful, to replicate it broadly throughout the Medicare program. At the same time, substantial uncertainty exists around exactly what an accountable care model would look like, and whether and how it could work. Additionally, questions remain about how accountable care reforms could be implemented successfully across the country.
In this paper we first briefly summarize the underlying concepts, overarching principles, and prospects for widespread implementation of the ACO model, including how it evolved from recent Medicare demonstrations. Next we highlight the role of primary care in these reforms, and the interactions between the ACO model and other payment and delivery system reforms. We conclude by describing some of the key challenges and how a national implementation strategy including Medicare and other public and private payers might address them.
What Are Accountable Care Organizations? An Update
With growing interest in accountable care organizations, several experts have advanced notions of how to define them,2⇓–4 and they largely agree on the core concepts. ACOs consist of providers who are jointly held accountable for achieving measured quality improvements and reductions in the rate of spending growth. Our definition emphasizes that these cost and quality improvements must achieve overall, per capita improvements in quality and cost, and that ACOs should have at least limited accountability for achieving these improvements while caring for a defined population of patients.
ACOs may involve a variety of provider configurations, ranging from integrated delivery systems and primary care medical groups to hospital-based systems and virtual networks of physicians such as independent practice associations. All accountable care organizations should have a strong base of primary care. Hospitals should be encouraged to participate, because improving hospital care is likely to be essential to success. But in contrast to others’ definitions, we believe that this need not be an absolute requirement for all ACOs. Rather, to encourage broad participation, we have called for testing a range of models, in accordance with local provider preferences and market circumstances, to identify the best ways to improve quality and lower costs.
Accountable care organizations can be implemented through different payment models. These could include opportunities to share in demonstrated savings within a fee-for-service environment, in which providers took on no new financial risk. They could also include limited or substantial capitation arrangements, in which payments were unrelated to the volume of services provided, to the intensity of service use, or to the frequency of face-to-face meetings, and in which providers took on some financial risk for poor-quality results or failure to control costs.5,6
Thus, accountable care organizations should have considerable flexibility in many aspects of design. At the same time, all variations would be based on these core defining principles: (1) Provider-led organizations with a strong base of primary care that are collectively accountable for quality and total per capita costs across the full continuum of care for a population of patients. (2) Payments linked to quality improvements that also reduce overall costs. (3) Reliable and progressively more sophisticated performance measurement, to support improvement and provide confidence that savings are achieved through improvements in care.
Exhibit 1 depicts the different conceptual levels and potential evolution of accountable care organizations.
Characteristics Of Accountable Care Organizations (ACOs) Across Levels
Based on this comprehensive definition, a wide variety of provider organizations—such as existing integrated delivery systems or other coordinated care arrangements involving hospitals, physicians, and long-term care and other providers—could be accountable care organizations. There would be transparency through well-specified cost and quality performance measurement. Payment incentives would be aligned with achieving better quality and lower costs for defined populations of patients.
Indeed, through the Brookings/Dartmouth Accountable Care Collaborative,7 we have established a “learning network” with approximately sixty provider systems across the country that have diverse organizational structures and market characteristics. These organizations are at different stages of learning and implementation. We have devoted additional technical assistance to a smaller number of organizations that are nearing the launch of ACOs in tandem with several commercial payers, with the expectation that they will eventually include Medicare and possibly other public payers.8
Building On Recent Medicare Demonstrations
The ACO model builds on similar initiatives that Medicare has implemented in the past several years. Starting in 2005, the Physician Group Practice Demonstration engaged ten provider organizations and physician networks, ranging from freestanding physician group practices to integrated delivery systems, in a “shared savings” reform. The providers in the demonstration continue to receive all of their usual fee-for-service payments. But they also receive bonus payments if their efforts to improve care through better care coordination and other delivery reforms translate into slower risk-adjusted health spending growth and improved performance on quality measures for the patients they serve. Participating providers are also held accountable for a portion of any excessive spending through reductions in future bonus payments.
All ten participating sites have achieved success on most quality measures. In the third year of the demonstration, five had achieved sufficient reductions in spending growth to allow them to obtain more than $25 million in shared-savings bonuses as their share of a total of more than $32 million in Medicare savings.9 Forthcoming analyses on performance for years four and five of the demonstration should shed light on whether and under what conditions providers not receiving bonus payments in the early years will eventually become successful in slowing spending growth.
