The Centers for Medicare & Medicaid Services (CMS) has recognized that addressing health care disparities and achieving health equity should drive our nation’s top health priorities.[1] According to the CMS Framework for Health Equity 2022-2023, health equity is defined as, “the attainment of the highest level of health for all people, where everyone has a fair and just opportunity to attain their optimal health regardless of race, ethnicity, disability, sexual orientation, gender identity, socioeconomic status, geography, preferred language, or other factors that affect access to care and health outcomes”. In order to eliminate health and health care disparities, there needs to be an effort to provide quality, equitable care to those in areas without access and availability to the services necessary to meet their health and social needs.
While CMS has long supported research efforts that led to the development of objective quality of care measures, there is a need to improve the methods of measuring health equity within the U.S. healthcare system. Underserved communities are a large focus of CMS’ plan to advance equity as these populations share a particular characteristic that has been systematically denied a full opportunity to participate in aspects of economic, social, and civic life.[2] Identifying underserved communities is crucial for ensuring that the necessary populations receive equitable care. In an effort to provide equitable health care and address the impact of the social determinants of health, CMS has begun to incorporate needs-based assessments into several of their value-based care models, including the use of the Area Deprivation Index.
Understanding the Area Deprivation Index[3]
The Area Deprivation Index (ADI) is a measure used to rank neighborhoods’ socioeconomic disadvantage at the state or national level. The ADI has been adapted and refined from the original measure developed by the Health Resources & Services Administration (HRSA) over three decades ago. This measure of a neighborhood is determined based off of 17 variables in the theoretical domains of income, education, employment, and housing quality. ADI uses variables such as median family income; percent below the federal poverty level, not adjusted geographically; median home value; median gross rent, and median monthly mortgage to determine the deprivation of a census block group. The index uses “neighborhoods” or census block groups as its sample unit. Census block groups are combinations of census blocks that are a subdivision of a census tract or block numbering area (BNA), and of which are the smallest geographic unit for which the decennial census tabulates and publishes sample data. The ADI of a neighborhood can be ranked either at the national or state level. Since the theoretical domains for which the ADI consists of result from census data, the accuracy of this data is dependent upon the American Community Survey (ACS) Five Year Estimates. ADI measurements have also been found to be largely influenced by the choice of geographic unit of interest.[4]
National rankings of neighborhoods utilize a scale from 1 to 100. Percentiles are created by ranking the ADI from low to high nationwide, with neighborhoods grouped into bins corresponding to each 1% of the total range. For example, Group 1 is the lowest ADI and indicates the lowest level of “disadvantage” nationally, while Group 100 has the highest-level ADI and the highest level of “disadvantage.” As an example, the Washington, D.C. census block groups with the ADI national rankings applied is shown below on the left.
At the state level, ADI is ranked in deciles from 1 to 10 without consideration of the national ADI of the same area. Similarly, Group 1 at the state level is the least disadvantaged and Group 10 is the most disadvantaged. For comparison, an example of the Washington, D.C. census block groups with the ADI state rankings applied is shown below in Figure 2.
Utilization of ADI in the ACO REACH Model[5]
CMMI has incorporated the ADI into the ACO Realizing Equity, Access and Community Health (REACH) Model by establishing the Health Equity Benchmark Adjustment (HEBA) through a five-part policy strategy to promote health equity starting in PY2023. In theory, the HEBA will increase the benchmark payments of ACOs servicing higher proportions of underserved beneficiaries. The ACO REACH Model’s social risk adjusting financial structure distributes resources based on individual-level and community-level assessments in an effort to improve quality of care and health outcomes for all aligned Medicare FFS beneficiaries. HEBA is a composite measure that incorporates the national ADI of a particular area of interest and the Dual Medicaid Status of an individual, whether a beneficiary is enrolled in both Medicare and Medicaid. The ADI measures local socioeconomic factors impacting health service disparities, while the Dual Medicaid Status captures the economic challenges at the beneficiary level.
Due to the nature of the two variables, ADI as a continuous variable and Dual Medicaid Status as a binary variable, and to ensure that each measure carries the same weight in the HEBA, the adjustment will be calculated in the following way: the ADI will be scored 1-99 based on percentile relative to the nation and a 25-point increase will be applied to the score for the dually eligible beneficiaries. For example, a dually eligible beneficiary residing in a census block group with an ADI in the 75th percentile would receive a score of 75+25, for a total of 100. Once calculated, all aligned beneficiaries will be stratified and the top decile, the most deprived, identified for an upward adjustment of $30 per beneficiary per month (PBPM). The bottom five deciles, the least deprived, will also be identified for a smaller downward adjustment of $6 PBPM. There will be no adjustments for the remaining deciles. These adjustments allow for a budget neutral approach that distributes funds from areas of low deprivation to areas of high deprivation. Each ACO in the program will receive a net benchmark adjustment based on the number of serviced beneficiaries aligned in each decile. By adjusting and distributing funds based on social risk, ACO REACH aims to address the health inequities that underserved communities experience in order to provide high-quality care for all Medicare beneficiaries.
