Figure 1d shows the required sample size by interaction effect size for comorbid disorders with a prevalence of 10%, 20%, and 30%, while the power is fixed at 80%. In order to detect an interaction effect size of 2.0 with a comorbid disorder of 30% prevalence, we need a sample of 4,987 subjects, we need 6,376 subjects to detect the same interaction with a comorbid disorder of 20% prevalence such as major depressive disorder, and we need 15,663 subjects to detect the same interaction with a comorbid disorder of 10% prevalence. The required sample size increases as detectable effects size and prevalence decrease.
