Sampling from different populations: Sociodemographic, clinical, and functional differences between samples of first episode psychosis individuals and clinical high-risk individuals who progressed to psychosis.

Over the past two decades, research and clinical resources on clinical high risk (CHR) for psychosis have both expanded, with goals to better understanding risk and protective factors on the course of illness and inform early intervention efforts. However, some studies have highlighted potential sampling bias among CHR research studies, raising questions about generalizability of findings and inequitable access to early detection and intervention. The current study sought to explore these questions by comparing 94 participants in a CHR longitudinal monitoring study across North America (NAPLS-2) who converted to syndromal psychosis over the course of the study (CHR-CV) to 171 participants who presented for treatment at a localized first-episode psychosis service (FES) after converting. CHR-CV participants were significantly more likely to be White and have a college-educated parent, while FES participants were more likely to be Black and first- or second-generation immigrants. On average, CHR-CV participants were younger at onset of attenuated positive symptoms, had a longer period of attenuated symptoms prior to conversion, and were more likely to be treated with antipsychotics prior to conversion compared to those in FES programs. After controlling for time since conversion, CHR-CV participants had higher global functioning and were less likely to have experienced recent psychiatric hospitalization. Findings suggest that CHR research and FES clinics may be sampling from different populations, although conclusions are limited by inconsistent sampling frames and methods. Integrated early detection that targets defined geographic catchments may deliver more epidemiologically representative samples to both CHR research and FES.