Health Affairs

The health care reform legislation enacted in March 2010 authorizes the Medicare program to contract with accountable care organizations (ACOs). These are networks of physicians and other providers that could work together to improve the quality of health care services and reduce costs for a defined patient population. This brief describes the ACO concept as set forth in the new legislation, discusses how ACOs might evolve over time, and reviews the challenges and opportunities facing health systems, physicians, administrators, insurers, patients, and policy makers as ACOs take shape.

Most public programs and private insurance plans pay for health care on a fee-for-service basis. This means that individual doctors, hospitals, and other providers are paid for each service they furnish to a patient. Critics of this system have long contended that it creates incentives for providers to furnish or order more services. And different providers who see the same patient often fail to coordinate their activities, leading to duplicative or conflicting treatments.

Over the years, there have been many efforts to promote integrated care systems, in which primary care physicians, specialists, and hospitals would work together to manage the overall care of their patients. Commonly cited prototypes include the Kaiser Permanente health plans, Mayo Clinic, and Cleveland Clinic. These systems own hospitals and employ salaried physicians. Their centralized organization means that providers within the systems can work together to improve quality and efficiency--for example, by exchanging patient information or developing and adhering to practice guidelines.

Coordinated Care: The same level of coordination has proved difficult to achieve when doctors and hospitals operate independently. Beginning in the 1970s, some physician groups or joint ventures between physicians and hospitals tried to operate as health insurers on their own. More often they contracted with health insurers to provide total care to an enrolled population. Many of these arrangements involved so-called capitation payment schemes: in return for a fixed monthly payment for each enrollee, the contracting group would accept financial liability for a range of covered services.

In effect, a "capitated group" took over the functions of an insurer, deciding which providers patients could see and what services would be furnished. The hope was that the group would find the best of way of managing care without going above a fixed financial ceiling. Although these arrangements still exist in some places, many consumers resisted network arrangements that restricted their choice of providers. There were also concerns that capitation would replace incentives to provide too many services with incentives to deny care.

The ACO has emerged over the last few years as a way of promoting integration while avoiding some of the perceived problems of past efforts. The concept began with the observations that physicians who are tied to a particular hospital often already function as a sort of informal network, and that their patients tend to stay within the network for most of their care. These facts suggested that groups consisting of one or more hospitals and doctors who use the hospitals, but aren't necessarily employed there, might be brought together in organized systems. Public and private payers could then hold these systems accountable by assessing whether they provided high-quality care to their usual patient population while reducing the unnecessary use of resources. Organizations that took steps to improve their performance would be financially rewarded; this would encourage further steps to improve care management, leading to further rewards and a steady evolution toward fully coordinated care systems.

Expanding Models And Structures: Discussions of ACOs have broadened from a focus on hospital-centered systems to include models based on physician practices--including large, multispecialty groups and independent practice associations (IPAs), which bring together solo practitioners and small physician groups in order to share resources and improve their bargaining power. And different people have advanced different ideas about how an ACO might operate--tightly or loosely structured, formed voluntarily or with the organization imposed on providers by Medicare or other insurers, and so on.

Because the ACO concept is a new one, it can be expected to evolve over time, as payers and providers learn which models work best. Exhibit 1 shows five delivery systems that could become models for ACOs.

Basic Features: The version of health reform legislation originally passed by the House would have given the Centers for Medicare and Medicaid Services (CMS) authority to pilot test a variety of different structural and payment approaches for ACOs. The Senate version that was enacted into law focused instead on one model that is now able to become a part of Medicare, not just a pilot program. The model embodies a few basic features proposed by some policy analysts:

• Invisible Enrollment. Patients who receive most of their care from ACO-affiliated providers would be treated as "assigned" to the ACO. At least at the outset, they would not be formally enrolled, would not be required to obtain services through the ACO, and might not even know the ACO existed. The assignment process would allow payers to define a population for which the ACO could be held accountable. Critics of this approach believe that patients should have a choice about participating in an arrangement that could reward providers for reducing services.
• Performance Measurement. Over some period of time, payers would collect data on utilization and costs for the ACO population and on measures of quality of care and population health. A provider could be required to meet minimum quality standards in order to continue to participate in the ACO. In addition, quality reporting requirements would encourage improvements in ACO-wide information systems, a key factor in developing coordinated care.
• Shared Savings. Spending for the population of patients in a particular ACO could be compared to targets based on past experience for the same patients, or to spending for similar patients in the community who were not assigned to the ACO. If the ACO was found to have saved money, it would receive some share of the savings. Just how the savings would be divided among the participating providers is a major question that each ACO will need to resolve on its own.
• Evolution Toward Stronger Incentives. In the beginning, there would be no downside risk: The ACO would not share in the losses if treatment of its patients cost more than expected, though this could change over time.

