The Centers for Medicare & Medicaid Services (CMS) recently announced that 32 health care organizations from around the country had signed on to the new Pioneer Accountable Care Organization (ACO) initiative, part of a broader push to incentivize ACOs in the Affordable Care act. ACOs provide financial incentives for healthcare teams—including primary care physicians, specialists, other health providers, and hospitals—to coordinate care, cut costs, and improve wellbeing among target patient populations.

Policy analysts contend that ACOs can contain costs and promote good outcomes because they align patient and practitioner incentives. By rewarding outcomes rather than the volume of health services performed, as in the traditional fee-for-service model, providers maximize their profits by improving patient health, which avoids unnecessary health services that increase cost. At once cutting costs and improving outcomes, ACOs seem like the magic bullet.

But there’s a catch: ACOs do little to address a crucial third issue in the healthcare dilemma—health equity. In fact, they may even exacerbate the problem.

The American health system produces mediocre outcomes for many times the cost of health systems in other high-income countries. In 2009, for example, Americans spent more than twice as much per person on healthcare than in the UK to live shorter lives. But perhaps more pressingly, the $8,000 per capita we spent in the US covered only 80 percent of our population, whereas the $3,500 per capita spent in the UK covered nearly the entire population. Part and parcel of this disparity in health coverage is a disparity in health outcomes: An African-American in the US can expect to live about 5 years shorter than the average White American. Among Whites alone, those with less than a high school education had a death rate 4.4 times higher than those with college degrees or greater in 2001—and that ratio appears to be rising.

What Education Reform Can Teach Health Reformers

To understand why ACOs aren’t likely to address our whopping health equity problem—and may make it worse—consider a lesson from our country’s education policy.

In 2001, “No Child Left Behind” (NCLB) was passed with bipartisan support as an Act of Congress. Proposed by President George W. Bush almost immediately after he assumed office, the legislation tied receipt of federal funds to school performance on standardized tests. Schools receiving federal funds were required to make “Adequate Yearly Progress” marks, or risk punitive measures.

What’s the problem with NCLB? Among others, school quality is not the only determinant of outcomes on standardized tests. In fact, there are a plethora of other factors that predict standardized test scores, including the number of parents in a child’s household, her annual household income, and the educational history of her parents, to name just a few. In this way, NCLB penalizes schools that serve lower income student populations that, despite any amount of improvement in facilities, administration, or teaching, are handicapped by the social determinants of educational outcomes opposing the success of their students.

How does this apply to the ACO model? Like schools in NCLB, ACOs are rewarded for improving outcomes. But healthcare quality, while an important piece, is not the only determinant of good health. The social epidemiologic literature is rife with data demonstrating the health influences that many of the same factors that predict testing outcomes—education, income, and neighborhood—may have on health. What’s more, these factors may also predict response to healthcare—such that the health improvements upon which ACO financial incentives are based may be substantially more difficult to attain in lower income patient populations than in richer ones.

What’s the upshot? Whereas public schools cannot move or restrict access, ACOs are private institutions that can do both—in lockstep with their financial incentives. Vexingly, then, the very incentives that tie provider compensation to patient outcomes to reduce costs and improve outcomes in the ACO model could also have the perverse consequence of incentivizing ACOs—and the health providers that comprise them—to turn away marginalized groups and to avoid locating in lower-income contexts. At the very least, these perverse incentives might dissuade providers already situated among poorer populations from forming ACOs, robbing this population of the potential outcome improvements these organizations could produce. Needless to say, this could exacerbate social disparities in health access and limit high quality healthcare among the people who need it most.

Current Efforts To Address Equity Concerns Are Inadequate

To be sure, policymakers have attempted to address this predicament. In the final version of CMS’s ACO regulations, ACOs are awarded a percentage of the savings they incur, assuming they meet 33 quality and outcome criteria. These savings are calculated using a risk adjustment tool that accounts for estimated patient costs as a product of disease history. In theory, then, ACOs shouldn’t be penalized for treating sicker, and therefore more expensive patients.

But this approach doesn’t adequately address the equity problem. First, payments are contingent on meeting quality and outcome criteria—criteria that are substantially more difficult to meet among poorer patients. For example, this list of 33 criteria includes influenza and pneumococcus immunizations, weight screenings and follow-up, tobacco cessation interventions, and depression screenings. Even in the setting of perfect availability, the poor are less likely to access these preventive services—and ACOs serving them are likely to be punished.

Second, while the hierarchical condition categories (HCC) classification scheme used in risk adjustment may account for projected costs associated with multiple underlying comorbidities, it doesn’t account for more transient complications in cost projections. As poorer patients are more likely to suffer these transient complications, HCC risk adjustment may systematically underestimate the costs of healthcare among the poor—rendering ACOs that serve poorer patient populations less likely to meet savings goals.

ACOs are certainly an attractive model to cut costs and to improve health outcomes. However, equity remains among the most important challenges in American healthcare—a challenge that ACOs may actually exacerbate. In order to avoid the potential widening of health disparities that could result from the widespread adoption of ACO payment models, quality criteria that ACOs must meet should be tailored to the socioeconomic realities of patient populations they serve, and risk adjustment mechanisms should account for the short-term complications that poor patients are more likely to suffer. Moreover, inequity must be a primary concern as we evaluate the performance of early ACOs and continue to improve the model.

Ignoring the equity challenges posed by ACOs—as was done with NCLB—could further marginalize our country’s most vulnerable.