- Design
- Systematic review, 17 studies (12 RCTs, 3 non-randomised trials, 2 cohorts), GRADE
- Population
- Adults hospitalised with alcohol use disorder; 1,917 in RCTs
- Primary outcome
- Linkage to outpatient AUD follow-up after discharge
- Effect
- 8 of 21 interventions improved linkage; RCT effects ranged −8% to +30% absolute
A systematic review in Annals of Internal Medicine (published 8 September) pooled 17 studies of 21 hospital-based interventions meant to get patients admitted with alcohol use disorder (AUD) to outpatient follow-up: 12 randomised trials (1,917 patients), three non-randomised trials and two cohort studies. Most interventions had several parts — behavioural (76%), care coordination (76%), social support (33%) and AUD medication (14%).
Only 8 of the 21 (38%) improved linkage to outpatient care. Across the randomised trials the effect ranged from an 8% absolute fall to a 30% absolute rise in follow-up. The interventions that worked generally combined behavioural support and care coordination with either AUD medication or social support. Certainty was low, because of risk of bias and imprecision.
The point for a ward physician is that a leaflet and a referral letter is the approach this evidence least supports. The admission is the moment the patient is in front of you, and the thing least often done — starting medication — is part of the combinations that worked.
- Document an AUDIT-C or equivalent on admission for anyone with an alcohol-related presentation
- Offer AUD medication before discharge, not as a suggestion for the clinic
- Book the follow-up appointment before the patient leaves; do not leave it to the patient
- Name the person — addiction team, nurse, social worker — who will contact the patient after discharge
- Check whether the patient has somewhere safe to go and anyone to support abstinence
Why it matters
Most hospital efforts to link AUD patients to follow-up fail, and the ones that work bundle coordination with treatment rather than advice alone.
Don't overread it
The evidence is low certainty and heterogeneous — it points to a direction, not a proven programme.
The statistics, in plain English
Eight of 21 interventions helping is not a pooled effect — the studies were too different to combine into one number. The range from an 8% fall to a 30% rise means some programmes did nothing or worse. Low GRADE certainty means further trials could easily change the picture.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for internal medicine, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free