Easy First Steps: Some people have suggested a "tiered" system. Organizations taking the first steps toward integration would operate under the shared savings approach. Easy-to-achieve spending targets would be set at the outset to encourage various types of organizations to participate. Some health systems are already highly integrated and could bear more financial risk at the outset, including capitation for some range of services. Others would move into this category as their systems and capacities developed. (Note that capitation might require formal enrollment of patients and restrictions on their use of non-ACO providers. Incentives for participation would then be needed for patients as well as providers.) Exhibit 2 illustrates a three-tiered approach to ACOs.

Early Results Mixed: Some of these concepts, such as invisible enrollment and shared savings, have been tested in a five-year Medicare Physician Group Practice demonstration project that began in 2005. Ten group practices, most of them hospital-affiliated, were permitted to receive bonus payments if they met quality standards and reduced costs. Although the full results are not yet available, the experience of the first three years was mixed: Some groups qualified for bonuses, but other groups found that costs for their patient populations grew faster than those for comparable Medicare beneficiaries in the same geographic area. Possible explanations include the groups' limited ability to manage the care of nonenrolled patients and the fact that participating providers were still paid on a fee-for-service basis, with continued incentives to increase service volume.
Although much of the discussion of ACOs so far has been in the context of Medicare, there is growing interest in extending the concept to patients covered by Medicaid and private insurance. Cooperation among multiple payers in promoting ACOs could have several possible advantages. Providers may be more likely to modify their practices if most of their patients--not just those with one type of coverage--are included in the ACO population. Efforts to improve care may be more effective if several payers are using uniform performance measures and quality standards. And a multipayer ACO may have enough patients to allow a meaningful focus on populations with special needs.

Serious Challenges Remain: Some analysts suggest that, in the current environment, ACOs will have serious challenges to overcome. Hospitals and physicians in some specialties benefit directly from maximizing the volume of services they provide; they may not see possible shared savings as enough to offset the revenue they would lose from a reduced use of services. Solo practitioners and small physician groups lack the data systems and organizational structures needed to form ACOs. In many communities, these providers have begun to band together in IPAs, but the groups will take time to develop into coherent systems and may require start-up capital.

In addition, ACOs may face legal hurdles, including antitrust laws and Medicare restrictions on various types of financial relations among providers. (The health reform law allows CMS to waive some of these rules for Medicare ACOs.) Another concern is that a few highly integrated systems could capture a large share of the market, increasing their bargaining power with private payers and reducing the potential for savings.

Possible Savings: How much can Medicare and other payers expect to save through the use of ACOs? The Congressional Budget Office estimates that the new Medicare ACO initiative will save about $5 billion in its first eight years. This represents a tiny fraction of total projected Medicare savings from the health reform law, most of which come from reductions in payments to providers and private Medicare Advantage plans. However, the CBO estimate assumes that the program will grow slowly and that most initial savings will go to the ACOs rather than being kept by the Medicare program.

This highlights a central issue in the design of ACO initiatives: If the goal is to encourage widespread participation, Medicare and other payers may need to offer substantial early incentives, and financial rewards will need to come quickly enough to compensate new organizations for their up-front investments in data and other systems. Longer-range savings will come if ACOs and other initiatives can help to change the culture of the medical care system.

The health reform law establishes a Medicare shared savings program for ACOs, to take effect no later than January 2012. This is not a demonstration or pilot project; the law makes contracts with ACOs a permanent option under Medicare. However, many of the specifics are left to the discretion of the secretary of the Department of Health and Human Services (HHS), which will allow the design of the program to evolve over time.

Eligible Organizations: Entities that could participate in Medicare as ACOs include group practices, independent practice associations or other networks of individual practitioners, partnerships of hospitals and professionals, hospitals that employ professionals, and other groups defined by the HHS secretary. Each ACO must have a formal legal structure that will allow it to receive shared savings payments and distribute them among providers, and it must show that it can meet quality and reporting standards to be developed by the secretary. An ACO must agree to at least a three-year contract and serve an assigned Medicare patient population of at least 5,000. (These requirements are meant to ensure a large enough sample and enough time for meaningful performance measurement.) Preference could be given to organizations that also develop ACO arrangements with payers other than Medicare.

