Categorical versus dimensional models of early psychosis.

ABSTRACT

AIM: Early psychosis is typically operationalized as a categorical construct by dividing people into one of three diagnostic statuses low-risk, clinical high-risk, and first episode psychosis. We empirically assess whether an alternative dimensional approach focused on observed symptom severity may be more desirable for clinical and research purposes. METHODS: Participants were 152 help-seeking youths ages 12-22 years old. Structured interview for psychosis risk syndromes interviews were used to obtain dimensional psychosis symptom severity ratings, and to classify participants by categorical psychosis risk status. Twenty-five participants were classified as having a diagnosable psychotic disorder, 52 participants as clinical high-risk, and 75 participants as help-seeking controls. We assessed the relation between categorical and dimensional measurements of psychosis severity, and then compared categorical versus dimensional psychosis severity in their ability to predict social and role functioning. RESULTS: On average, dimensional psychosis symptom severity increased along with categorical risk status (help-seeking control < clinical high-risk < diagnosable psychotic disorder). There was, however, considerable overlap between categories, with people at clinical high-risk being particularly hard to distinguish from people with diagnosable psychotic disorders on the basis of symptom severity. Dimensional symptom severity was more predictive of functioning than categorical risk status. CONCLUSIONS: Categorical risk status and psychosis symptom severity are related but not interchangeable, and dimensional models of psychosis may be more predictive of functional outcomes. Adopting a dimensional rather than categorical approach to the psychosis risk spectrum may facilitate better predictive models and a richer theoretical understanding of early psychosis.