Time-to-first help-seeking and time-to-first treatment were tested in Cox regression models that included all subjects, with dummy variables representing the silent, generation X, and millennial generations with the baby boomer generation again as referent. Four models were run: 1) years from birth to first help-seeking behavior, 2) years from first AUD onset to first help-seeking behavior, 3) years from birth to first treatment utilization, 4) years from first AUD onset to first treatment utilization. Three planned interactions of sex with generation were tested in each model to assess whether females in a given generation had different rates of help-seeking behaviors or treatment utilization than males, relative to their baby boomer counterparts; interactions that were significant at p ≤ .017 after Bonferroni correction were included in the final model. When interactions were significant, separate terms for males and females within the generation were created to facilitate interpretation. These reparamaterized models had the same number of parameters as the models with the interaction terms. Once final models with significant interactions were determined, tests of proportional hazards, i.e. the assumption that the risk
