Objective: To investigate the risk of suicide following contact with mental health services and whether it has changed over time.Method: Record linkage was used to obtain the records of previous hospital admissions and mental health service contacts for deaths due to suicide in the period 1980-98. Standardized incidence rates were calculated for the general population and for users of mental health services. Proportional hazards regression was used to assess risk factors within the cohort of people with mental health service contact.Results: Suicide risk was significantly increased in users of mental health services (rate ratio 6.66 in males and 7.52 in females). Suicide risk was highest in the first 7 days after discharge from in-patient care, and decreased exponentially with time since discharge. Suicide risk in users Of mental health services has increased over the study period.Conclusion: These results highlight the importance of adequate followup of patients discharged from in-patient services, and the need for adequate resources for community-based services.