Health Policy$ense

What is an ACO? McWilliams Responds to ACO Study Critique

A few months ago, I spoke with LDI Senior Fellow Lawton R. Burns about a newly-published study in JAMA entitled, "Changes in Health Care Spending and Quality for Medicare Beneficiaries Associated With a Commercial ACO Contract."  Burns challenged the findings by arguing that the contract under study was more like the HMOs of the 1990s than an ACO of today.  That post, “Still Looking for That Unicorn: Latest ACO Study 'Misleading'” is here.

J. Michael McWilliams, the lead author of the JAMA study, responded to this critique, and here we publish his comments in their entirety.  We include a response from Burns as well. This exchange highlights an even larger issue—whether “ACO” is becoming an umbrella term encompassing varying degrees of risk-sharing and network “openness.” Going forward, it will be important to define and test the salient characteristics of the specific model presented as ACO, or ACO-like.  We think this is a valuable debate, and welcome other contributions (use comments section).

The Blue Cross Blue Shield of Massachusetts’ (BCBSMA) Alternative Quality Contract (AQC) is not a “closed network HMO with global capitation” as described by Professor Burns. The AQC covers enrollees in BCBSMA HMO plans whose PCPs are members of provider organizations that have agreed to the terms of the AQC; eleven organizations had entered the AQC by the end of our study in 2010. Although the HMO plan overall is “closed network,” members who choose a PCP in an AQC-participating organization are not confined to the organization’s providers.  Enrollees in the AQC continue to have full access to the much broader BCBSMA HMO provider network and face no additional cost-sharing for care from providers not participating in the AQC.  HMO plan enrollees may not even be aware they are “in” the AQC; their benefits and choice of providers remain the same as HMO enrollees whose PCPs are not members of participating provider groups.

In addition, the AQC payment model is not based on a global prospective payment (capitation) but rather a global budget.  Participating provider groups continue to be paid on a fee-for-service basis.  Fee-for-service spending at the end of a contract year is compared to the budget to determine shared savings or losses.

Accordingly, as noted in our paper, the features of the AQC are remarkably similar to ACO contracts in Medicare, particularly Pioneer ACO contracts that already include downside risk-sharing features.  (Of note, all participants in the Medicare Shared Savings Program are also expected to assume downside risk by their second contract period if they stay in the Program).  Moreover, the recent wave of “ACO” risk-contracting between commercial insurers and provider organizations includes many contracts that are similar to the AQC.  Many of these contracts include no downside risk (shared savings only), and those implemented within PPO plans require patient assignment rules like those used in the Medicare programs.  Our classification of the AQC as a “commercial ACO contract” is therefore accurate.

Our appropriate characterization of the AQC notwithstanding, we are careful to note in our study that the findings may not generalize.  Spillover effects of Medicare ACOs on care for non-Medicare patients may differ, but our findings nevertheless suggest that spillovers are important to consider.  Thus, our study is neither “misleading” nor a “misrepresentation.”  Likening ACOs to magical unicorns might be, but that was not one of our conclusions.

               - J. Michael McWilliams, MD, PhD

Burns’ response:

As described in a 2011 NEJM article, the AQC contract differs from an ACO in at least three important ways: the use of primary care physician gatekeepers, restrictions on out-of-network care, and provider-borne risk. BCBS implemented the AQC in its health-maintenance-organization (HMO) and point-of-service plans. Enrollees in these plans are required to designate a primary care physician as a gatekeeper, who must authorize referrals for care by a specialist or admissions to a hospital that is outside of the HMO network. ACOs, especially those in the Medicare arena, do not have a designated primary care physician for the patient and do not provide a gatekeeping function. Patients are free to self-refer to any specialist and may (indeed often do) seek care outside of the geographic market.

Why is this important? Out-of-network utilization leads to costs that are not under the control of the AQC, but for which the AQC is responsible under its global risk payment from the HMO. Thus, the AQC has downside risk for such expensive utilization. The global budget, with upside/downside risk sharing, distinguishes the AQC from most Medicare ACO contracts, where providers have elected only upside risk.

               - Lawton R. Burns, PhD