Comparing Alternative Approaches to Care Management Prioritization: A Prospective Comparative Cohort Study of Acute Care Utilization and Equity Among Medicaid Beneficiaries

OBJECTIVE: To test whether prioritizing Medicaid beneficiaries for outreach by population health teams based on predicted benefit from care management services reduces acute care utilization more effectively than traditional prioritization approaches based on predicted risk of future acute care utilization. DATA SOURCES/STUDY SETTING: Healthcare utilization data among 9266 adult Medicaid beneficiaries assigned to primary care practices in Washington State (May-December 2024). STUDY DESIGN: Prospective cohort study. A difference-in-differences analysis was conducted to compare patients between two clusters of clinics using alternative prioritization strategies for population health care management outreach: one cluster using risk-based prioritization (N = 4572 patients prioritized by predicted probability of future acute care utilization without intervention), while the other changed from risk-based to benefit-based prioritization (N = 4694 prioritized by predicted reduction in probability of future acute care utilization with versus without intervention). DATA COLLECTION/EXTRACTION METHODS: Claims data were collected from the Medicaid health plans of patients, while emergency visit and hospitalization notices were obtained from an admit, discharge and transfer data feed service. PRINCIPAL FINDINGS: In the intention-to-treat analysis of all 9266 patients, benefit-based prioritization reduced acute care visits by 92.4 per 1000 member-months compared with risk-based prioritization (95% CI: -113.0, -72.0; p < 0.001). Among the 2845 successfully engaged patients, benefit-based prioritization was associated with 208.4 fewer visits per 1000 member-months (95% CI: -284.0, -133.0; p < 0.001). Outreach and engagement rates were similar between groups, and effects were also consistent across sex and race/ethnicity subgroups. CONCLUSIONS: Prioritizing Medicaid care management outreach based on predicted benefit from such services, rather than predicted risk of acute care utilization alone, was associated with similar engagement rates but substantially lower acute care utilization. This approach may improve the effectiveness of resource-constrained Medicaid programs.