Moving Forward With Accountable Care Organizations: Some Answers, More Questions

Our health care system is moving from traditional fee-for-service payment to value-based alternative payment models, such as accountable care organizations (ACOs). With passage of the Medicare Access and CHIP Reauthorization Act, Congress has made clear its intention to accelerate this transition. Because there is strong bipartisan support for these market-based approaches to improving health system performance, it is an opportune time to reflect on what we know about the effects ACOs have had on clinical care and challenges that have been identified.

During the past 4 years, we have made great strides in advancing our understanding of the changes that hospitals and physician groups implement in response to changes in payment and how providers perform under alternative payment models. Accountable care organization contracts, one type of alternative payment model, hold groups of health care professionals responsible for the cost and quality of care delivered to patients. Accountable care organization programs attract a diverse array of providers with differing legal and governance structures, varied contracts, and mixed capabilities that span service dimensions. These capabilities may include care management personnel and programs, adoption of advanced analytics (eg, to estimate the risk of hospitalization), or support for shared decision making.¹ On average, ACO patients are spending modestly less on health care services^2,3 and are associated with improved patient satisfaction and other patient-reported measures, with gains often concentrated in high-need, high-cost populations.^4,5 Previous exposure to risk-bearing contracts,⁶ greater numbers of dually eligible or disabled patients,⁶ and higher initial financial benchmarks have been associated with greater savings. At the same time, these results mask variation in performance, with some ACOs saving while others spend over their benchmark following ACO formation.^2,4 In addition, these modest savings have come at a net cost to the Medicare program (both program costs and shared savings rewards) and, although some quality measures have improved, many have not.

The research presented in this issue of JAMA Internal Medicine provides insights from 3 different ACO payment models: the Medicare Shared Savings Program (MSSP), Colorado’s Accountable Care Collaborative, and Oregon’s Coordinated Care Organizations. All 3 programs show some degree of success, although the results continue to be modest in magnitude. In their study of the association between ACO participation and Medicare spending through 2014, McWilliams et al⁷ found reductions in inpatient and skilled nursing spending among the 2012 cohort of MSSP ACOs caused by changes in the path of care for ACO patients (reductions in the use of and length of stay at skilled nursing facilities). These reductions did not require full integration or ownership of the nursing homes, suggesting that improvement may be due to how ACO physicians managed the care of their patients in independent hospitals and nursing homes. In studying the association between the Colorado and Oregon ACO programs and spending, access, and utilization, McConnell et al⁸ found improvements in quality and cost performance in both of the state-based Medicaid ACO models. Oregon ACOs accept full financial risk for their patient population and must manage all care within a global budget. Participating Colorado groups receive per member per month funding to coordinate care, but are paid fee-for-service and do not have upside or downside financial risk. Equivalent reductions in spending were achieved in Colorado without significant financial risk or capital investment.

These findings are relevant to 3 major considerations facing policymakers: (1) concerns about the harms of consolidation, (2) the amount of risk bearing needed to produce changes in behavior, and (3) how to manage potential conflicts between alternative payment models. Consolidation between clinical providers, such as purchase of physician groups by hospital systems, and striking the proper balance between facilitating clinical integration and limiting market power is a major concern. Physician practices are evolving toward a model that may require ongoing investment in both technology (eg, electronic health records with decision support and registries) and personnel (eg, care coordinators) to enable team-based care. At the same time, independent and smaller physician groups may have greater capacity to innovate and reduce avoidable facility-based utilization. Many believe that continued survival of independent practices is critical, yet economic incentives favor consolidation due to higher reimbursement for the same service in a hospital-based office compared with a physician office. Hospital-physician group consolidation can also raise prices in less-competitive commercial markets through increased bargaining power vis-à-vis insurers. The McWilliams et al⁷ study suggests that financial integration between physician groups and hospitals is not necessary to achieve improved post-acute care. These findings should encourage antitrust regulators to consider carefully whether settings across the continuum—from outpatient to inpatient to post-acute care—require financial integration to provide clinical coordination. Other potential approaches to improving the viability of independent physician groups should be considered, including the equalization of payment rates across sites of care or direct financial support for independent practices as used in the advanced payment model.

