Replication in the CATS sample25 was carried out using regression analyses that controlled for CUD, opioid use disorder (ascertainment criterion), lifetime use of illicit drugs other than cannabis, age, sex, and the first 9 genetic principal components. To assess differences in findings across the COGA and CATS samples, we used the following formula to calculate a z score for each outcome: Z= |βCOGA−βCATS|/(SECOGA2+SECATS2), where βCOGAis the association between the schizophrenia PRS and the outcome in COGA, and SECOGAis the standard error of the regression coefficient.
