• Katherine J Sapra, Wenya Yang, Nancy Barker Walczak, and Stephen S Cha.
• Center for Medicare and Medicaid Innovation, Centers for Medicare and Medicaid Services, Baltimore, Maryland.
• Popul Health Manag. 2020 Feb 1; 23 (1): 12-19.

AbstractThe objective of this observational longitudinal study of Maryland fee-for-service Medicare beneficiaries (2015-2016) was to investigate whether using data on neighborhood socioeconomic disadvantage in addition to individual clinical risk data improves identification of high-cost Medicare beneficiaries. Neighborhood socioeconomic disadvantage is measured using the Area Deprivation Index (ADI), a validated composite measure based on publically-available US census data (2011-2015) for Maryland census block groups. Hierarchical Condition Categories (HCC) score, health care utilization, and spending were obtained from Centers for Medicare & Medicaid Services Chronic Condition Warehouse beneficiary file and Part A and Part B claims data (2015). Total cost of care (TCOC) was calculated for 2016. Descriptive and multivariate analyses were performed to examine the relationship of residency in neighborhoods with high ADI and subsequent year health care spending. Among 615,637 Maryland Medicare fee-for-service beneficiaries, those living in neighborhoods with the greatest disadvantage vs. the least disadvantage incur significantly greater costs in the subsequent year (ADI Quintile 5 $12,439 versus Quintile 1 $8920, P < .001). Clinical risk exacerbates this disparity. Among beneficiaries in the highest HCC score quintile, costs are 27% ($5458, P < .001) higher among beneficiaries in the highest compared with the lowest ADI quintiles without risk adjustment and 24% ($4599, P < .001) higher with risk adjustment. Several sensitivity analyses found the relationship between ADI and TCOC robust. Association between neighborhood socioeconomic disadvantage and health care cost is most pronounced among the most clinically complex Maryland Medicare beneficiaries. Using ADI in combination with HCC score may facilitate more precise targeting of care management resources.

21 keywords...

hide…