Building on and extending the Physician Group Practice Demonstration, the Medicare Health Care Quality Demonstration, established by the 2003 Medicare Prescription Drug, Improvement, and Modernization Act, is testing similar payment and quality improvement reforms in other delivery settings. Under this demonstration, regional provider-led organizations will qualify for shared-savings payments if they are able to improve care quality while slowing health spending for Medicare beneficiaries.
In that context, two large regional demonstrations were implemented in 2010: the Indiana Health Information Exchange, and Community Care of North Carolina. These collaborations involve a wide range of providers and multiple payers, including Medicaid, state employee health benefit programs, major private insurers, and now Medicare. Yet the underlying principles and program structure are similar to the Physician Group Practice Demonstration. Bonus payments are linked to performance on quality measures and demonstrated slower spending growth for a defined population of beneficiaries.
Accountable Care In Reform Legislation
In the wake of health reform, many of the provider systems in our learning network expect to implement accountable care reforms within Medicare. The recently enacted legislation will extend past demonstrations in several ways.
Pilots:
First, ACOs that meet certain criteria, such as having a sufficient minimum number of participating beneficiaries to produce statistically reliable predictions of future spending, will be able to implement accountable care pilots outside of the traditional Centers for Medicare and Medicaid Services (CMS) demonstration process.3,10 This provision is intended to permit a larger number of groups with a range of different characteristics to participate in these reforms more quickly than has occurred within the established research and demonstration process.
Payment Models:
Second, the legislation supports piloting a broader range of Medicare ACO payment models than those available under current accountable care demonstrations. These include a “one-sided” shared-savings model, which would entail no performance risk to providers even if they experience higher costs or if they do not achieve quality performance goals. Also possible are “two-sided” or “symmetric” payment models that would give providers an opportunity to receive proportionately larger bonus payments in exchange for accountability for costs that greatly exceed preset goals.
Partial Capitation:
Finally, a range of “partial capitation” models can also be established, which would replace a portion of the fee-for-service payments with a flat payment plus bonuses and penalties based on whether the predetermined cost and quality benchmarks are achieved. The legislation also supports new evaluation methods to reward ACOs for slowing spending growth. Current demonstration methods generally compare providers to trends in a local control population. By contrast, the new law augments this comparison with a pre-post budget projection approach that uses actuarial methods based on historical spending and utilization data of the actual group, to establish a clear and explicit quantitative target against which ACOs could track their performance over time. This approach addresses the increasingly challenging problem of identifying a “control” population not subject to health care reforms while still achieving savings, and potentially enables more widespread and timely adoption of reforms that prove effective.
CBO Scoring:
In the final reform bill, these legislative provisions were “scored” by the Congressional Budget Office (CBO) as saving approximately $5 billion in the ten years after enactment.11 This is clearly not “bending the cost curve,” yet it represents a step beyond other incremental “process reforms” in care delivery (such as health IT, care coordination, or disease management implemented alone) that are not regarded by the CBO as having sufficient evidence to conclude that they would reliably reduce costs in themselves. The estimate of some direct savings from accountable care organizations is promising, but it highlights that the organizations should be viewed as an element of broader reform implementation and that better evidence is needed on how such reforms can lead to the greatest improvements in care. We turn to these issues in the remainder of the paper.
Accountable Care Organizations And Health Care Reform
Implementation of accountable care organizations is likely to be more effective if it can be aligned with a range of other reforms that also increase the emphasis on and support for improving quality and reducing costs. This is particularly true for primary care–oriented reforms such as the patient-centered medical home.
Support For Primary Care
Primary care is central to the successful implementation of ACOs. First, implementation requires participation of primary care physicians who agree to be held jointly accountable for improving quality and lowering spending growth for the population they serve. Although the organization itself is ultimately held accountable for all costs related to this defined patient population—including costs for providers not participating in the ACO—the model itself is rooted in existing relationships between primary care physicians and their patients.
Second, reforms that support primary care can leverage accountable care, and vice versa. For example, medical homes typically involve additional payments to primary care physicians each month in exchange for physicians’ leading prevention, disease management, and care coordination activities that reflect best practices in primary care. These payments are intended to support the time and investment required to coordinate care effectively, with the expectation that they will result in better outcomes and lower costs for the patients involved. However, analysts have expressed concern that savings achieved through implementing medical homes may not reliably offset the costs of the medical home payments to the participating primary care providers on a long-term and sustainable basis.12 These concerns have unquestionably been a barrier to more widespread adoption.
