Background: Repetition of hospital-treated self-poisoning and admission to psychiatric hospital are both common in individuals who self-poison. Aims: To evaluate efficacy of postcard intervention after 5 years. Method: A randomised controlled trial of individuals who have self-poisoned: postcard intervention (eight in 12 months) plus treatment as usual v. treatment as usual. Our primary outcomes were self-poisoning admissions and psychiatric admissions (proportions and event rates). Results: There was no difference between groups for any repeat-episode self-poisoning admission (intervention group: 24.9%, 95% CI 20.6-29.5; control group: 27.2%, 95% CI 22.8-31.8) but there was a significant reduction in event rates (incidence risk ratio (IRR) = 0.54, 95% CI 0.37-0.81), saving 306 bed days. There was no difference for any psychiatric admission (intervention group: 38.1%, 95% CI 33.1-43.2; control group: 35.5%, 95% CI 30.8-40.5) but there was a significant reduction in event rates (IRR = 0.66, 95% CI 0.47-0.91), saving 2565 bed days. Conclusions: A postcard intervention halved self-poisoning events and reduced psychiatric admissions by a third after 5 years. Substantial savings occurred in general hospital and psychiatric hospital bed days.