The Patient Protection and Affordable Care Act is serving as a catalyst in health care for new approaches to measuring performance and value, promoting wider adoption of health information technology (HIT), and developing models for delivering and paying for care more effectively and efficiently. In particular, the Medicare Shared Savings Program created by the legislation establishes financial incentives for accountable care organizations (ACOs) to provide coordinated, well-integrated care. Anticipation of the program has caused a flurry of activity among providers, purchasers, and payers.
Providers and payers recognize that for ACOs to reach their potential, there is a need for payment models other than the fee-for-service approach dominant today. As the new ACOs form, payers are establishing shared-savings programs and other payment models in an effort to create financial incentives for high-quality care. Payers are also considering payment methods that confer a portion of the financial risk to the provider, seeking to create stronger incentives than shared savings only; in fact, the proposed rule for the Shared Savings Program includes a shared-risk component. But while many providers and payers prepare to participate in ACOs, there is minimal evidence about what it takes for ACOs to succeed, including the payment models—shared-risk or otherwise—that will most appropriately support them.
This report summarizes research on ACO shared-risk payment models conducted by Catalyst for Payment Reform and Booz Allen Hamilton. The focus is on private sector payment models that meet criteria along three dimensions: provider risk, inclusion of services, and incentives for quality. Models of interest are those that include a provider risk-sharing component, address the broad array or full continuum of patient care/services, and provide meaningful quality incentives. While we found dozens of ACO initiatives, only eight met the criteria for inclusion in this study.
The research uncovered several key findings:
- Payer–provider shared-risk models are in an early developmental phase; there are few operational shared-risk models aside from the traditional capitated HMO model.
- There are varying definitions of shared risk, and shared-risk initiatives use a variety of program designs.
- Providers do not currently have the infrastructure required to take on and manage risk successfully, though some payers are providing infrastructure and other support to providers.
- Shared-risk models have typically evolved from shared-savings programs.
It is particularly important to note the very early stage of development of these population-based, shared-risk contracts. We found few operational shared-risk payment models; most are in development or at an early stage of implementation. Of those that are operational, many launched only in the first quarter of 2011 (three of the eight models). A variety of approaches to shared risk exist. Exhibit ES-1 summarizes the four main approaches we found.
While this work focused on shared-risk models that fall into the definitions in the bottom two rows of this table, it is the models in which providers take on whole or partial financial risk for a patient population—and move away from the fee-for-service "chassis"— that are arguably of greatest interest. All of the studied models include the fee-for-service "chassis"; we were unable to find any models currently in place that both move away from fee-for-service and include financial risk to the provider for a patient population.
The fact that providers lack the infrastructure they need to take on and manage risk successfully was also a common theme among the initiatives we studied. Providers do not have the data they need about the clinical or financial experience of their patients to manage patient care and financial risk effectively—the HIT structure necessary to coordinate care among providers is at varying levels of implementation. Providers also face operational and structural challenges related to the ACO model of care, which demand more coordinated, efficient processes. Many of the initiatives we studied try to mitigate some of these challenges by providing case management, disease management, and risk management support to providers.
Although this research uncovered several key findings about the development of ACOs and the payment models to support them, it is too early to identify which payment models best align incentives for ACOs with high-quality, high-value care. The majority of the payment models we studied have a fee-for-service foundation; however, shared risk combined with other base payment approaches can be even more robust. Capitation, bundled payments, and global budgets place the responsibility for managing financial risk more squarely on the shoulders of providers. Though we did not uncover any existing initiatives that employ one of these payment methods and meet the other research criteria, some initiatives anticipate more aggressive shared-risk payment models in their future. These burgeoning models and the experiments they embody will inevitably instruct us.
While many unknowns remain, this research advances our collective understanding and identifies several elements of ACO development to monitor and learn from over time. Through the Center for Medicare and Medicaid Innovation, there might even be opportunities to build on these experiments with public–private partnerships, strengthening the power of the pilots and making it potentially easier to detect their impact. Alignment between Medicare and the private sector will be critical to strengthen the impact of each other’s reforms, sending consistent signals to and intensifying incentives for health care providers to improve quality and reduce costs. It is also critical to assure that Medicare reforms do not simply lead providers to shift additional costs from Medicare to private purchasers and payers but, instead, make health care more affordable for all. It is unlikely that there will ultimately be a "one size fits all" solution; it will be important to learn which models work best in which situations.