Several prior studies, mostly based on small samples, have examined the patient characteristics that predict low risk drinking (i.e., controlled drinking)[11–15], and findings have been mixed. For example, lower severity of alcohol dependence has been identified as a predictor of low risk drinking in many studies[15–17], but not others[12, 14, 18]. Similar inconsistencies have been found for the role of psychiatric problems and mood disturbances[19–21], as well as baseline drinking patterns[14, 22], social network drinking [23, 24], and various demographic factors[11, 23, 25]. Importantly, these prior studies included small samples from only one treatment site, relied on static definitions of low risk drinking (e.g., never exceeding a drinking threshold), and have mostly focused, with only two exceptions[12, 14], on long term outcomes after varying lengths of treatment[13 15–16, 19]. With respect to the last two points, clinicians are not typically interested in patients never exceeding a certain drinking threshold and are often far more interested in the overall patterns of drinking during treatment. To address these limitations, the current study examined which patient characteristics predicted low risk drinking patterns during