Implementing a medical home and accountable care organization at the same time could address budgetary concerns while also providing more incentives for overall care coordination. In a simultaneous model, medical home payments that primary care physicians receive each month would initially be paid out of a separate funding pool, just as public and private payers make medical home payments today. Any bonuses under an ACO would be available after a given performance period, once the organization’s budget is reconciled, and could in turn be used for future medical home support.
For example, in Community Care of North Carolina, one of the accountable care demonstrations mentioned above, incremental medical home payments have been financed through separate funds supported by Medicaid. Any future shared savings from the demonstration with Medicare will support additional infrastructure and quality improvement investments to help participating providers achieve further improvements in care.
Successfully coordinating medical home reforms within ACOs requires addressing some technical implementation issues. For example, the two types of care settings often use different methods for determining “attribution” of patients—that is, identifying which patients are associated with providers in each program for the purposes of measuring performance. The technical differences reflect the somewhat different aims of attribution based on beneficiaries’ past patterns of care.
Medical homes need to assign patients to a specific, current individual provider for medical home services. Consequently, patients are often attributed to medical homes based on the most recent physician visits. Accountable care organizations, by contrast, are designed to foster shared accountability for overall quality and costs encompassing a larger range of providers, including primary care physicians. Thus, attribution should reflect the overall pattern of care for a set of providers to facilitate performance measurement involving outcomes along the care continuum.
In the long run, payers and providers may benefit from the development of a single, consistent attribution method for both medical home and ACO reforms.
Finally, because of their potential for improving health and avoiding costly complications of chronic diseases, many of the performance measures used by accountable care organizations relate to performance in primary care. We turn to these measures next.
A Path To Comprehensive Performance Measurement
A core principle and design feature for all ACOs is the implementation of a robust quality measurement strategy. Such a strategy should help ensure, and make the public confident, that any cost savings are attributable to actual improvements in care.
Exhibit 2 illustrates a set of quality measures, drawn from our interactions with the providers and other stakeholders participating in our Accountable Care Organization Collaborative. As the exhibit shows, these measures encompass key national priorities for performance improvement that have been endorsed by a wide range of stakeholders, including consumer groups.13
Examples Of Performance Measures In Accountable Care Organizations (ACOs)—Beginning, Intermediate, And Advanced Accountability
Most providers today have limited health IT capabilities. When they initially establish an accountable care organization, they are likely to depend at first on measures based on administrative data—such as use of cancer screening (preventive care), lipid and blood sugar testing for patients with diabetes (chronic care), and hospital inpatient or emergency department use (use of care). Over time, participating providers are expected to undertake more sophisticated steps to improve care, which will require more advanced health IT capabilities and involve clinical data—such as adequate control of lipids, blood sugar, and blood pressure for patients with diabetes. These information systems will support more advanced measures. Particularly if they are aligned with the measures used in complementary payment or quality improvement initiatives, they can support further improvements in care.
Interactions With Other Payment Reforms
Accountable care organizations and medical homes both contain structures that support new and unproven payment mechanisms. Many other types of payment reforms of interest to policy makers also have not yet been widely implemented or evaluated, including bundled payments for episodes of care and payments to reduce readmissions. Some evidence suggests that these reforms may be more likely to slow cost growth if they are tied to overall accountability for producing better results.14 For example, discharge planning to prevent avoidable readmissions is likely to be more effective if not only primary care physicians but also the hospital, long-term care facility, and surgeons and other specialists involved in the care of a patient get paid more when they work together to make the discharge planning succeed.
Similarly, episode-based payment reforms may work more effectively if they are coupled with initiatives and incentives that pay more when reductions in the rates of some types of episodes (such as surgeries for chronic diseases or hospitalizations for heart disease) actually occur. Implementing these payment reforms within ACOs can yield both up-front funding for providers to invest in the delivery improvements that can help the ACOs succeed, and the confidence that additional short-term payments are more likely to translate into long-term cost savings.
Accountable care implementation can also be aligned with other performance measurement initiatives to improve the consistency and clinical relevance of performance measures. In particular, the American Recovery and Reinvestment Act (ARRA) of 2009 provided for an estimated $29 billion by 2016 in Medicare and Medicaid payments to physicians and hospitals for the “meaningful use” of health IT to improve care.15 The U.S. Department of Health and Human Services has reaffirmed its emphasis on paying for the use of health IT to improve care, not simply for functional capabilities.15 Accountable care performance measures represent a relatively straightforward way to measure whether these investments actually result in improvements in care.
