What's the risk? The frequency of false positives in psychosis risk screening instruments.

Abstract

One way to facilitate early intervention in psychosis is by improving early identification of individuals with psychosis risk through screening. However, screening instruments for psychosis risk have not incorporated the low base rates of psychosis risk and psychosis into their psychometric properties, resulting in an overestimation of the instruments' utility. This study evaluated the false positives of psychosis risk screening instruments using methods that incorporate base rates. Five psychosis risk screening instruments were selected: the Prodromal Questionnaire (PQ), the Prodromal Questionnaire-Brief Version, the Prodromal Questionnaire-16-item version, the Abbreviated Youth Psychosis at Risk Questionnaire, and the PRIME screen. We calculated the long-run probability of false positives for the author-recommended cutoff scores of each instrument. The incorporation of base rates revealed high frequencies of false positives for all psychosis risk screening instruments despite all selected instruments demonstrating acceptable sensitivity and specificity. Using base rates that reflect psychosis risk in a help-seeking population and in the general population, all instruments resulted in ≥ 95% false positives for individuals scoring at the author-recommended cutoff score and ≥ 81% false positives for individuals scoring at or above the author-recommended cutoff score. The PQ had the highest predictive power for a help-seeking sample, demonstrating that the instrument provides 17% additional diagnostic information beyond chance. Screening instruments for psychosis risk should incorporate base rates into instrument validation and cutoff score determination in order to improve identification of psychosis risk in United States mental health settings where these individuals are likely to present. (PsycInfo Database Record (c) 2021 APA, all rights reserved).