3360 participants were retained for analysis. The random intercept for family grouping accounted for 16.30% of the variance in divorce. Results from the models of divorce as a function of AUD, PRS, and psychiatric comorbidities are summarized in Table 4. AUD, CAN, TOB, and MDD were associated with higher likelihood of divorce. The alcohol problems PRS was not associated with divorce. Relative to baby boomers, odds of divorce were 54% lower for those in the silent generation, and 70% lower in millennials. Odds of divorce were 42% higher among females. Odds of divorce were 2% higher per year of age (at assessment). In the interactive effects model, no multiplicative interaction terms were associated with the likelihood of divorce (Supplement, Table S6).