Addressing Barriers To Effective Implementation Of Accountable Care
Several policy concerns and technical issues must be addressed to effectively bridge the gap between our current limited knowledge about how best to build on current accountable care demonstrations. As we have noted, one barrier is the lack of common understanding about what an accountable care organization is. Although we believe that accountable care must focus on the goal of population-level performance improvement, the preceding discussion should make clear that a wide range of policy and organizational reforms to improve care can and should complement this focus.
Another potential barrier is the lack of knowledge and experience on the part of both providers and payers in establishing the organizational and legal structures required to implement accountable care payment and performance measurement reforms. And a related and increasingly important barrier is current uncertainty about legal and regulatory issues surrounding provider coordination and engagement with multiple payers. In particular, such coordination may reinforce antitrust concerns about creating more consolidation that leads to higher costs.16
Finally, even if these issues are addressed, the success of an ACO will depend on whether it is able to support providers in achieving meaningful clinical improvements. Successful implementation will thus require ongoing learning, not only about the effectiveness of different approaches to reorganization, payment, and clinical improvement in different markets, but also about how local contextual factors influence the success of different accountable care models.
Moving Forward: Overcoming Barriers
Exhibit 3 summarizes steps that can be taken to address barriers to effective implementation of accountable care and develop better evidence on how accountable care reforms can have a greater impact on improving care and reducing costs.
Barriers And Approaches To Implementing Accountable Care Organizations (ACOs)
National Leadership
A national pilot program effort should be guided by clear and transparent principles emphasizing the importance of having providers and payers align payment incentives around the goals of improved care and lower costs. These principles can be supported and made concrete through a clear and comprehensive pilot evaluation strategy, with public and private participation.
Consistent Specifications
This pilot strategy, and ACO implementation more generally, does not require that all pilot programs be identical. In fact, a range of models could be successful, and more evidence is needed on these variations. However, consistency in definitions of cost and quality measures across all participating payers will foster both more informative evaluations, in that different pilots can be better compared, and wider adoption, in that providers will have more clarity about needed care improvements and better support when they achieve measurable improvements.
Building on the illustrative starter set of initial performance measures outlined in Exhibit 2, future pilot programs should promote the development of more comprehensive performance measures that encompass patients’ experiences of care, health outcomes, and the overall costs of care, including more specific and actionable cost and efficiency measures. Accountable care measures should also be aligned with those in other payment reforms—including medical homes, clinical transformation activities, and support for health IT.
Technical Support For Implementation
A national pilot strategy should include technical assistance and opportunities to exchange experiences with best practices to help providers and payers make important actuarial, contracting, and other technical decisions. Our efforts to support ACO pilots through our learning network suggest that agreeing to consistent performance measures, and provider-payer negotiations to execute contract changes, requires a strong initial commitment of time and leadership from participating stakeholders.
Payers may need assistance in implementing reliable, transparent, timely, and valid measures of quality and cost. Providers may need assistance in achieving improvements in measured performance, setting budget benchmarks, and addressing legal and other issues.
Support For Clinical Transformation
Clinical transformation is the linchpin of ACOs’ success, and it does not happen automatically by simply changing payment arrangements and measuring performance. Rather, it requires effective investment in infrastructure, process and organizational redesign, and other clinical activities to achieve delivery reforms that can actually produce needed improvements in care (for example, enhanced preventive care, better care transitions, and chronic disease management).
Accountable care measurement and payments simply help make these clinical improvements sustainable. Providers will benefit from efforts to identify best practices in reforms to combine both primary care improvements and better overall care coordination with specialists, hospitals, postacute care providers, and, ideally, providers of long-term services and supports.
Ongoing Evaluation
A number of provider systems in our learning network are already negotiating shared-savings contracts and implementing quality measures with private payers. These and other accountable care implementations will lead to better evidence, especially with consistent quality and cost measures.
However, a more straightforward process for Medicare and Medicaid participation in multipayer accountable care pilots could accelerate the successful development of the “learning” pilot program we describe in this paper. More regular and well-understood conditions for public participation, with well-defined evaluation mechanisms, would not only provide stronger incentives to improve care and reduce costs but would also provide more evidence to support beneficiaries who have the most to gain from systematic care improvements—older patients with multiple chronic conditions, frailty, or serious illnesses.
