Do the Components of Attenuated Positive Symptoms Truly Represent One Construct?

Abstract

Abstract Background and Hypotheses Psychosis-risk inventories, like the Structured Interview for Psychosis-Risk Syndromes (SIPS), utilize symptom components and coalesce the information into a single-severity rating. These components include frequency, duration, in-the-moment conviction, retrospective insight, distress, and effect on social/role functioning. While combining components distills a great deal of important information into one practical symptom rating, this approach may mask important details of the greater clinical picture. Study Design Individuals at clinical high risk for psychosis (n = 115) were assessed with the SIPS Score Separable Components (SSSC) scale, created to accompany the SIPS positive items by dividing each item into the 7 components identified above. The latent structure of the SSSC was identified with an exploratory factor analysis (EFA). The factors were followed up with validation analyses including hypothesized cognitive, functioning, and symptom measures. Finally, clinical utility analyses were conducted to understand relationships between psychosis risk and common comorbidities. Study Results EFA revealed that the SSSC had 3 interpretable factors with the appropriate fit (rmsr = 0.018, TLI = 0.921): Conviction (in-the-moment conviction, retrospective insight), Distress-Impairment (distress, social/role functioning), and Frequency/Duration (frequency, duration). Conviction was minimally valid, Distress-Impairment had excellent validity, and Frequency/Duration was not related to any of the candidate validators. Conviction significantly predicted elevated psychosis risk. Distress-Impairment was related to common comorbid symptoms. Notably, the factors associated more strongly with clinical features than the traditional SIPS scores. Conclusions The SSSC offers a supplemental approach to single-severity ratings, providing useful clinical insight, mechanistic understanding, and the potential for better capturing heterogeneity in this population.