BACKGROUND AND HYPOTHESIS: Using data from the EU-GEI Work Package 2 (EU-GEI WP2) programme, we sought to test several hypotheses related to gaps in our knowledge of associations between childhood adversities and psychosis. STUDY DESIGN: EU-GEI WP2 comprises incidence and case-control studies of first-episode psychosis conducted in 17 sites in 6 countries. In each site, over 2-year periods, we identified and collected relevant data from individuals aged 18-64 with a first-episode psychosis and with no history of psychosis. Missing data were imputed. We used multi-level logistic regression to test our hypotheses. STUDY RESULTS: In total, 1071 cases and 1497 controls were included. We found variations in the prevalence and the magnitude of associations between any adversity and psychosis by place (eg, odds ratios ranged from 0.4 Cuenca, Spain to 12.1 Madrid, Spain). The weighted percentages reporting adversities in control samples were associated with site incidence rates (eg, 3+ adversities: Spearman's rho 0.56, P .025). We found variations in the magnitude of associations by sex (eg, effect of physical and sexual abuse stronger among women), by age of exposure, and by severity and frequency of adversities (eg, largest odds ratios for adversities involving hostility, threat, and violence). CONCLUSIONS: Variations across populations in prevalence and effects of adversities may contribute to variations in rates of psychosis. Variations in effects by sex and age of onset may point to sex-specific mechanisms and to developmentally sensitive periods. Adversities involving severe threat, hostility, and violence may have the largest effects on risk of psychosis.
Morgan et al. (Fri,) studied this question.