Utilization of ADI in the Medicare Shared Saving Program[6]
CMS made changes to Medicare Shared Savings Program (MSSP) in the Calendar Year (CY) 2023 Medicare Physician Fee Schedule Final Rule (the “Final Rule”) in order to align the program with CMS’ value-based strategy of growth, alignment and equity. The Final Rule made several changes to the MSSP based on feedback from health care providers in underserved populations. One of the principal concerns mentioned was the need for upfront capital to make the necessary investments to succeed in accountable care. In response, CMS finalized policies to advance shared savings payments, referred to as advance investment payments, for ACOs that did not have experience with performance-based risk Medicare ACO initiatives, that are new to the MSSP, and that serve underserved populations. In order to distribute these advance investment payments on a needs-basis, CMS will increase payments in ACOs with more beneficiaries who are: 1) enrolled in the Medicare Part D low-income subsidy (LIS); 2) are dually eligible for Medicare and Medicaid; 3) live in areas of deprivation, measured with the ADI, or a combination of any of these three social risk factors. The purpose of these advance investment payments for certain qualifying ACOs is to increase participation in accountable care models in underserved communities.
By providing advanced investment payments, CMS hopes to incentivize entities in rural and underserved areas to join together as ACOs, encourage the development of infrastructure to allow providers to succeed in MSSP, and promote equity by addressing beneficiaries’ health and social needs. According to the final policy, an eligible ACO may receive a one-time fixed payment of $250,000 and quarterly payments for the first two years of the five-year agreement period. Eligibility for quarterly payments is determined by the presence of one of the three social risk factors, highlighted above, in each assigned beneficiary. The amount of these quarterly payments will be based on a score set to 100 if the beneficiary is enrolled in the LIS or has a dual eligibility status and otherwise set to the ADI national percentile rank (an integer between 1 and 100) of the census block group in which the eligible beneficiary resides. Higher payment amounts are awarded to ACOs for assigned beneficiaries with a higher risk factors-based score. ACOs will not receive a payment for beneficiaries scoring below a risk-based factor score of 25. Advance investment payments must be used to improve ACOs’ quality and efficiency of items and services provided to beneficiaries through increased staffing, improved health care infrastructure, and the provision of accountable care for underserved beneficiaries.
Limitations of ADI as a Measure of Health Inequity
Accounting for social factors in risk adjustment is necessary for ensuring accurate Medicare performance assessments and encouraging equity within health care systems. As shown above, a common approach to incorporating social factors in risk adjustment is through community-level social risk factors, like ADI, which are widely available and relatively easy to implement. [7] Despite inclusion of these social risk measures in programs such as ACO REACH and MSSP, community-level social risk factors may not address payment disparities for some beneficiaries with high-levels of social risk.[8] The ADI national percentile rank has drawn particular attention recently due to claims suggesting that the index may mask the inequities and poor health outcomes of certain communities. Further research efforts should examine the effectiveness of the national ADI as a social risk factor index.
While ADI can serve as a proxy for poor community health outcomes in rural communities and lower-cost urban areas with low life expectancy, in areas with high property values and cost of living, but low life expectancy, ADI may mask health and social inequities.[9] Examples of these communities include neighborhoods in Washington, D.C. and New York City where the ADI national percentile rank misrepresents the health of individuals due to high property value. These underserved communities experience apparent health outcome disparities when examined individually.[10] Based on this observation, CMS should examine the national ADI ranking’s impact on health equity efforts underway in value-based care models. For example, under the ACO REACH model it is unlikely these areas would fall under the highest deciles, and therefore ACOs in these underserved communities would not receive the additional resources and funds necessary to provide high-quality care.[11] A similar issue would occur under the MSSP policies, where some of these areas would be considered advantaged and ACOs’ in these communities may not receive advance investment payments to cover the start-up costs of establishing the program to provide care for these underserved populations.[12] Many of the public comments on the MSSP CY 2023 PFS Proposed Rule recognize these limitations of the ADI national ranking and recommend that CMS continue to develop and test new measures to identify disadvantaged areas. According to a RAND Health report commissioned by the HHS, none of the existing health equity indices, including ADI, were ideal for policies directed at addressing social determinants of health (SDOH) or health related social needs (HRSNs). The report proposed that HHS examine how other indices could more accurately allocate funds to SDOH or HRSNs at the geographic level.[13] By masking the health disparities prevalent in communities with high property values and costs of living, but low life expectancy, using ADI may be limiting CMS’ goal to advance equitable health care.
