Suicide rates in psychiatric in-patients: an application of record linkage to mental health research

David Lawrence, D'Arcy Holman, A.V. Jablensky, Stuart Fuller

Research output: Contribution to journalArticlepeer-review

40 Citations (Scopus)


Objective: To study trends in the rate of suicide in psychiatric patients in Western Australia. To examine the associations of suicide with demographic and clinical factors.Methods: A population-based cohort of 52,010 individuals whose first psychiatric admission occurred in 1980-95 was identified from the Health Services Research Linked Database. There were 471 deaths by suicide by 31 December 1995. Age standardised suicide rates per 1,000 person-years at risk were calculated. Suicide rates in the first year after a patients first admission were also examined and a proportional hazards regression analysis was performed to examine risk factors for suicide.Results: Male psychiatric patients were 3.4 times more likely to commit suicide than female patients (95% CI 2.76-4.24). Younger patients were at higher risk than older patients, and patients with extended periods of in-patient treatment were at more than double the risk of short-stay patients. Over the 16-year period, the rate of suicide in the first year after first psychiatric admission was found to increase by 3.4% a year (95% CI -0.7-7.6%).Conclusions: The findings confirm that psychiatric patients are at high risk of suicide. Patient outcomes in terms of risk of suicide after hospital discharge have deteriorated.Implications: Improvements are needed in the provision of community support to high risk psychiatric patients. Further work should be done to identify patients at highest risk of suicide.
Original languageEnglish
Pages (from-to)468-470
JournalAustralian and New Zealand Journal of Public Health
Issue number5
Publication statusPublished - 1999


Dive into the research topics of 'Suicide rates in psychiatric in-patients: an application of record linkage to mental health research'. Together they form a unique fingerprint.

Cite this