Self-harm risk assessment — medical first then structure

Sequence: medical stabilisation and antidotes first (paracetamol level/NAC, ECG, gas, glucose) — confirm sequence against local toxicology protocol. Do not complete psychosocial risk assessment while intoxicated — observe until sober enough for reliable history, per local psychiatry-ED protocol.

Structured interview: ideation (intensity/frequency), plan (specificity/lethality), intent (now vs at time of act), access to means, protective factors, triggers, psychiatric history. Prior attempt history is a studied reattempt risk factor in the held meta-analysis.^[1]

Tools: SAD PERSONS shows high specificity but low sensitivity (nearly half of the suicide group classified low-risk) — structured clinical judgement drives disposition, not a score alone.^[3]

Disposition: high (active intent + plan + means) → admit, 1:1, means removal, consider MHA; moderate → urgent psychiatry before leave + safety plan; lower risk with strong supports → safety plan, crisis contacts, rapid community follow-up. Safety-planning interventions are supported for reducing ideation/behaviours and hospitalization in the held review.^[2] Confirm tiers against local psychiatry-ED protocol.

Means restriction: firearms, stockpiled medicines, ligature risks — confirm the key steps against local psychiatry-ED protocol.