Editor’s Note: The post below argues that accountable care organizations (ACOs) represent a critical step away from volume-driven health care payment and toward better health and better care at lower cost. In addition to Aaron McKethan and Mark McClellan of the Engelberg Center for Health Reform at the Brookings Institution (pictures and bios above), the post is coauthored by Elliott Fisher and Jonathan Skinner of Dartmouth College. Fisher is Director of the Center for Population Health at The Dartmouth Institute for Health Policy and Clinical Practice and Professor of Medicine and of Community and Family Medicine at Dartmouth Medical School. Skinner is the John Sloan Dickey Third Century Professor of Economics at Dartmouth College; he is also a professor at the Dartmouth Institute for Health Policy and Clinical Practice, and a professor of Community and Family Medicine at Dartmouth Medical School.
This post responds to a post by Jeff Goldsmith which argues that ACOs are “not ready for prime time.”
Accountable care organizations (ACOs) are provider collaborations that integrate groups of physicians, hospitals, and other providers around the ability to receive shared-savings bonuses by achieving measured quality targets and demonstrating real reductions in overall spending growth for a defined population of patients. ACOs can be organized in a number of provider configurations with different payer participants. They can also feature different payment incentives ranging from “one-sided” shared savings within a fee-for-service environment, to a range of limited or substantial capitation arrangements with quality bonuses.
Unlike traditional pay-for-performance (P4P) programs, which often add to total costs by paying out incremental financial bonuses in exchange for meeting certain benchmarks on process measures, ACOs place a much greater emphasis on measuring and rewarding results at the level of a population of patients – not at the level of particular services or episodes that may or may not add up to higher-value care.
In the current health care reform environment, ACOs are being considered as an element of reform to achieve urgently needed improvements in the quality and cost of health care. According to the Congressional Budget Office (CBO), ACOs can produce savings for Medicare ; indeed, many payers, including the Centers for Medicare and Medicaid Services (CMS), are currently implementing ACO-type payment reforms. As Jeff Goldsmith notes, it is important to identify any problems and issues with these reforms to ensure that current efforts build on past experience. On behalf of the Brookings/Dartmouth ACO Collaborative, which is designed to support providers in implementing ACO reforms, we also use this opportunity to correct some misperceptions of how ACOs work and highlight how they are designed to overcome challenges in previous payment reform efforts.
Building on Past Experience
Some have argued that ACOs could be vulnerable to the same challenges that have contributed to the failure of past hospital/physician partnerships, including the inability to effectively manage significant risk. Yet ACOs are not “risk bearing” in the traditional insurance sense, which makes them quite unlike provider-sponsored organizations (PSOs) and capitated groups of the 1990s. One of the advantages of ACOs is that they can be developed to accommodate a range of different levels of financial risk depending on the capacity and organizational experience of participating providers. As noted above, this can range from one-sided “shared savings only” payments when ACOs achieve quality goals and beat their spending benchmarks, to other arrangements in which participating providers share in a greater proportion of savings but also voluntarily bear some downside risk as well. We envision a full spectrum of ACO pilots emerging with a range of payment arrangements that are based on demonstrated provider capacity and experience.
Other hospital/physician efforts have suffered from the lack of infrastructure for providing and using reliable and timely cost and quality information to support care management. Such infrastructure failures are instructive for future reform efforts. That is why we have been collaborating with CMS, private health plans, and medical systems to identify and support the technical infrastructure needed to deliver timely, consistent performance information to ACOs and other providers. This infrastructure should be aligned with other ongoing efforts. For example, the upcoming CMS funding for health IT implementation provides a critical opportunity to support the adoption of health IT systems not based on technical capabilities or functional standards, but based on the meaningful use of technology to actually coordinate and improve care.
The lack of collegiality and collaboration between physicians and hospitals also raises questions about whether ACOs can be successful. We have heard many stories from providers around the country that our fragmented payment systems reinforce the fragmented care delivery system that is a major impediment to their efforts to work together to deliver high-value care. Hence it is critical to get on a path to reform payment systems and regulations that pose a barrier to such collaboration. ACOs should be part of that path.
Other critiques of ACOs involve the role of patients and their health care choices and decisions. A popular though somewhat misguided interpretation of the 1990s experience with managed care is that patients reject efforts to restrict their choice of providers. Yet the ACO model was developed to expressly avoid inhibiting patients’ access to providers. As we are implementing them, ACOs maintain patient coverage for all of their providers. What changes is that those providers have the potential to share in the benefits of delivering care to their patients more effectively and efficiently.
No Disruption To Doctor-Patient Relationships
There have been some concerns expressed about “assigning” Medicare beneficiaries to ACOs without their knowledge and consent, implying that ACOs would somehow disrupt existing doctor-patient relationships or health plan choices. It is true that to provide insights into patterns of patient care, ACOs use patient attribution methods to set annual spending benchmarks and track overall spending at an ACO population level. But these measures are derived from actual practice patterns driven by patient choice and ongoing provider referrals. Nothing about the ACO model locks in patients to any specific provider or group of providers, and nothing should prevent patients from benefiting significantly from more access to meaningful quality information and better, more coordinated care. This is why CMS has had success in implementing ACO-type approaches in demonstration programs: rather than taking anything away from beneficiaries, ACO payments provide an additional source of provider support to improve care in ways that are not feasible under the current payment system.
