Prevalence and predictors of depressive symptoms in non-affective psychosis patients in Jockey Club Early Psychosis (JCEP) Project in Hong Kong

This journal suppl. entitled: Abstracts of the 4th Biennial Schizophrenia International Research ConferenceBACKGROUND: Depressive symptoms are frequently found in the early stage of psychotic disorders and are associated with heightened risk of suicide. In this study, we aimed to examine the prevalence and factors predictive of the occurrence of depressive symptoms in adult patients presenting with first-episode non-affective psychosis to JCEP project in Hong Kong. METHODS: A total of 343 patients aged 26 to 55 years presenting with first-episode non-affective psychosis (DSM-IV schizophrenia, schizophreniform disorder, delusional disorder, brief psychotic disorder or psychosis NOS) to Jockey Club Early Psychosis (JCEP) project were recruited. Patients who scored 6 or above in Calgary Depression Scale (CDS) were classified as having clinically significant depressive symptoms. Premorbid functioning was examined by Premorbid Adjustment Scale (PAS). Psychopathology was assessed with Positive and Negative Syndrome Scale (PANSS). Global functioning was measured using Social and Occupational Functioning Assessment Scale (SOFAS). Medication Compliance Questionnaire (MCQ) was used to assess patients’ attitude to treatment. Drug-induced Parkinsonism side-effects were evaluated using Simpson Angus Scale (SAS). A series of univariate analyses were conducted to determine relationship of “depressed clinical status” with potential predictive variables, followed by multivariate binary logistic regression to identify predictors of depressive symptoms. RESULTS: Fifty-two patients (14.8%) had clinically significant depressive symptoms at study entry. No significant difference between depressed and non-depressed patients in age at entry and onset, sex, premorbid adjustment and duration of untreated psychosis. Univariate analyses revealed significant associations of “depressive clinical status” with educational attainment (p<0.05), log-transformed duration of untreated illness (LogDUI) (p<0.05), SOFAS rating (p<0.05), PANSS positive (p<0.05) and negative symptom (p<0.05) subscale scores, SAS score (p<0.05) and MCQ attitude subscale score (p<0.05). Multivariate logistic regression indicated that more severe positive and negative symptoms, worse drug-induced Parkinsonism side-effects, and poorer attitude to medication treatment independently predicted occurrence of clinically significant depressive symptoms (Nagelkerke R2=0.215, p<0.001). Depressed patients had significantly poorer global functioning than non-depressed counterparts (t=3.6, p