By Elliott Fisher, M.D., and Stephen Shortell, Ph.D.
A year ago, accountable care organizations (ACOs) were little more than a concept that offered both promise and peril in the reform of health care delivery. Now, with the proliferation of private payer ACOs, the new Medicare Shared Savings Program ACOs, and the Pioneer ACOs, there will soon be several hundred health care organizations with payment contracts in place that meet the key criteria of the ACO model: responsibility for a defined population of patients and financial incentives that reward improving care and slowing cost growth. (See the map below for all Medicare ACOs.)
But if ACOs are to achieve their promise, we must acknowledge that we still have much to learn in at least four areas:
1) Contract design. The structure of the payment incentives will be key, and little is known about optimal designs. Issues include: the degree of risk taken on by participating providers; how the rewards are calculated and shared; how payment thresholds and caps are set, if at all; and how these might vary for organizations at different stages of development and in different parts of the country.
2) Organizational capabilities. While some certifying organizations and learning networks are measuring presumed requirements for success of an ACO, such as advanced health information technology, care management capability, and leadership and governance, the difficult work of validating these to determine whether and when they predict actual performance remains to be done. Further, we need to learn how these tools and processes are best implemented in different organizational settings operating in different markets.
3) Impact on patients. The ACO model should meet the needs of all those served by the organization, but special attention should be paid to those most likely to benefit from coordinated care: those who are sick, frail, poor, or have serious mental illness. The extent to which ACOs can meet the needs of these vulnerable populations is of particular importance, given the likely expansion of insurance coverage for these groups in 2014.
4) Impact on community-level health and costs. In addition to examining ACOs' impact on patients, it will be important to assess their impact on the health status and health care costs of the broader community. For example, performance measures and payment models might encourage providers to form partnerships with health departments, schools, and community-based organizations to reduce the burden of illness in their communities by addressing the underlying social and environmental determinants of health. Many observers are concerned that ACOs that contract with Medicare could take advantage of any market power they gain from consolidation to shift costs to private payers by raising prices for the privately insured. For this and other reasons, tracking community-level health care costs will be important.
To learn from ACOs' early experiences, three challenges will need to be overcome.
First, common definitions and measures of a core set of contract attributes and organizational capabilities will be needed. Examples of these include electronic health record functionality, the use of care management processes, and quality improvement measures. A number of ACO assessment tools are currently in use or under development, including several surveys (American Medical Group Association (AMGA), Premier, Dartmouth-Berkeley, Brookings-Dartmouth, UC-Berkeley ACO Safety Net Readiness Assessment, HRET Hospital Assessment of ACO Readiness); site visit–based assessments (Premier; AMGA, Brookings-Dartmouth); data collection processes (the Medicare Shared Savings Program and Premier application processes); the UC-Berkeley National Survey of Physician Organizations and the National Committee for Quality Assurance ACO certification program. While each of these groups will want to assess different domains of ACO capabilities to meet their own needs, much will be gained by forging agreement on a set of measures for a core set of ACO capabilities. Without this, it will be difficult if not impossible to compare findings across studies and cumulative knowledge will be seriously compromised.
There also needs to be consensus on at least a core set of performance measures so we can learn whether ACOs are successful. The 33 measures included in the federal ACO programs are a good starting point for measuring the impact of ACOs on covered populations and communities. These should be augmented with measures of cost and resource use and, as soon as possible, more advanced outcome measures such as patient-reported functional health status.
The second major challenge will be to track performance at both the ACO and community levels. This will require collecting and merging data from all of the public and private payers potentially affected by these contracts—from those whose populations are covered by the ACO contract and those who are not. Release of provider- and plan-specific pricing information raises issues of contractual commitment and competitive advantages on the one hand and antitrust concerns on the other. But without at least some common information on the quality of care, resource use, and relative pricing on the part of ACOs, it will be impossible to assess their performance. And without community-level aggregation, we will be hard pressed to know whether the new payment model is having an impact on what matters: the quality and affordability of care and the health of our communities.
A third challenge is to create transparency in sharing data and results among all ACO participants. While legitimate proprietary interests should be respected, greater learning will occur if those involved exchange data and results. This will require discussion among the federal government, consulting firms, think tanks, learning networks, and the academic research community. The goal should be to develop ground rules or guiding principles that balance the legitimate self-interest of participants with the need for shared learning to improve health care quality, promote population health, and control costs.
Success in meeting these challenges will depend on commitment by private and public stakeholders to craft a path forward that meets their interests as well as the public good. Further, a process is needed to coordinate data collection initiatives, beyond what individual organizations can do on their own. This will require funding, perhaps from the Centers for Medicare and Medicaid Services, foundations, and other groups that have invested in efforts to improve health and health care.
The goals of accountable care—supporting providers' efforts to work together to achieve better care, better health, and lower costs—are compelling. But translating principles into practice requires learning. Let’s not miss the opportunity.
View Accountable Care Organizations in a larger map