Mid-Atlantic Health Law TOPICS

Background hero atmospheric image for ACOs: Altogether Curious & Odd

ACOs: Altogether Curious & Odd

A version of this article was published in The Daily Record on February 7, 2012.

On March 31, 2011, the Centers for Medicare and Medicaid Services (CMS) released its long-awaited proposed rules for the creation and operation of Accountable Care Organizations (ACOs).

Federal Health Care Reform mandated the creation of ACOs, which are clinically and administratively integrated health care super-entities, formed under state laws and possessing individual tax identification numbers. ACOs are to promote evidence-based practices and to coordinate care across providers, suppliers, and professionals participating in the ACO. Each ACO is “accountable” for the cost and quality of treatment to certain Medicare fee-for-service beneficiaries.

Congress’ stated goal for the ACO program is to reduce cost and increase quality in the Medicare program specifically, and throughout the health care system generally, by introducing risk-sharing and performance metrics into the traditional fee-for-service model, rewarding providers for decreasing per-beneficiary costs by “sharing the savings,” while punishing low-quality providers or providers that deliver fragmented or unnecessary care.

Beyond that broad goal, Congress left open many concrete questions about how ACOs would operate. CMS’ proposed rule provides answers to these questions, and the important features of the proposed rule are summarized below.

A. Sharing the Savings with CMS

CMS has defined “savings” to mean the amount CMS spends for ACO-included beneficiaries under a benchmark that CMS expects to spend for those beneficiaries on a risk and inflation adjusted basis.

ACOs may initially select from two models of shared savings based on risk preference. The first model is two-sided, with the ACO and Medicare sharing both savings and losses. The second model is one-sided, and allows the ACO to share in savings with the Medicare program, but does not require the ACO to share losses. The one-sided model is only available for the first two years of the three year initial agreement period between an ACO and CMS.

Under the one-sided model, ACOs may receive up to 50% of savings (52.5% for Federally Qualified Health Centers (FQHCs) and Rural Health Clinics (RHCs)) up to 7.5% of the ACO’s benchmark. To control for random statistical variance, sharing is triggered after 2-3.9% savings from the benchmark (the minimum saving rate), with the exact percentage of the minimum savings rate decreasing with increases in the size of the ACO’s beneficiary population.

Under the two-sided model, ACOs may receive up to 60% of savings (65% for FQHCs and RHCs) up to 10% of the ACO’s benchmark. Minimum savings rates and minimum loss rates are set at 2% regardless of the size of the ACO’s beneficiary population. However, if the 2% threshold is met, then all losses or savings, including the first 2%, are included in the calculation of losses or savings. During the three year agreement, loss sharing is capped at 5% of the benchmark in year one, 7.5% in year two, and 10% in year three.

ACOs participating in the two-sided model must provide CMS with a surety bond, an escrow, or a letter of credit, so that CMS knows it will get paid by the ACO if the ACO’s beneficiaries receive more care than the benchmark. CMS may also withhold up to 25% of shared savings payments in the first two years as a reserve against future losses.

Under both models, ACOs will only share savings if they also score well on five quality metrics, namely, (1) patient experience; (2) care coordination; (3) patient safety; (4) preventative health; and (5) at-risk/elderly health. High quality scores can also decrease shared losses under the two-sided model.

In light of the foregoing, if Medicare paid the following amounts in regard to beneficiaries assigned to a particular ACO without FQHC or RHC participants: $110 Million in year one, $100 Million in year two, and $110 Million in year three, while CMS benchmarked the ACO spending at $100, $110, and $130 Million, respectively, then CMS would assign a $10 Million loss in year one, a $10 Million savings in year two, and a $20 Million savings in year three.

Under the one-sided model, the ACO in the foregoing example would get $0 in year one, up to $3.9 Million in year two, and up to $10.8 Million in year three, for a three-year total of up to $14.7 Million in shared savings.

Under the two-sided model, the ACO would get up to $5 Million in losses in year one, because of the benchmark cap, while sharing up to $6 Million and $12 Million, respectively, in years two and three, with a three-year net total of perhaps $13 Million.

B. Sharing the Savings within the ACO

ACOs may create their own plans to distribute the shared savings payments within the ACO, subject to CMS approval. As part of its application to participate, an ACO must describe how it plans to use shared savings payments, including the criteria it plans to employ for distributing shared savings among its participants. The ACO must also describe how the proposed plan will achieve the specific goals of the “Shared Savings Program,” such as improved care and outcomes, and lower cost growth.

Since distribution of Medicare payments among ACO participants implicates the anti-kickback and self-referral laws, CMS will waive those restrictions on a case-by-case basis for the distribution of shared savings payments within the ACO, and waive those restrictions for payments to ACO participants for other activities “necessary for and directly related to” ACO participation. Flows of funds between ACO participants will also be allowed if they meet other existing exceptions to the self-referral laws, such as the exception for bona fide employment relationships.

C. Beneficiary Selection

Congress did not mandate a specific method for allocating fee-for-service Medicare beneficiaries to ACOs. In response, CMS has decided to assign beneficiaries to ACOs on an entirely passive basis: a beneficiary belongs to the ACO of the primary care physician from which the beneficiary receives the plurality of primary care services during a year as determined by primary care services charges, if that primary care physician belongs to an ACO.

The implication of this method is that for any given year, an ACO does not know which of its Medicare patients constitute its beneficiary population, that is, the individuals whose care the ACO is responsible for, until after the end of the relevant year.

Presumably, this passive, post-care allocation serves two purposes: (1) it ensures that beneficiaries retain the right to choose their health care professionals, rather than forcing them to use only ACO professionals, and (2) a year-end allocation forces the ACO to treat all Medicare patients as if they belonged to the ACO, preventing the ACO from discriminating against its ordinary fee-for-service Medicare patients.

D. Antitrust Guidance

Simultaneously, with CMS’ release of its proposed ACO rules, the Federal Trade Commission and the Antitrust Division of the Department of Justice, the agencies tasked with enforcing federal antitrust laws, released a statement outlining how the agencies will treat ACO formation and subsequent coordination within ACOs. That guidance further complicates an already complicated program, and that antitrust guidance will be described in a subsequent issue of TOPICS.

E. Conclusion

The proposed ACO rule is just that, proposed. While parties can now initiate “pilot” ACOs in accordance with the proposed rule, CMS may change the proposed ACO rule prior to its final adoption.

In any case, if the final rule is similar to the proposed rule, then, on the one hand, it is possible that ACOs may eventually join the ranks of other forgotten acronyms. As presently structured, the connection between behavior and reward may be so attenuated that ACOs never gain the traction needed to bend the health care cost curve materially.

On the other hand, in retrospect, ACOs may be seen as a significant first step in a philosophical shift in regard to how health care providers deliver care. Instead of worrying about a discrete episode of care, in the future health care providers may routinely act as if they are accountable for a continuum of care.


September 20, 2011




Rosen, Barry F.


Health Care