Clinical insight in first-episode psychosis: Clinical, neurocognitive and metacognitive predictors

Objective: Poor insight is a major problem in psychosis, being detrimental for treatment compliance and recovery. Previous studies have identified various correlates of insight impairment, mostly in chronic samples. The current study aimed to determine clinical, neurocognitive, metacognitive, and socio-cognitive predictors of insight in first-episode psychosis.Methods: Regression analyses of different insight dimensions were conducted in 190 patients with first-episode psychosis. Measures of clinical symptoms, neurocognition, metacognition, social cognition, and 'jumping to conclusions' bias were entered as predictors. Results: Delusions, disorganisation, and certain negative symptoms were associated with unawareness in various domains, while depression was associated with greater awareness of illness. Deficit in theory of mind and selfreflective processes, as well as a 'jumping to conclusions' bias, contributed to poor insight. Several neuropsychological scores also contributed to this but their contribution was no longer observed in regression analyses that included all the previously identified clinical and cognitive predictors. A measure of perseverative errors was still associated with unawareness and misattribution of symptoms.Conclusion: In models that account for 28 % to 50 % of the variance, poor insight in first-episode psychosis is mainly associated with delusions and certain negative symptoms. At the cognitive level it does not appear to result from neuropsychological impairment but rather from altered reasoning bias and dysfunction inmetacognitive processes. Therapeutic strategies specifically directed at these mechanisms could help improve the evolution of insight in first episode psychosis.