Lay Summary

We used statistics to explain differences in responses on a clinical questionnaire in 342 young people (12-15 years old) with a range of mental health disorders who were also at ultra-high risk of developing psychosis. We found a shared underlying factor of poor mental health in individuals’ responses and high overlap among seemingly different mental health domains. Our findings suggest that youth at ultra-high risk for psychosis share a general symptom domain and this may improve our understanding of mental health disorders.

Background

Research has increasingly shifted its focus from categorical to dimensional conceptualizations of mental disorders. This is supported by the high amount of overlap among disorders, particularly psychosis spectrum disorders, which cut across traditional diagnostic boundaries. The growing body of evidence on transdiagnostic symptom dimensions has stimulated debates about aetiology and symptom trajectories. While there is evidence for a general factor of psychopathology in service users with schizophrenia, schizoaffective disorder, and psychotic bipolar I disorder, transdiagnostic dimensions of psychopathology have not been replicated in young individuals at ultra-high risk (UHR) for psychosis. The aims of the proposed analyses are to investigate: 1) whether transdiagnostic dimensions of psychopathology can be replicated in UHR individuals; 2) the diagnostic utility for classifying UHR individuals correctly into criteria of a) UHR (trait vulnerability, attenuated psychotic symptoms, BLIPS) and b) comorbid DSM diagnoses; and 3) associations between demographic, clinical, and social variables and transdiagnostic dimensions at baseline.

Methods

Multidimensional item-response modeling was conducted on symptom ratings of the brief psychiatric rating scale (BPRS) at baseline in the staged treatment in early psychosis (STEP) trial, which aims to determine the most effective type, timing, and sequence of interventions in the UHR population.

Results

In total, 342 help-seeking young people enrolled in the study. A bifactor model with one general symptom dimension and four specific dimensions of positive symptoms, negative symptoms, affect, and activation provided the best model fit. This lends support to the notion of a shared general factor across the risk syndrome, pointing at a pluripotent risk state, while simultaneously recognizing the contribution of single, domain-specific factors.

Conclusion

These findings shed light on the dimensionality of symptoms in youth at UHR for psychosis and highlight the importance of further investigating transdiagnostic phenotypes at developmentally early stages of psychopathology.