a. Random intercept and, for the continuous outcome, time effects were included for each participant with an unstructured covariance pattern for serial correlation. No heavy drinking defined as the absence of any heavy drinking days in the past 30 days, where a heavy drinking days is defined for men ≥5 drinks per day and for women as ≥4 drinks per day.
b. Bars reflect the associated 95% confidence intervals for the point estimates.