The development of the community deprivation index and its application to accountable care organizations

Abstract

There is strong interest among policymakers to adjust for area-level deprivation when making payments to providers because such areas have traditionally been underserved. The Medicare Accountable Care Organization Realizing Equity, Access, and Community Health (ACO REACH) model provides higher payments to ACOs serving areas with greater deprivation. Area Deprivation Index (ADI) is the primary component to measure deprivation for ACO REACH. The ADI is a commonly used deprivation index, but there are concerns about its methodology, primarily its use of nonstandardized deprivation factors. Prior research indicates the ADI is mainly determined by home values, which does not allow it to adequately capture deprivation in urban areas. This paper revises and updates the ADI, using American Community Survey data to compute a census block group deprivation index, the Community Deprivation Index (CDI). The CDI standardizes the deprivation factors to be unit neutral, applies statistical shrinkage to account for the imprecise measurement of the factors, updates several factors, and reweights the factors using the most recently available data. Validation tests suggest the CDI exhibits higher correlations with several health outcome/utilization measures than the ADI. The CDI will better serve policymakers by improving identification of urban areas with higher deprivation.