S115. EVALUATION OF THE CLINICAL UTILITY OF SYMPTOM DIMENSIONS ON LONG-TERM CLINICAL AND FUNCTIONAL OUTCOMES IN FIRST EPISODE PSYCHOSIS

BACKGROUND: Current clinical utility of diagnostic categories in patients with psychosis is in debate. Alternatively, symptom-based dimensional approaches are suggested, but research on their utility and longitudinal stability is at its early phases, showing lack of consistencies. The aims of this prospective study are: 1) to test the stability of structure of symptom dimensions in first episode psychosis patients; 2) to explore the utility of symptom dimensions in predicting clinical and functional outcomes. METHODS: This study included a total of 208 with first episode of psychosis-spectrum disorders aged 18–65 years who presented to psychiatric services in South London, recruited as part of the Genetics and Psychosis Outcome (GAP) case-control study and EU-GEI multicentre case-control study. A subsample of 114 patients were traced after a mean of 6.5 years. Psychopathology was assessed at baseline and at follow-up using OPCRIT and tested with bifactor model, encompassing one general psychosis dimension and five specific symptom dimensions (positive, negative, disorganisation, mania, and depression). Follow up measures on functional outcome (assessed with GAF scale) and clinical outcomes (number and total length of hospitalisations) were derived from clinical records. Predictor role of baseline symptom dimensions was tested by multiple linear regression to predict global functioning; and by negative binomial regression for length of hospitalisation and number of hospital admissions. RESULTS: Factor loadings of disorganisation dimension were most likely to change longitudinally while loadings of positive dimension were most stable. Regarding dimension stability over time, all positive, disorganised and depressive symptoms significantly improved over time while manic and negative symptoms did not significantly differ. In terms of prediction of outcome, baseline manic symptoms were associated with reduced risk of hospitalisation (adj OR 1.56; 95% CI 1.01–2.38), reduced length of hospitalisation (adj IRR= 0.73; 95% CI 0.56–0.95) and better global functioning (β=5.21; 95% CI 2.46–7.95) at follow up. Similarly, depressive symptoms were associated with reduced length of hospitalisation (adj IRR= 0.77; CI 0.61–0.97). On the other hand, baseline positive symptoms were associated with increased risk of hospitalisation (adj OR 1.93; 95% CI 1.25–2.96). No other significant associations were found between the rest of symptom dimensions and outcomes. DISCUSSION: This study provides new evidence on the longitudinal stability of bifactor model of psychosis and, shows that all except manic and negative symptoms significantly improved over time. Whereas affective symptoms (including mania and depression) were associated with good prognosis, positive symptoms seem to predict poor clinical outcomes. The particular and different influence of affective and psychotic symptoms on long-term functional and clinical outcomes may have therapeutic implications and support the potential clinical utility of incorporating symptom-based approach in further outcome research.