Evaluation of verbal list learning as a predictor of psychosis

Aim We explored whether cognitive performance, and verbal learning in particular, predicts psychosis or psychiatric hospitalizations among unselected first‐admission adolescent patients in general psychiatric care. Methods Up to 152 adolescents aged 15–18 were interviewed with the SIPS, tested with a cognitive test battery in the beginning of their psychiatric treatment, and followed for a maximum of 9 years (median 4.5 years). Results The composite factors of processing speed, verbal performance and visuospatial performance did not predict psychosis (n = 7) or all‐cause psychiatric hospitalizations (n = 26) beyond psychosis risk symptoms. However, those who developed psychosis performed worse on California Verbal Learning Test (CVLT) compared to other adolescents. Lower scores of CVLT immediate recall predicted psychosis (P = .003, HR = 1.13 per CVLT point decrease). However, when general verbal ability was adjusted for, CVLT did not reach significance. Conclusions Impaired verbal list learning may predict psychosis also among adolescent psychiatric patients not preselected for psychosis risk suspicion.