Accountable care organizations are defined by their accountability for the cost and quality of care, but the degree of accountability for spending varies dramatically across contracts. Some, such as the Oregon Medicaid contract, are global budgets with substantial downside risk (where the organization is held responsible for spending over the budget). Others, such as the MSSP, are shared savings programs where participating organizations receive bonus payments for spending below a budget and bear no downside risk. The amount of financial risk bearing necessary to achieve behavior change is an important area of inquiry, in part because so little is known. The incentives in most existing alternative payment models, including ACOs, are commonly considered insufficient to result in behavior change.⁹ However, the Colorado study suggests that strong incentives may not be necessary.⁸ Rather, the Colorado medical home model improved value through supporting providers with coaching, connecting members with nonmedical services, and providing feedback on costs, utilization, and outcomes. When compared with the “high-powered” Oregon model, which uses global budgets and full financial risk, Colorado’s program, which includes no accountability for the total cost of care, produced similar reductions in spending over the first 2 years. The Centers for Medicare & Medicaid Services should be complimented for rapidly iterating on the risk frameworks in the MSSP and for the recent rule creating a path for more modest risk bearing (proposed Track 1+).¹⁰

A third issue concerns potential conflicts between Medicare alternative payment models and the poorly understood interplay of incentives across these reforms. Currently, hospitals retain savings within inpatient-initiated bundles for patients attributed to an MSSP ACO. Given McWilliams et al’s⁷ finding that reductions in post-acute care use were largely due to changes in care for ACO patients, and ACOs’ broader accountability for total cost of care, this distribution of savings should be reconsidered with ACOs receiving savings in attributed patients. Although incentivizing value within a bundle, bundled payment models do not eliminate the incentive to provide more bundles. Accountable care organizations have incentives to assess whether a specific bundled service (eg, joint replacement) is necessary, as well as to coordinate and reduce costs for the services within the bundle; they should be rewarded for both. As the Centers for Medicare & Medicaid Services expands alternative payment models, it is possible that bundled payment programs will undermine overall savings.

In summary, we still have much to learn. Accountable care organizations have been established across diverse market settings, using a multitude of organizational structures and approaches to governance and operations, and this heterogeneity is reflected in the heterogeneity of their performance. The 2 articles published in this issue add to a growing body of evidence on overall performance, several dimensions of quality, and spending. Nevertheless, we know little about the effects of ACOs on patients’ health and quality of life. Perhaps most important for ACO leaders and the long-term success of these programs, we know little about the key ACO capabilities that are important to ensuring their success in different organizational or market contexts. Although the Centers for Medicare & Medicaid Services has conducted rigorous evaluations of the Pioneer program, generalizable findings tailored to organizational contexts are few. A long-term commitment to alternative payment model evaluation is necessary to ensure effective, sustainable payment and delivery system reform.

Corresponding Author: Carrie H. Colla, PhD, Geisel School of Medicine, The Dartmouth Institute for Health Policy and Clinical Practice, One Medical Center Drive, Lebanon, NH 03766 (carrie.h.colla@dartmouth.edu).

Published Online: February 13, 2017. doi:10.1001/jamainternmed.2016.9122

Conflict of Interest Disclosures: Dr Fisher receives personal fees from Hospital for Special Surgery, Price Waterhouse Cooper, Lancaster General Health, Christiana Care Health System, American College of Pathologists, Angiodynamics (a for-profit company), Blue Cross, Blue Shield of Louisiana, National Confederation of General Insurance, Private Pension and Life, Supplementary Health and Capitalization Companies, Brazil, Blue Cross, Blue Shield of South Carolina, and Vizient, Signature Health Care. Dr Fisher is an unpaid member of the board of directors of the Institute for Healthcare Improvement and the Fannie E. Rippel Foundation. No other disclosures were reported.

Additional Contributions: Greg Kennedy, MS, and Courtney Stachowski, MPH (Dartmouth Institute for Health Policy and Clinical Practice), provided invaluable research assistance with this commentary; there was no financial compensation.