Policymakers, payers and health care leaders agree that the current fee-for-service pay structure is creating unsustainable growth in U.S. health care costs. As a result, many are touting the accountable care organization as a way to bend the health care cost curve and encourage care coordination.

An ACO accepts responsibility for the cost and quality of the care its providers deliver to a specific population of enrolled patients. ACOs also coordinate care among multiple providers, and in doing so they can address some of the shortcomings associated with the fee-for-service payment system.

But the success of the ACO model resides in fostering clinical excellence and continual improvement. These goals may be accomplished best by incentivizing hospitals, physicians, post-acute care facilities and other providers to coordinate care while collecting and analyzing data on costs and outcomes.

Capabilities for ACO Participation

Although health care and health policy leaders have embraced the ACO concept, there are no national indicators of how many hospitals are participating in ACOs and what their current capabilities are in care management, financial management, information management and performance improvement. The Health Research & Educational Trust recently fielded a survey to determine the characteristics and processes of hospitals in ACO development. This project was supported by The Commonwealth Fund.

Data for this study were derived from a representative sample of 4,973 short-term, general acute care hospitals within the United States, as identified through the American Hospital Association's Annual Survey. This study did not include several types of hospitals, such as psychiatric hospitals, long-term care facilities, rehabilitative hospitals, children's hospitals, cancer hospitals and critical access hospitals, as they were not eligible for ACO participation. This study achieved an overall response rate of 34 percent.

Of the 1,672 hospitals responding to the survey, 1.2 percent reported that they were part of an ACO, 2.0 percent have established an ACO, and 9.6 percent indicated that they were working to become an ACO. While hospitals participating in an ACO and hospitals not exploring the ACO model did not vary significantly in terms of care coordination practices, hospitals participating in ACOs tended to be larger, more urban, more likely to belong to a centralized health system, better able to detect readmissions, and more frequently trained their leaders in continuous quality improvement practices.

Care Coordination and Readmissions Management

Our survey analysis revealed several additional themes. First, hospitals reportedly expect their revenue sources from risk-based financial reimbursements to nearly double over the next two years (from 9.6 percent to 18.5 percent). Bundled payments (physician plus hospital services) are expected to increase from 4.2 percent to 10.3 percent, and partial and global capitation payments are expected to increase from 5.4 percent to 8.2 percent.

Additionally, a majority of hospitals are actively engaged in numerous care coordination efforts, though there is variation in the use of specific practices. Care coordination practices reported to be widely used or standard practices are medication reconciliation (82 percent); sharing clinical information between settings of care (66 percent); and telephonic outreach to discharged patients (42 percent). The ACO membership status of the hospital had little impact on the implementation of various care coordination practices.

Hospitals are managing readmissions information in a variety of ways. Approximately one-third of hospitals not exploring the ACO model do not detect readmissions. This compares with non-detection rates of 18 percent for hospitals preparing to participate in an ACO and 14 percent for hospitals that are now participating in an ACO. Less than 10 percent (5 percent of hospitals not exploring the ACO model; 8 percent of hospitals preparing to become an ACO; and 8 percent of existing ACOs) can detect all readmissions. More hospitals that are ACOs can detect readmissions (57 percent) than hospitals preparing to participate in an ACO (54 percent) or hospitals not exploring the ACO model (48 percent).

Characteristics of Established ACOs

It appears that ACOs are maturing toward a fully coordinated accountable care model. At this stage of development, ACOs have an average of 128,300 patients, employ a number of processes to identify patients for population health, are actively participating in exchanging data through health information exchanges, and are monitoring and sharing a variety of performance data. 

Moreover, there is evidence that ACOs are striving to improve the quality of their services by using valid measures of their performance and making results available to the public and participating providers. Far more ACOs (84 percent) have an organized program to train clinical leaders in continuous quality improvement than non-ACOs (54 percent). Half of ACOs track and routinely share performance against measures with all members of the ACO. Of those currently sharing performance data, 91 percent are providing utilization measures by each setting of care as well as clinical quality measures by each setting of care. Eighty-seven percent are providing financial measures by each setting of care, and an equal percentage is providing patient satisfaction measures by setting of care.

Perceived Barriers

The study also identified perceived barriers. Hospitals preparing to participate in an ACO, as well as hospitals now participating in ACOs, said the most challenging barrier they perceived was reducing clinical variation. They said the least challenging barrier they perceived was developing a workable governance structure.


Hospitals preparing to participate in ACOs were far more likely to perceive the following four barriers as an "extreme challenge" than were hospitals that already are participating in ACOs:

  • aligning financial incentives (21 percent for hospitals preparing to participate in an ACO versus 6 percent for hospitals now participating in ACOs);
  • accessing capital and investing on a systemwide basis (16 percent for hospitals preparing to participate in an ACO versus 4 percent for hospitals now participating in ACOs);
  • raising start-up capital (19 percent for hospitals preparing to participate in an ACO versus 4 percent for hospitals now participating in ACOs);
  • increasing the size of the covered patient population (16 percent of hospitals preparing to participate in an ACO versus 4 percent for hospitals now participating in ACOs).

Finally, it appears that hospitals seeking to participate in an ACO are largely prepared to do so. They have the organizational and operational characteristics: They have established governance and legal structures, and they have the capabilities to manage their financial resources; share savings; and provide primary, acute and post-acute care.

Overall, our snapshot of ACO membership revealed that only a small portion of hospitals report that they belong to an ACO or that they are actively working to join an ACO. This number is likely to grow, as there is more evidence available to support the ACO model. Thus, it may be necessary to supplement this study by re-examining the differences in practices and structures of ACO hospitals. In the meantime, these findings suggest that the ACO model of care appears feasible from the hospital perspective as many hospitals interested in joining an ACO appear prepared to do so.

Kevin Kenward, Ph.D., is the director of research, and Nathan Bostick, M.A., M.P.P., is a senior researcher, both at the Health Research & Educational Trust.