The newly enacted health reform legislation represents a strong foundation for advancing further implementation of accountable care organizations. Because these organizations focus on the bottom-line goals of better quality and lower costs, it is likely that support for this implementation and evaluation infrastructure will also make it easier to understand the quality and cost effects of a much broader range of health care policy reforms that complement ACOs.
Improved Antitrust Monitoring And Evaluation
Some recent studies have suggested that consolidation of providers into networks may be leading to increased costs.16 Such consolidation is generally accompanied by claims that it will lead to improvements in efficiency and quality. But in the absence of consistent and relevant measures of performance, which are critical for ACOs, evaluating these claims is difficult. This reality complicates the ability of antitrust authorities to challenge consolidations that increase costs. Some experts have suggested imposing price controls or expanding government provision of coverage, but these approaches may also have adverse consequences for overall quality and efficiency of care, and may not reduce spending if utilization rises.
Accountable care reforms could provide an alternative way of addressing the anticompetitive risk of consolidation. In particular, because the purpose of regulatory monitoring of provider (and other) integration and consolidation is to protect consumers, the transparent cost and quality measurement activities available through ACOs should help clarify whether consumers are receiving better care at a lower overall cost. In addition, consistent measures would facilitate payer and provider competition on value, rather than simply on fee-for-service price discounts. Clearly, the evaluation of accountable care reforms can help guide antitrust policy.
Concluding Comments
Achieving major, lasting changes in our current health care system has not been and will not be easy, even after passage of health reform legislation. Effective health reform requires real and significant changes in health care delivery, not simply incremental policy changes at the federal and state levels. It requires leadership from physicians and other providers at the community and regional levels, and support for better ways to deliver and finance care from public and private payers. This will not happen through legislative reforms alone, or by hoping that adding payments for more services will reliably achieve reductions in costs.
Only by fostering real accountability for both quality and costs—effectively linking payments with appropriate investments in infrastructure and process improvements—will we be able to make the transition to a health care system that better addresses major gaps in performance and makes critical clinical and process transformation feasible and sustainable. Developing and testing accountable care organizations, alone and in combination with other reforms such as patient-centered medical homes, represents a critical step away from purely volume-driven payments and toward payments emphasizing value. ACOs offer incremental modifications of current payments on the one hand, but the foundation for payments that provide fundamentally better financial support for clinical transformation on the other hand.
Beyond effective pilot programs of accountable care organizations established by the new heath care reform law, we have highlighted numerous public and private initiatives that could be coordinated with these reforms, including payments for e-prescribing, enhanced payments for reporting on quality measures, subsidies for the meaningful use of electronic health records, and other practical technical assistance and related resources from the economic stimulus legislation and potentially from further legislative proposals. These individual incentives may each be relatively small, but adding them up could produce significant (if temporary) revenue streams that could help upgrade provider capabilities and create an environment for medical practice that, cumulatively, is much more supportive of high-value health care than the status quo.
Now that national health care reform legislation has been enacted, Medicare, Medicaid, and, most importantly, providers and private payers should take coordinated steps to test alternatives to relying on payments based on volume and intensity of services. This is health care reform that pays for what we really want: helping providers and patients achieve better health and better care at a lower cost.
Acknowledgments
The authors gratefully acknowledge the collaboration and expertise of John Bertko, Cindy Chen, Larry Kocot, Steve Lieberman, and Mark Zezza.
NOTES
- ↵The Patient Protection and Affordable Care Act of 2010 includes a number of provisions that establish accountable care organizations in Medicare and other federal programs and that encourage public-private collaboration for accountable care. These include a “shared savings” accountable care program in Section 3022 and additional flexibility for implementing accountable care in Section 10307.
- ↵
- ↵
- ↵
- ↵As other analysts have noted, it may be advisable to encourage one-sided bonus models (such as shared savings) that transition over time toward models with greater accountability for costs and quality.
- ↵
- ↵
- ↵For example, the following providers, a subset of our larger Accountable Care Organization Collaborative, are currently implementing accountable care reforms: Carilion Clinic, Roanoke, VA; Norton Healthcare, Louisville, KY; and TMC HealthCare, Tucson, AZ.
- ↵
- ↵Initial actuarial analyses suggest that, for reliable measurement of spending patterns, an accountable care organization will need to serve a primary care population of at least 5,000 Medicare beneficiaries, or at least 15,000 patients under age sixty-five. Further analysis is warranted.
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
