The validity and utility of risk assessment for inpatient suicide

It is widely assumed that identifying clinical risk factors can allow us to determine which patients are at high risk of suicide while in hospital, and that identifying those patients can help prevent inpatient suicide. We aimed to examine the validity and utility of categorizing psychiatric patients to be at either high or low risk of committing suicide while in hospital.

The assumption that high-risk categorizations are valid was examined by comparing factors included in high-risk models derived from individual studies of inpatient suicide with the results of a meta-analysis of factors associated with inpatient suicide. A valid high-risk model was then applied to a hypothetical clinical setting in order to test the assumption that high-risk categorizations are useful.

The existing models for assessing whether inpatients are at high risk of suicide all include one or more factors that were not found to be associated with inpatient suicide by meta-analysis and were probably chance associations. Depressed mood and a prior history of self-harm are the only well-established independent risk factors for inpatient suicide. Using these risk factors to classify patients as being at high or low risk would prevent few, if any, suicides, and would come at a considerable cost in terms of more restrictive care of many patients and the reduced level of care available to the remaining patients.

Risk categorization of individual patients has no role to play in preventing the suicide of psychiatric inpatients.