Psychosis screening questionnaire: Exploring its factor structure among South African adults

Abstract

Background: Early detection of psychosis improves treatment outcomes, but there is limited research evaluating the validity of psychosis screening instruments, particularly in low-resourced countries. Aim: This study aims to assess the construct validity and psychometric properties of the psychosis screening questionnaire (PSQ) in South Africa. Setting: This study was conducted at several health centres in the Western and Eastern Cape provinces in South Africa. Methods: The sample consisted of 2591 South African adults participating as controls in a multi-country case-control study of psychiatric genetics. Using confirmatory factor analysis and item response theory, we evaluated the psychometric properties of the PSQ. Results: Approximately 11% of the participants endorsed at least one psychotic experience on the PSQ, and almost half of them (49%) occurred within the last 12 months. A unidimensional model demonstrated good fit (root mean square error of approximation [RMSEA] = 0.023, comparative fit index [CFI] = 0.977 and Tucker–Lewis Index [TLI] = 0.954). The mania item had the weakest association with a single latent factor (standardised factor loading = 0.14). Model fit improved after removing the mania item (RMSEA = 0.025, CFI = 0.991 and TLI = 0.972). With item response theory analysis, the PSQ provided more information at higher latent trait levels. Conclusion: Consistent with prior literature, the PSQ demonstrated a unidimensional factor structure among South Africans. In our study, the PSQ in screening for psychosis performed better without the mania item, but future criterion validity studies are warranted. Contribution: This study highlights that PSQ can be used to screen for early psychosis.