alcohol use between age 18 and 30, as previously described (Dick et al., 2014; Li et al., 2017). Age was centered at 21, where alcohol consumption started to stabilize (Dick et al., 2014; Li et al., 2017). In addition to age, sex and self-reported racial identity were included as covariates to control for the baseline differences in alcohol consumption and AD by sex and race. For race, two dummy variables were created from self-reported racial identity to represent African Americans (AA) or other races, respectively, with European Americans (EA) as the reference group. Native Americans, Asians, Pacific Islanders, and unknown racial identification were combined into other races. First-degree autoregressive covariance structure was used in all models to incorporate repeated assessments within individuals. Parameters were estimated by the GEE estimator in SAS software version 9.4 (SAS Institute, Cary, NC) using the GENMOD procedure with the REPEATED option.
