Latent Class Analysis to Represent Social Determinant of Health Risk Groups in the Medicaid Cohort of the District of Columbia.

Objective: To develop distinct social risk profiles based on social determinants of health (SDH) information and to determine whether these social risk groups varied in terms of health, health care utilization, and costs.
Methods: We prospectively enrolled 8943 beneficiaries insured by the District of Columbia Medicaid program between September 2017 and December 2018. Participants completed a SDH survey and we obtained their Medicaid claims data for a 2-year period before study enrollment. We used latent class analysis (LCA) to identify distinct social risk profiles based on their SDH responses. We assessed the relationship among different SDH as well as the relationship among the social risk classes and health, health care use and costs.
Results: The majority of SDH were moderately to strongly correlated with one another. LCA yielded 4 distinct social risk groups. Group 1 reported the least social risks with the most employed. Group 2 was distinguished by financial strain and housing instability with fewer employed. Group 3 were mostly unemployed with limited car and internet access. Group 4 had the most social risks and most unemployed. The social risk groups demonstrated meaningful differences in health, acute care utilization, and health care costs with group 1 having the best health outcomes and group 4 the worst (P<0.05).
Conclusions: LCA is a practical method of aggregating correlated SDH data into a finite number of distinct social risk groups. Understanding the constellation of social challenges that patients face is critical when attempting to address their social needs and improve health outcomes.
Competing Interests: The authors declare no conflict of interest.
(Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.)