The law does not specify how beneficiaries will be assigned to each ACO. However, CMS has indicated that it will look at whether beneficiaries obtain most of their primary care from an ACO physician. Beneficiaries will remain free to use non-ACO providers, and payments for services by both ACO and non-ACO providers will be continue to be made on a fee-for-service basis. A spending benchmark will be set for each ACO, based on its assigned beneficiaries' past Medicare expenditures. The ACO will be deemed to have achieved savings if it keeps spending growth for its population below average per capita spending growth for all Medicare beneficiaries. If the ACO also meets the secretary's quality performance standards, it will receive some share of the calculated savings. How savings will be divided between the ACO and Medicare is left unspecified.

Other Payment Models: The HHS secretary is authorized, but not required, to use other payment models. The law specifically mentions partial capitation, defined as an arrangement under which highly integrated care systems would assume the full financial risk of providing some range of Medicare services (such as all physician services, or all services under Medicare Part B) in return for a fixed monthly payment per beneficiary. Another option would be "risk corridors," under which the ACO's potential for profit or loss would be limited.

The health reform law also authorizes experimentation with ACOs under Medicaid. A five-year demonstration project to begin in 2012 would allow participating states to contract with organizations made up of pediatric providers. More broadly, the law establishes a new Center for Medicare and Medicaid Innovation within CMS. This center will test a variety of new payment and delivery models for both programs. Some of the possible models might be seen as extensions of the ACO concept, including risk-based, comprehensive payment for groups of providers and coordinated care programs for the chronically ill and for people at risk of institutionalization.

CMS plans to issue a Notice of Proposed Rulemaking (that is, a draft regulation) for the shared savings program for ACOs in the fall of 2010. At this time, it is soliciting input from providers, patient advocacy groups, and other stakeholders. Until the full ACO program becomes operational in 2012, the law allows CMS to begin ACO contracts with the provider groups that participated in the Physician Group Practice demonstration project. CMS is having discussions with these groups and is likely to begin preparing the initial ACO contracts soon.

Meanwhile, health organizations are preparing for ACO formation on their own or by participating in joint initiatives, such as ACO collaboratives developed by Premier, Inc., a national alliance of 2,300 hospitals, and the ACO Learning Network sponsored by the Brookings Institution and the Dartmouth Institute for Health Policy and Clinical Practice.

Centers for Medicare and Medicaid Services, "Medicare 'Accountable Care Organizations' Shared Savings Program--New Section 1899 of Title XVIII: Preliminary Questions & Answers."

Congressional Budget Office, "Budget Options: Vol. I; Health Care," December 2008.

Congressional Budget Office, Letter to Rep. Nancy Pelosi, Speaker of the House of Representatives, March 20, 2010.

Department of Health and Human Services, "Report to Congress: Physician Group Practice Demonstration Evaluation Report," September 2009.

Devers, Kelly, and Robert Berenson, "Can Accountable Care Organizations Improve the Value of Health Care by Solving the Cost and Quality Quandaries?" (Washington, DC: Urban Institute, October 2009).

Fisher, Elliott S., Douglas O. Staiger, Julie P.W. Bynum, and Daniel J. Gottlieb, "Creating Accountable Care Organizations: The Extended Hospital Medical Staff," Health Affairs 26, no. 1 (2007): w44-57 (published online December 5, 2006).

Medicare Payment Advisory Commission, "Report to the Congress: Improving Incentives in the Medicare Program," Chapter 2, "Accountable Care Organizations," June 2009.

Shortell, Stephen M., Lawrence P. Casalino, and Elliott S. Fisher, "How the Center for Medicare and Medicaid Innovation Should Test Accountable Care Organizations," Health Affairs 29, no. 7 (2010): 1293-98.

Exhibit 1 of this Health Policy Brief published July 27, 2010 erroneously listed Atrius Health as an independent practice association; it is a multispecialty group practice. In addition, Hill Physicians Group (southern California) should have been listed as Hill Physicians Medical Group (northern California). Exhibit 1 has been corrected in this version.

Written by
Mark Merlis
Health Policy Consultant

Editorial review by
Robert A. Berenson
Institute Fellow
Urban Institute

Elliott S. Fisher
Director
Population Health and Policy
Center for Population Health
Dartmouth Institute for Health Policy
and Clinical Practice

Susan Dentzer
Editor-in-Chief
Health Affairs

Ted Agres
Senior Editor
Health Affairs

Health Policy Briefs are produced under a partnership of Health Affairs and the Robert Wood Johnson Foundation.

Cite as:
"Health Policy Brief: Accountable Care Organizations," Health Affairs, July 27, 2010.

Sign up for free policy briefs at:
www.healthaffairs.org/healthpolicybriefs