Co-Occurring Substance Use Disorders and 30-Day Psychiatric Readmission in Older Medicare Beneficiaries: A Retrospective Cohort Study

Background and Purpose To characterize the co-occurrence of substance use disorder (SUD) and primary psychiatric disorders in a Medicare inpatient population and quantify their combined effect on two hospital quality indicators: length of stay (LOS) and 30- and 90-day all-cause readmissions, to inform care coordination and quality improvement. Methods A retrospective cohort analysis was conducted using three-step hierarchical multivariable regression models. These models tested main effects and dual-diagnosis interactions while controlling for demographics, admission type, and comorbidities. Active SUD, primary psychotic disorders, primary mood disorders, and substance-induced psychiatric episodes were identified using ICD-10-CM codes (excluding remission codes). Extreme length of stay (LOS) outliers were universally capped at a maximum of 36 days across all diagnostics. Setting: United States of America, Centers for Medicare & Medicaid Services (CMS) patients across the country. Participants: 1,219,325 admissions from the CMS Limited Data Set (LDS) inpatient claims files (5% sample, 2022-2024). Outcome Measures: Hospital LOS and 30- and 90-day all-cause readmissions. Results Primary mood disorders were present in 19.1% of admissions, while SUD, active alcohol use disorder (AUD), and primary psychotic disorders each appeared in approximately 3%. Co-occurrence was significant across all combinations (p < 0.001). Primary psychotic disorder was a significant driver of prolonged LOS (b = 4.45, p < 0.001). Interactions for primary psychotic disorders with active alcohol (b = -1.19) and active SUD (b = -0.90) were sub-additive (i.e., having both conditions does not double the stay, but rather plateaus due to discharge capacity constraints), yielding net LOS increases of ∼3.3 and ∼3.5 days. Conversely, interactions with primary mood disorders were super-additive, prolonging stays. Significant multiplicative interactions for 30-day readmission odds were observed across combinations, most notably for active alcohol with primary mood disorder (OR = 1.26, 95% CI = 1.18–1.34) Conclusion While co-occurring SUD and primary psychiatric disorders prolong inpatient stays, they interact sub-additively; the dual-diagnosis patient population faces elevated 30-day readmission risk and risky care transitions; therefore, quality improvement strategies for integrated medical-behavioral care, structured discharge planning, and post-acute care coordination are needed.