Every year, alcohol use disorder (AUD) generates millions of emergency department (ED) visits and hospital admissions, costing the U.S. health sector over $90 billion. These hospital admissions are critical opportunities to start patients on addiction pharmacotherapy, but factors like medication non-adherence and post-discharge relapse contribute to frequent re-admissions. Two single-dose interventions are well suited to facilitate treatment retention and prevent re-admissions due to their prolonged, adherence-independent effects: extended-release (XR) naltrexone injection and intravenous (IV) ketamine infusion. These have not been thoroughly investigated in the hospital setting among high-utilizer, safety-net populations. Therefore, the investigators aim to:

1. Test the feasibility of randomizing hospitalized patients (n=45-60, age 18-65) with multiple AUD-related admissions to treatment with either extended-release (XR) naltrexone, intravenous (IV) ketamine, or no single-dose medication, all with enhanced linkage to care. Feasibility outcomes such as recruitment rate, patient acceptability, post-discharge follow-up rate, and adverse events will help to identify key lessons for a future comparative effectiveness study.
2. Estimate the 30-day re-admission rate for patients randomized to treatment with XR naltrexone, with IV ketamine, or no single-dose medication, all with enhanced linkage to care. The investigators hypothesize that the re-admission rate will be lower for each of the two single-dose medication groups than for the "linkage-alone" group.

Inclusion Criteria:

• Age 18-65
• 1+ alcohol-related* admission(s) or emergency department visit(s) in past 12 mo.
• Has insurance (public or private)
• Seen by inpatient addiction consult service

Exclusion Criteria:

• Known or suspected active COVID-19 infection
• Hepatic: AST/ALT >5x upper-limit of normal, decompensated liver failure
• Renal: Glomerular filtration rate <30ml/min
• Cardiovascular: History of acute coronary syndrome, cerebrovascular event, hypertensive crisis, known cardiomyopathy
• Known elevated intracranial pressure
• Thrombocytopenia (<50/microliter)
• Active moderate/severe withdrawal (based on hospital withdrawal protocol)
• Active delirium (alcohol-related or otherwise)
• Already enrolled in study
• XR naltrexone or IV ketamine in last 30 days
• Known intolerance to naltrexone or ketamine
• Other active severe substance use disorder (tobacco, cannabis excluded)
• Pregnant or breast-feeding, or planning.
• Opioids: chronic, recent (<24h), or anticipated
• Unstable psychiatric illness (active psychosis, active suicidality)
• Moving from region within 30-days of discharge
• Discharge to acute/residential treatment
• Involuntary hold

Immediate effects of IM naltrexone or IV ketamine were assessed immediately after administration in the hospital. Adverse events by PRISE questionnaire were assessed at follow-up visit (targeted at 1 week post-discharge) for those who attended. Deaths and serious adverse events would be detected at any point within 30 days post-discharge from electronic health record alerts, direct patient communication with study team, or other staff communication.

Hospital admission or emergency department visits were not automatically considered serious adverse events. This is because these were the primary/secondary outcomes targeted by the intervention itself.