Concern is growing that CMS’ new equity tool could backfire

BY: DAN GOLDBERG | 04/17/2023 08:08 PM EDT

The Biden administration, in the name of equity, may inadvertently be penalizing low-income people and communities of color.

The Centers for Medicare and Medicaid Services implemented a new payment model this year that aims to award doctors and medical groups more money to treat underserved populations with greater medical needs.

But to determine which patients are most needy, the government is using a somewhat wonky formula that many public health experts worry discounts people living in urban neighborhoods and may pull money from communities that need it most.

At stake are billions of dollars that the government uses to incentivize medical groups to work with low-income populations.

The technical indicator gaining favor inside CMS is called the Area Deprivation Index, and it uses several factors, including home value, to measure socio-economic disadvantage in every census block in America.

Its ability to capture neighborhood-level detail is its appeal, but in parts of the country where extreme wealth and poverty live side-by-side, the metric can produce distorted results because home values tend to be higher in urban areas, even in places rife with poverty and poor health outcomes.

“The ADI is, in effect, a blunt instrument,” said Susan Dentzer, president and CEO of America’s Physician Groups, which represents about 360 physician groups across the country. “If you’re going to have an equity provision in a model, you want the provision to make sense and to truly pinpoint populations in need within communities, rather than masking those populations because of high housing costs.”

The Area Deprivation Index — which was created by the Health Resources and Services Administration decades ago and then refined by researchers at the University of Wisconsin-Madison — is an important part of the Biden administration’s ACO REACH program, a Medicare pilot that offers provider groups a set amount of money each month to treat a roster of underserved patients.

Under the program, CMS provides an additional $30 per month for patients who are considered in the top 10 percent of most disadvantaged and takes $6 per month from patients in the bottom half.

“The ACO REACH Model promotes health equity,” Liz Fowler, director of the CMS Innovation Center, said in an email. “ADI, which is captured at the census block level, is a widely-accepted measure of social risk. It includes multiple measures of disadvantage such as local socioeconomic factors correlated with health disparities and underservice.”

The agency’s thinking is that neighborhoods with higher ADI scores are home to people with greater medical and social needs and, therefore, it makes sense for the government to pay extra money to the Accountable Care Organizations — groups of doctors, hospitals and other health care providers who provide coordinated care — that serve patients in those areas.

But parts of Washington, New York City and San Francisco have some of the worst health outcomes in the country, yet home values can remain relatively high because of their proximity to wealth and city centers. That can, in some instances, lower their ADI score.

The Tenderloin in San Francisco, for example, has multimillion-dollar homes and a homelessness problem that all get averaged together in the ADI without really reflecting the acute needs of people in the area.

Predominantly Black neighborhoods in southeast Washington have high rates of poverty and low life expectancy but the ADI remains relatively low, according to a recent article in Health Affairs, which found that under CMS’ policy, “no neighborhood in Washington D.C., would be considered disadvantaged.”

Similarly, the South Bronx in New York is one of the least wealthy congressional districts in the country and has among the lowest life expectancies. But its ADI score masks some of the inequality that every doctor in New York City knows exists.

Mount Sinai Health System’s Accountable Care Organization considered joining ACO REACH but passed, in part, because the patients it serves would be considered “advantaged” under the formula, which would likely penalize the ACO.

Maria Alexander, vice president of population health operations at Mount Sinai and a co-author of the Health Affairs paper, said the organization explained to the Center for Medicare and Medicaid Innovation that the government might be “overlooking” a problem. The administration was amicable but did not change its stance.

“I worry that the train has left the station,” Alexander said.

Another concern is that CMS pays ACOs based, in part, on the predicted medical expenses of its patient roster, and ADI can, in some instances, mask the medical needs of a low-income population, essentially mistaking a lack of access to medical services for a lack of need.

Using the ADI to measure social risk factors in models that determine payments to health plans could paradoxically worsen payment disparities, according to a recent JAMA article.

“We found that if you add the ADI to existing models to predict spending, and then adjust payments based on those predictions, you might think you’re advancing equity but, in reality, you may be diverting money away from the more impoverished areas,” said Suhas Gondi, a resident physician at Brigham and Women’s Hospital in Massachusetts and co-author of the JAMA paper.

CMS is aware of the critiques and said it may adjust the model in time. But it defended its use of ADI, noting that when combined with dual eligibility status — which looks at whether a patient is also on Medicaid — it provides a relatively complete picture of an individual’s ability to access high-quality care.

Still, the agency acknowledged it may need to be tweaked at some point in the future.

“CMS will continuously monitor the performance of the ACO REACH’s health equity benchmark adjustment and engage with external groups regarding other variables that might be included as part of the adjustment measurement,” Fowler said. “As demonstrated in the past, if future adjustments are needed, the CMS Innovation Center will make them.”

In a September RAND Health report commissioned by HHS, the assistant secretary for planning and evaluation recommended “continued study of how these indices would target funds, as well as development of indices that more directly target funds to [health-related social needs] at the geographic level.”

To be sure, every measure has its own flaws — replace ADI with life expectancy or the social vulnerability index, and new problems will likely arise.

But that doesn’t mean the problems with ADI should be overlooked.

“It’s easy to think you’re doing the right thing until you test it empirically and find that you might actually have the opposite effect,” Gondi said.