Alternative Approaches to Measuring Social Risk Factors
In order to more accurately reflect the health disparities in certain underserved communities, several recommendations for social-risk factor measures have been proposed in place of ADI. One suggestion is to use an absolute measure such as small-area life-expectancy estimates rather than a relative measure. The U.S. Small-area Life Expectancy Estimates Project (USALEEP) produced estimates from most U.S. census tracts for the 2010-2015 period of the life expectancy at birth, the average number of years a person can expect to live.[14] Another proposal is to change the geographic scale on which ADI is measured. One study found that utilizing a local ADI estimate with a radius of 10 kilometers had the strongest association with hospitalization rates.[15] As the geographic scale increased, to 20-km, 30-km, and a regional area, the strength of the association with hospitalization rates decreased. The findings from this study support the need for locally sensitive relative deprivation measures.[16] While this study suggests that national and regional approaches may not properly reflect the impact of area deprivation on health outcomes in smaller study areas, further research is necessary to determine the applicability of a local area deprivation tool in other parts of the U.S. Another study found that the ADI is driven by variables with dollar values, with median home value carrying more weight than the other 16 variables in the calculation.[17] Using the raw values for monetary variables adds weight to the four variables (median home value, median family income, median monthly mortgage, and median gross rent), compared to the other ADI variables which are measured as proportions. This study recommends using standardized data to ensure that the monetary variables do not predominant the ADI. In doing so, the ADI can more accurately represent the social-risk factors of a census block group. Regardless of the proposal, further research must be conducted to ensure that changes to the ADI will highlight the inequities underserved communities face throughout the U.S., so resources and funds can be properly distributed to those in need.
[1] Centers for Medicare & Medicaid Services, CMS Framework for Health Equity 2022-2023. (Baltimore, MD: Centers for Medicare & Medicaid Services, 2022),5-7, https://www.cms.gov/files/document/cms-framework-health-equity-2022.pdf
[2] Centers for Medicare & Medicaid Services, ACO Realizing Equity, Access, and Community Health (REACH) Model (Baltimore, MD: Centers for Medicare & Medicaid Services, 2022), 1-116, https://innovation.cms.gov/media/document/aco-reach-rfa
[3] “About the Neighborhood Atlas,” University of Wisconsin, May 1, 2018, https://www.neighborhoodatlas.medicine.wisc.edu/
[4] University of Wisconsin, “About the Neighborhood Atlas.”
[5] Centers for Medicare & Medicaid Services, ACO Realizing Equity, Access, and Community Health (REACH) Model, 1-116.
[6] Centers for Medicare & Medicaid Services, Calendar Year (CY) 2023 Medicare Physician Fee Schedule Final Rule- Medicare Shared Savings Program (Baltimore, MD: Centers for Medicare & Medicaid Services, 2022), 1-13, https://www.cms.gov/files/document/mssp-fact-sheet-cy-2023-pfs-final-rule.pdf
[7] Powers BW, Figueroa JF, Canterberry M, et al. “Association Between Community-Level Social Risk and Spending Among Medicare Beneficiaries: Implications for Social Risk Adjustment and Health Equity,” JAMA Health Forum 4, no.3 (March 2023). doi:10.1001/jamahealthforum.2023.0266
[8] Id.
[9] Azar, K.M.J, Alexander, M., Smits, K., et al. “ACO Benchmarks Based on Areas Deprivation Index Mask Inequities,” Health Affairs Forefront, (February, 2023). DOI: 10.1377/forefront.20230215.8850
[10] “U.S. Small-area Life Expectancy Estimates Project – USALEEP,” Centers for Disease Control and Prevention, accessed May 3, 2023, https://www.cdc.gov/nchs/nvss/usaleep/usaleep.html
[11]Azar, K.M.J, Alexander, M., Smits, K., et al. “ACO Benchmarks Based on Areas Deprivation Index Mask Inequities,” Health Affairs Forefront, (February, 2023). DOI: 10.1377/forefront.20230215.8850
[12] Id.
[13] Health & Human Services, Landscape of Area-Level Deprivation Measures and Other Approaches to Account for Social Risk and Social Determinants of Health in Health Care Payments, (Washington, DC: Department of Health & Human Services, 2022), 1-106, https://aspe.hhs.gov/sites/default/files/documents/8dc674c904723bf8a5ce4cfc8d3dcdaa/Area-Level-SDOH-Indices-Report.pdf?source=email
[14] “U.S. Small-area Life Expectancy Estimates Project – USALEEP,” Centers for Disease Control and Prevention, accessed May 3, 2023, https://www.cdc.gov/nchs/nvss/usaleep/usaleep.html
[15] Maroko AR, Doan TM, Arno PS, Hubel M, Yi S, Viola D. “Integrating Social Determinants of Health With Treatment and Prevention: A New Tool to Assess Local Area Deprivation,” Prev Chronic Dis 13, no. 160221 (September, 2016). DOI: http://dx.doi.org/10.5888/pcd13.160221external icon.
[16]Maroko AR, Doan TM, Arno PS, Hubel M, Yi S, Viola D. “Integrating Social Determinants of Health With Treatment and Prevention: A New Tool to Assess Local Area Deprivation.”
[17] Hannan, E., Wu, Y., Cozzens, K., & Anderson, B., “The Neighborhood Atlas Area Deprivation Index for Measuring Socioeconomic Status: A Overemphasis On Home Value,” Health Affaris 42, no. 5 (May 2023): 702-709, https://www.healthaffairs.org/doi/epdf/10.1377/hlthaff.2022.01406