The risk of self-harm and violent behaviour by psychiatric patients is an important clinical research topic. However, a literature search on measurement of patients’ structured self-reported assessment of future risk of violent, suicidal or self mutilating behaviour failed to disclose any published empirical research. In the acute psychiatric ward at Ålesund Hospital in Norway we conducted a study of risk assessment of self harm and violent behaviour at admittance and at discharge, followed by prospective measurement of occurred episodes.

All admitted patients during one year (489 persons with 716 stays) were included. At admittance and at discharge the Violence Risk Screening-10 (V-RISK-10), items taping suicidal thoughts and behaviour, and patients’ structured self-reported risk of self-harm and violent behaviour (583 self-reports) were scored (six items: no risk-, little-, moderate-, high-, very high risk, don't know the risk, will not answer). The patients were asked to give a blood sample to measure lipids and serotonin. Incidents of aggressive behaviour (against others and self-directed) were monitored during hospital stay (phase 1) and 3, 6, 9 and 12 months after discharge (phase 2).

Preliminary results pertaining to the relationship between the patients’ predictions of self-harm and violent behaviour and occurred episodes during the subsequent hospital stay (phase 1) and the first three months after discharge (phase 2), indicate that the predictive validity of patients’ self-report was highly significant both for violent behaviour, suicidal behaviour and self mutilation in phase 1 and phase 2. Effect sizes ranged from moderate to high.