A case-control study of 92 cases of in-patient suicides

Background: A significant number of patients committed suicide while receiving in-patient treatment in psychiatric hospitals. Most previous studies on psychiatric in-patient suicides were conducted in the West. This study aimed to describe the characteristics and identify risk factors of suicides occurring during psychiatric in-patient care in Hong Kong. Method: The case record data of suicide cases (Coroner's verdicts of suicides and undetermined deaths) from all public psychiatric hospitals in the entire region within a 3 years' period (N=93) were compared with matched controls. Results: In-patient Suicide rate was 269/100,000 admissions. Majority had schizophrenia. Suicide usually occurred after the first month of admission, during leave, and by jump from heights. There were little case-control differences in treatment received. Multiple conditional logistic regression found 5 risk factors: previous history of deliberate self-harm (OR-4.60, 95% CI= 1.57-13.5): admitted because of suicidal behaviour (OR=3.92, 95% CI= 1.3-11.9); depressive symptoms at time of suicide (OR=8.53! 95% CI=1.4-52); away without leave at anytime during index admission (OR= 17, 95% CI=1.76-163); and extrapyramidal side effects/akathisia at time of suicide (OR= 10.8, 95% CI= 1.75-66.7). Limitations: Retrospective case record review depended on non-standardized and variable quality of case notes entry. Matching for hospitals in this study would make the comparison between hospitals impossible. Although this is the second largest case-control study of psychiatric in-patient suicide, the estimated power suggested subtle risk factors would be missed. Conclusion: Majority of in-patient suicides occurred at a time of perceived low risk. A high sensitivity to the risk of suicide and vigorous treatment of depressive symptoms were indicated. The care processes during the index admission could bear strong influences on the risk of in-patient suicides. (c) 2005 Elsevier B.V. All rights reserved.