AIMS: To trial two novel methods of recruiting people who experience non-fatal heroin overdose through the ambulance service. SETTING: Melbourne and Sydney, Australia. METHODS: In Melbourne potential participants were given numbered contact cards by ambulance paramedics after revival, while in Sydney potential participants were approached after revival by a researcher who travelled with ambulance paramedics to the overdose scene. RESULTS: In Melbourne 281 cards were distributed during the period 1 June 1998-31 December 1998 and a subsequent contact rate of 24% was achieved with 14% attending a subsequent interview. In Sydney there were 170 initial contacts of which 139 (82%) answered a series of questions asked at the scene (the remainder either ineligible or incapable of answering questions) with 48 (35%) also attending for follow-up interviews. CONCLUSIONS: Recruitment through contact with ambulance services is a novel method of recruiting heroin users for research into non-fatal heroin overdose with advantages over other methods of sampling for research on non-fatal heroin overdose.