Initiation of Medications for Alcohol Use Disorder Among Hospitalized Veterans : A Retrospective Cohort Study

Introduction

Hospitalizations for alcohol use disorder (AUD) present critical opportunities to initiate evidence-based pharmacotherapy. This retrospective cohort study examines medication initiation patterns among veterans hospitalized with AUD across the Veterans Health Administration (VHA) system. AUD affects nearly 15 million U.S. adults, yet medications like naltrexone and acamprosate remain underutilized despite strong evidence supporting their efficacy in reducing heavy drinking days and maintaining abstinence.

Research Methodology

The study analyzed 29,041 AUD hospitalizations across 142 VHA facilities during 2022-2023. Researchers tracked initiation of FDA-approved medications for alcohol use disorder (MAUD) either during hospitalization or within 7 days post-discharge. Using multivariable logistic regression models, the team examined associations between MAUD initiation and factors including: patient demographics, clinical characteristics (frailty status, co-occurring opioid use disorder), service type (psychiatry vs. general medicine), ICU admission, and addiction specialty consultations. Statistical analyses accounted for hospital-level variations through fixed effects modeling.

Key Findings

MAUD initiation occurred in only 30.8% of eligible hospitalizations (8,932 cases). Naltrexone was the most prescribed medication (57.9%), followed by acamprosate (16.5%) and injectable naltrexone (13.9%). Timing analysis revealed 69.6% of initiations occurred during hospitalization, with 97.7% of these patients continuing prescriptions within 30 days post-discharge. Significant disparities emerged: older veterans (≥65 years) had 22% lower initiation odds compared to younger patients. American Indian/Alaska Native veterans showed 18% lower odds than White veterans. Other negative predictors included male gender, frailty, concurrent opioid use disorder, and ICU admission. Positive predictors included addiction specialty consultations (2.4-fold higher odds) and psychiatry service care (1.8-fold higher than general medicine). Hospital-level variation was substantial, with median initiation rates of 29.9% (IQR 22.6-36.3%).

Clinical Implications and Barriers

The low overall initiation rate (30.8%) highlights systemic barriers including: knowledge gaps about MAUD efficacy, misconceptions about patient readiness, and workflow challenges during care transitions. The dramatic variation between hospitals (range: 10.2% to 58.7% initiation rates) suggests institutional practices significantly influence prescribing patterns. Addiction consultations demonstrated the strongest positive effect, emphasizing their value in complex cases. The high continuation rate (97.7%) among inpatient starters indicates hospitalization effectively overcomes initial treatment barriers. The study reveals opportunities to: implement standardized AUD order sets, expand addiction consultation services, develop telehealth bridges between inpatient and outpatient care, and address implicit biases affecting prescribing for older patients and racial minorities.

Conclusion and Future Directions

This study identifies actionable strategies to improve MAUD access: hospitals should prioritize addiction consultations, psychiatry service collaborations, and standardized discharge protocols. The documented racial, age, and gender disparities necessitate targeted interventions. Future quality improvement initiatives should examine best practices from high-performing hospitals and develop implementation toolkits for broader dissemination. Given AUD’s significant health burden and high hospitalization costs, increasing MAUD initiation represents a critical opportunity to improve veteran health outcomes and reduce healthcare utilization.

Funding and Disclosures

Primary funding provided by the U.S. Department of Veterans Affairs and National Institute on Aging. Citation: Anderson TS, et al. Annals of Internal Medicine. 2026; PMID: 42081820. Full study available at: https://pubmed.ncbi.nlm.nih.gov/42081820/