Objective: It remains unclear whether very early onset psychosis (VEOP; ≤12years of age) and early onset psychosis (EOP; onset 13-17years of age) are homogeneous in their clinical presentation. We investigated the predictive value of age of psychosis onset for severity, functioning and demographic variation by: 1) comparing groups based on traditional cut-offs for age of psychosis onset, and 2) using receiver operating characteristic (ROC)-curve calculations, without a priori age of onset cut-offs.
Method: Participants were 88 (45 female, 43 male) children and adolescents with a recent onset of psychosis (age range=6.7-17.5years; M=13.74, SD=2.37).
Results: The VEOP group had significantly shorter duration of untreated illness and untreated psychosis, and lower functioning than the EOP group. The VEOP and EOP groups did not differ significantly on gender proportion, urbanicity, psychotic diagnosis, family history of psychotic disorder, psychotic, depressive and anxiety symptoms or IQ. When applying ROC-curves to the lowest three quartiles of positive psychotic symptoms scores, the optimal age-cut-off was 14.0years (sensitivity=0.62; specificity=0.75). For the highest quartile of functioning scores, the optimal differentiating cut-off for age of psychosis onset was 14.7years (sensitivity=0.71; specificity=0.70).
Conclusions: Larger samples of patients, assessed at presentation and followed-up, are necessary to clearly examine clinical presentation and outcome as a function of social and neural development to better understand if the differentiation between VEOP and EOP is justified. This will aid the development of predictive diagnostic tools, more accurate prognosis prediction, and age-tailored therapeutic interventions.
Keywords: Childhood onset; First-episode psychosis; ROC-curves; Schizophrenia; Very early and early onset psychosis.
Copyright © 2015 Elsevier B.V. All rights reserved.