In its recent 2023 Advance Payment Notice, CMS proposed the development of a Health Equity Index as an enhancement to its Medicare Advantage Star Ratings program. CMS states that “As we further explore this option, we are considering what other data are available and what other social risk factors might be appropriate to include over time. For example, we are considering the feasibility and utility of incorporating the Area Deprivation Index (ADI) into the health equity index.” CMS has also proposed to use the ADI as part of a payment adjustment in its newly redesigned Global and Professional Direct Contracting Model renamed REACH (ACO Realizing Equity, Access, and Community Health).
The ADI has been extensively used in academic research and purports to incorporate a wide-array of social risk factors (SRFs) at a detailed geographic level (census tract). However, we encourage all stakeholders to take a step back and evaluate what the ADI actually measures in order to ensure it is the most appropriate tool for CMS payment policy. Without such a critical assessment, policy makers may steamroll this measure through to “check the box” on incorporating health equity into the Star Ratings and other CMS programs. MAOs have seen this play before regarding health outcomes, where substantial weight is given each year to Health Outcomes Survey measures whose results are generally random, unactionable and useless.
What is the ADI?
The ADI was first published by Gopal K. Singh in 2003 in order to see if geographic social risk factors could be linked to mortality risk. The author found that there was indeed higher mortality in areas that scored as the most deprived in this index. The published index includes 17 factors such as median household income, family poverty rate, percent of the population below 150% of poverty, average rent, median home value and the percent employed in white collar jobs. These data are obtained from the US Census Bureau, which allows very fine geographic level analysis. Because, as you would imagine, many of these factors correlate highly with one another, a factor analysis and principal-components analysis were used in index construction.
To understand how the ADI uses this information to calculate a score, an excellent reference can be found in this 2016 article by Knighton AJ et. al. titled “Introduction of an Area Deprivation Index Measuring Patient Socioeconomic Status in an Integrated Health System: Implications for Population Health”. As illustrated in the calculation detailed in this article, surprisingly most of the factors have essentially no contribution to the index score—the overwhelming major driver of the results are from median housing prices. The effect of this can be see by examining the mapping atlas of these results. Here it is obvious that areas such as Los Angeles and Boston with very high housing prices have almost no disadvantaged census blocks, while some states in the south are almost entirely composed of disadvantaged census blocks.
We were not expecting this finding given the extensive academic research that has been conducted using the ADI; however, as one might imagine, in many cases housing prices are a good way to indicate what neighborhoods are the most or least desirable as newer housing, good schools, access to services, and high-quality transportation options will all drive up the cost of housing (especially when comparing areas within a metropolitan area to another). These same factors that increase housing prices are also likely to be associated with a population with lower SRFs. Thus, at least when examining the results at a local level, the index appears to provide face validity.
The Use of ADI in Medicare Advantage
The Medicare Advantage Star Ratings program allows Medicare Advantage Organizations (MAOs) that obtain certain scores on a range of quality measures to earn a 5% bonus payment (the Quality Bonus Payment or QBP). Given the size of MA (over 40% of Medicare in 2022), the Star Ratings program creates billions of dollars in incentives across MA to increase quality performance. CMS has always viewed the Star Ratings program as a powerful means to advance its policy agenda—whether that be the incorporation of many more compliance-oriented measures during the Obama administration or the increased weight of member satisfaction measures during the Trump administration. Thus, it is unsurprising that the Biden administration is looking to advance its own top CMS policy priority of health equity through changes to the MA Star Ratings.
Per the Advanced Notice, the Health Equity Index could be implemented as a separate new measure, or as a replacement for the r-factor (a scoring bonus MAOs can earn by obtaining consistently high performance across measures). More specifically CMS describes such an index as the following:
“The distribution of contract performance on each measure for each SRF [social risk factor] would be separated into thirds, with the top third of contracts receiving 1 point, the middle third of contracts receiving 0 points, and the bottom third of contracts receiving -1 point. The index could then be calculated as the weighted sum of points across all measures included in the index using the Star Ratings measure weights divided by the weighted sum of the number of eligible measures to calculate the index. Contract performance on the index would vary from -1.0 (performance was in the bottom third for each included measure) to 1.0 (performance was in the top third for each included measure).”
It is not entirely clear exactly how the ADI would technically be incorporated, but for example, contracts that serve a larger portion of members in areas that are disadvantaged could be eligible for this measure. This would probably require the ADI to be calculated at the county-level given the nature of other CMS data.
Is the ADI Right for You?
As with many questions, the answer here is not straightforward. From a technical perspective the ADI appears to have several fundamental flaws (not all of which are covered here). On the other hand, the net effect of the ADI, especially if it was used in a measure that replaces the r-factor, may result in more plans that do not currently earn QBPs getting those payments. In comparison to the HOS Improving or Maintaining Physical and Mental Health measures, the results of a health equity index using the ADI would likely be more predictable and consistent over time, but, like the HOS measures, changes would be unlikely to be associated with changes in what it purports to measure: health equity.
*At ZAHealth we make the technical understandable. We go beyond the “what” and help illuminate the “why”—leading to deeper insights and more successful strategy. Please reach out to Adam Zavadil today to discuss your project.