100 INPATIENT SUICIDES

Background The study aimed to define the characteristics and assess th e clinical predictability and possible prevention of psychiatric in-pa tient suicides. Method The coroner's files on atl suicides in the Grea ter Montreal Region from 1 April 1986 to 31 March 1991 were examined. The medical records of each case of suspected in-patient suicide were then reviewed and rated for predictability and preventive measures tak en. Results A total of 3079 suicides were recorded over this five-year period (mean annual rate of 16.4 per 100 000 inhabitants). Of these, 104 (3.4%) involved hospital in-patients. Nearly half(48%) of these in -patient suicides occurred outside the hospital setting. The methods m ost frequently employed were hang ng (3 6%) and jumping from high plac es (24%). Patients suffering from an affective disorder (45%) or schiz ophrenia (35%) comprised the majority ofthe sample. Suicides were sign ificantly more predictable in general hospital psychiatric wards. Suic ide prevention measures did not differ significantly across settings. Conclusions The majorityof in-patient suicides were not highly predict able For highly predictable suicides, the results underline the import ance of actively treating and protecting these patients.