Opioid Overdoses Among High-Risk Medicaid Members: Healthcare Cost, Service Utilization, and Risk Factor Analysis

Abstract

Research Objective: Identify risk factors associated with opioid overdoses among three high-risk populations of Medicaid members related to cost and service utilization. Study Design: Repeated cross-sectional study using five years of Massachusetts Medicaid (MassHealth) claims and state agency data. Population Studied: MassHealth members aged 11-64 years considered to be high-risk (homeless, unstably housed, and/or criminal justice-involved) and in need of support services, especially those with extensive behavioral health (BH) needs. These three populations were identified as being particularly vulnerable to non-fatal and/or fatal opioid overdoses. Principal Findings: MassHealth members who were both justice-involved and unstably housed were at much higher risk of an opioid overdose than the MassHealth population overall, especially those with a substance use disorder (SUD) or a serious mental illness (SMI). Experiencing both homelessness and justice involvement substantially compounded members’ non-fatal overdose risk, regardless of BH diagnosis. Co-occurring SUD/SMI was a key driver of high overdose prevalence, particularly among the justice-involved. Compared to MassHealth members in general, those with justice involvement and unstable housing had costs that were 50-65% higher; members who experienced homelessness had triple the costs. Healthcare service use both before and after an overdose was relatively low, including the timeframe between multiple non-fatal overdoses. In multivariate analyses, all three high-risk factors (i.e., populations) were significantly related to an increased opioid overdose risk after controlling for additional risk factors (BH diagnoses, chronic medical conditions, and demographic characteristics). Males and whites were more likely to have an opioid overdose; those with diabetes or hypertension were less likely. These results were similar when assessing various opioid overdose outcomes (non-fatal and/or fatal). Conclusions: These findings helped inform MassHealth’s understanding of its members’ experiences regarding medical and BH services, especially among high-risk populations with an opioid overdose. The identification of risk factors most predictive of a subsequent overdose may help address the needs of these high-risk groups. For most of the populations studied, prevalence of co-occurring BH diagnoses was much higher than MassHealth members in general and appeared to impact opioid overdose rates. Most members received services for 1-2 months in both the pre- and post-overdose periods; service use was relatively low in the year following a non-fatal overdose, suggesting retention was also low. Multivariate analyses consistently showed that gender and race were significantly associated with increased overdose risk. Implications for Policy or Practice: Understanding opioid overdose risk factors and identifying service utilization gaps and missed opportunities are important. As payment reforms evolve under the umbrella of accountable care organizations, BH community partnership models are key for collaborating with healthcare and social service providers, and community resources for care management, care coordination, and referrals to support services. Our study initially developed an in-depth descriptive analysis of individuals with SUD, SMI, or both identified as being at high risk for an opioid overdose. Understanding service trajectory and outcomes through additional analyses was critical for planning and prioritizing appropriate services. As payors are actively making decisions about effective systems of care, they are particularly interested in understanding the need for community-based and residential services, particularly for those with housing instability and/or criminal justice involvement.