While ACOs place an emphasis on provider incentives, we would like to see ACO pilots extend to the consumer side of the medical marketplace to better engage patients in their own decisions about health and health care. This can take a number of forms. In the short term, consumers can get reliable and timely information to help them make better health care decisions. But it could also include more active steps like sharing a portion of savings with consumers when they participate in systems that are delivering high-quality, cost-effective care. This latter example could be similar to “Centers of Excellence” programs that pass a portion of savings on to consumers in the form of lower premiums or to tiered provider networks that give patients lower copayments when they choose high-quality, cost-effective providers. But such consumer incentives are not an ACO requirement.
Some analysts have considered a mandatory Medicare ACO program, but we have always advocated for voluntary ACO pilots. We believe a set of strong, voluntary Medicare ACO pilots, implemented in coordination with efforts of other public and private payers, offers the best opportunity to learn which payment incentives and quality improvement steps are most effective. Further, we support clear CMS authority to cancel ineffective ACO pilots based on our belief that ACOs must earn their bonuses. This authority should also allow CMS to extend and expand pilots when they prove successful. Only through the incremental accumulation of evidence on the experience of ACO pilots will we know how best to transform Medicare payments around real accountability for results.
To be sure, barriers to the success of the ACO model must not be underestimated. Physician leadership will be critical not only to reassure the public that reforms are intended to improve care but also to provide peer-to-peer feedback and support for the implementation of practice changes and care management. ACOs will need to have some level of organizational structure to ensure that performance data are collected and effective and that timely feedback to physicians is routinely provided. ACOs will further need to ensure improved performance measurement to provide assurance that changes in payment incentives will not lead to stinting on needed care. As the degree of risk borne by ACOs increases, the need for regulation of the financial security of these organizations will also increase. And despite new ACO incentives, financial and regulatory policies continue to create pressures against focusing on the best results for patients at the lowest cost. But as we have recently learned from health systems in high-performing regions of the U.S., these barriers are far from insurmountable.
Fortunately, a number of providers and payers are stepping forward to implement accountable care reforms that will promote greater responsibility for patient care across provider silos. This includes provider organizations like AultCare in Canton, Ohio, and the Carilion Clinic in Roanoke, Virginia. In addition to these more integrated systems, our Brookings-Dartmouth “ACO Learning Network” includes physician-hospital organizations in rural communities; groups of primary care providers collaborating with community hospitals in several suburban areas of the country; groups of large, competing hospitals in a major metropolitan area; and prestigious academic medical centers. While these providers reside within different market characteristics, they view the ACO model as a viable institutional mechanism that can help drive greater coordination and value-increasing integration, combining financial incentives with data and information to improve care.
Similarly, several new state policy initiatives are emerging that will implement ACO-like models into state purchasing and encourage delivery system reform more broadly. Leaders in Colorado, Minnesota, Oregon, and the Commonwealth of Massachusetts, for example, are seeking new approaches to contain the escalation of cost growth using ACOs as a component of their reform efforts. These efforts should be encouraged, supported with robust technical assistance, and carefully evaluated over time.
These and other ACO pilots would be much more effective if they include Medicare. Indeed, Medicare’s existing participation in some ACO-like demonstrations is promising. For example, the Medicare Physician Group Practice (PGP) demonstration offers ACO-style shared savings and quality improvement incentives for a number of providers in mostly integrated delivery systems. Building on the PGP demo, we believe new Medicare ACO pilots can help us learn whether and under what conditions similar payment, care coordination, and quality improvement steps can be taken in other market environments or provider arrangements. More recently, the new Medicare Health Quality Demonstration offers a similar payment and quality improvement structure in a multi-payer framework. All of the ACO pilots for which we are providing technical assistance through our ACO Network similarly include collaborations with multiple commercial payers along with, hopefully soon, Medicare and Medicaid.
Similarly, a wide range of private-sector efforts involve different combinations of care coordination programs, quality measurement and improvement efforts, and provider payment incentives. While many of these key elements of ACOs are not new, ACOs seek to integrate these reforms under a cohesive organizational framework leading to demonstrated better health and lower costs for a population of patients, without a radical shift away from current financing systems. This framework can be used to bring together and enhance the impact of other reforms, including medical homes, coordinated care networks, bundled payments, and other payment and delivery system reforms receiving increasing attention in the health reform debate.
Congress has an important opportunity to provide new authority and support to CMS for Medicare and Medicaid to participate in and support multi-payer ACO pilots at the regional level. To further strengthen the evidence base for accountability reforms, Congress should commit sufficient resources to ensure comprehensive performance measurement that encompasses patients’ experience of care, health outcomes, and the overall costs of care as well as the infrastructure to provide needed data to support ACO pilots. And CMS will need a much stronger capacity to support the timely implementation, evaluation, and rapid adaptation of a full range of payment reforms focused on improving quality and slowing cost growth, including ACOs.
Achieving major delivery system reform will not be quick or easy. But it should be very clear by now that it will not happen under the status quo of existing payment systems, or through patches or add-ons to those systems. Only by fostering real accountability for results will be able to address the major gaps in quality and costs in our health care system. We believe ACOs represent a critical step in moving away from purely volume-driven payments to payments based on what we really want to support: better health and better care at a lower cost.