Prediction of suicidal behavior in young people presenting with first-episode psychosis in Hong Kong: a 3-year follow-up study

This journal suppl. entitled: Abstracts of the 4th Biennial Schizophrenia International Research ConferencePoster no. M3BACKGROUND: Suicide behaviours are common in the early stage of psychotic disorders. The present study aimed to examine the rate and predictors of suicidal behaviours in the initial 3 years of treatment for patients presenting with first-episode psychosis to a territory-wide specialized early intervention program, namely EASY (Early Assessment Service for Young people with psychosis) in Hong Kong. METHODS: Seven hundred patients aged 15-25 years presenting with first episode psychosis (based on ICD-10 criteria) to EASY program between July 2001 and August 2003 were included in the study. Of the initial cohort, 546 completed the 3-year follow-up and thus constituted the final study sample. Demographics, past history of suicidal attempts, substance abuse, baseline and follow-up clinical and functioning variables were collected via systematic medical record review based on standardized protocol. Suicidal behaviour was defined as either attempted or committed suicide. A series of univariate logistic regression analyses were performed to examine the relationship of suicidal behaviour (yes/no in 3-year follow-up) with potential predictor variables, followed by multivariate regression model to determine independent predictors of suicidal behaviours. RESULTS: There were no significant differences between completers and non-completers in socio-demographics, past suicidal attempt, baseline clinical and functional measures with the exception that completers were significantly more likely to have schizophrenia diagnosis. By the end of 3-year follow-up, 11.2% (n=61) of patients exhibited suicidal behaviour over the study period, including 1.3% (n=7) committing suicide. Univariate analyses revealed that hospitalization at intake (OR=0.45, p<0.05), past history of substance abuse (OR=0.43, p<0.05), pre-treatment suicidal attempt (OR=0.38, p<0.01), and baseline social functioning (OR=0.97, p<0.01) were identified as the risk factors for suicidal behaviours. Multivariate regression demonstrated that only pre-treatment suicidal attempt, past history of substance abuse and baseline functioning independently predicted the occurrence of suicidal behaviours during 3-year follow-up period (Nagelkerke R2 = 0.064, chi-square= 21.8, p