ACO regs, round 1

After a two-and-a-half-month delay, the long-awaited regulations for accountable care organizations (ACOs) were released on March 31. Policy wonks, health lawyers, and hospital consultants were so eager to read what the Centers for Medicare & Medicaid Services (CMS) had written that they almost crashed the website the afternoon it was released.

Now that the dust has settled and people have had a chance to read all 429 pages, the excitement is likely to be substantially subdued. For an administration that has promised increased flexibility and a willingness to consider midcourse adjustments in new delivery strategies, the proposed rule can only be regarded as a disappointment. Affected and interested parties have until the end of May to write comments to CMS, and it is likely that many will do so.

What’s Required?

Hospitals and physicians will be able to form new organizations called ACOs to share savings produced from traditional Medicare fee-for-service payments under Parts A and B, without being subject to penalties for violating antitrust, anti-kickback, and self-referral laws, provided they meet a number of requirements. It’s the challenge of the requirements and the minimum savings that needs to be achieved prior to the shared savings that is causing so much concern.

Under the rule, an ACO would be required to assume responsibility for meeting all the health- care needs of a minimum of 5,000 beneficiaries for at least three years. Each patient would be assigned to the primary care physician in the ACO who provides most of the patient’s care. A patient could be assigned either retrospectively or prospectively to an ACO, based on the patient’s use of primary care services. But there is no enrollment process, as there is with managed care. Patients would be notified as to whether the physicians they are accustomed to seeing are part of an ACO, and patients would always have the option of choosing another primary care physician if they wish to do so—no limitations would be imposed on patients regarding the use of physicians, whether or not they participate in an ACO.

To share in any savings, the ACO would have to meet both quality standards and a minimum savings level. The quality standards include 65 measures, about 20 percent of which are likely to be available from claims data, with the remainder coming from medical records or survey information. The minimum savings requirement to receive a portion of the shared savings would be 5 percent of a benchmark, with the benchmark being based on past experience. CMS proposes to withhold 25 percent of the shared savings until the end of each year, which would be forfeited if the ACO were to drop out or be terminated for any reason.

Two different types of risk models—a one-sided risk model and a two-sided risk model—would be available. In the one-sided risk model, the ACO would share savings only for the first two years, although it would be at risk for sharing savings and losses in the third year. In the two-sided model, risks and losses would be shared all three years. The potential savings are greater in the two-sided model, but the potential for loss is obviously also greater.

Any marketing materials used by the ACO would need to be preapproved by CMS.

The PGP Experience

Because the proposed ACO program is modeled after CMS’s Physician Group Practice (PGP) demonstration, which also provided for shared savings after minimum quality and savings thresholds had been met, the requirements included in the ACO proposed rule are particularly perplexing.

The PGP demonstration included 10 large physician group practices and ran from 2005 to 2008. These practices were multispecialty groups with well-known names, such as the Marshfield Clinic, Geisinger Health System, Park Nicollet, and Billings Clinic. By contrast, because many ACOs are likely to consist of organizations newly formed for the purpose of becoming an ACO, few of these organizations are likely to have the depth of experience that the PGP participants possessed as they embarked on the demonstration.

Nonetheless, even with all of their experience, only two of the 10 PGP participants were able to attain better than a 2 percent savings threshold the first year of the demo, and only half were able to surpass this savings threshold after three years. Although all 10 were able to meet the quality standards imposed as an initial screen, those standards were less extensive, and therefore less burdensome, than the 65 being proposed for ACOs.

In addition, the start-up costs for the PGP participants were significant. As a result of their experience, the Government Accounting Office estimates an expected average start-up and first-year operating cost for an ACO of $1.7 million.

What’s Next?

Given the difficulty of the 10 experienced PGP participants in meeting a 2 percent savings threshold, the requirements in the proposed regulations for a minimum of a 5 percent savings with a 25 percent set-aside are perplexing. The financial risks imposed on the likely much-less-experienced ACO project participants seem disproportionate to their likelihood of gain—particularly given that all would have to assume downside risk in year three (probably even before they have received information from CMS on their year-one experience) in addition to meeting all the other proposed requirements. Furthermore, the possibility of retrospective assignment of patients to ACOs would make an ACO vulnerable to the behavior of patients who have been outside of the organization’s control, disregarding the fact that ACO had no way of knowing it would be responsible for the patients.

It is not unusual that the government would want to protect itself from unplanned financial risk, but the onerous nature of several of the major provisions of the ACO proposed rule is inconsistent with the frequently stated desire of the Obama administration to offer alternative models that move away from the current system of fragmented, uncoordinated care.

Perhaps the idea of ACOs should have been explored first as a pilot program, for which it would be expected that the selection of organizations would be limited and selective. In that way, CMS might have been able to expand on the experience of the 10 large PGP participants by extending it to 50 or 100 newly established and less-experienced organizations for perhaps a commitment even exceeding three years, but without all of the excessive protections the government seemed to feel necessary to include in this regulation. The window for comments is open. All those organizations and individuals that have been waiting to form their ACOs with the expectation of participating in this new program should make their views known. With all of the hype being given to ACOs, it would be a shame to repeat the experience we had with provider-sponsored organizations in the 1990s.


Gail R. Wilensky, PhD, is a senior fellow at Project HOPE; a former administrator of HCFA, now CMS; and a former chair of the Medicare Payment Advisory Commission ([email protected]).

Gail Wilensky, Ph.D.

Categories