Using Factor Mixture Models to Evaluate the Type A/B Classification of Alcohol Use Disorders in a Heterogeneous Treatment Sample.

BACKGROUND The type A/B classification model for alcohol use disorders (AUDs) has received considerable empirical support. However, few studies examine the underlying latent structure of this subtyping model, which has been challenged as a dichotomization of a single drinking severity dimension. Type B, relative to type A, alcoholics represent those with early age of onset, greater familial risk, and worse outcomes from alcohol use. METHODS We examined the latent structure of the type A/B model using categorical, dimensional, and factor mixture models in a mixed-gender community treatment-seeking sample of adults with an AUD. RESULTS Factor analytic models identified 2 factors (drinking severity/externalizing psychopathology and internalizing psychopathology) underlying the type A/B indicators. A factor mixture model with 2 dimensions and 3 classes emerged as the best overall fitting model. The classes reflected a type A class and 2 type B classes (B1 and B2) that differed on the respective level of drinking severity/externalizing pathology and internalizing pathology. Type B1 had a greater prevalence of women and more internalizing pathology and B2 had a greater prevalence of men and more drinking severity/externalizing pathology. The 2-factor, 3-class model also exhibited predictive validity by explaining significant variance in 12-month drinking and drug use outcomes. CONCLUSIONS The model identified in this study may provide a basis for examining different sources of heterogeneity in the course and outcome of AUDs.