Babor T F, Dolinsky Z S, Meyer R E, Hesselbrock M, Hofmann M, Tennen H
Department of Psychiatry, University of Connecticut School of Medicine, Farmington 06030-1410.
Br J Addict. 1992 Oct;87(10):1415-31. doi: 10.1111/j.1360-0443.1992.tb01921.x.
This study evaluated the discriminative power and predictive validity of five common typological schemes used to classify alcoholics for theoretical or clinical purposes. A heterogenous sample of 321 alcoholics was classified according to primary vs secondary alcoholism, parental alcoholism, Fellinek's gamma-delta distinction, gender, and subtypes derived from MMPI profiles. A prospective longitudinal cohort design was employed to compare the relative ability of these typologies to differentiate alcoholics according to natural history, alcohol-related consequences, response to treatment, and post-treatment adjustment. The findings indicate that while 'one-dimensional' typologies discriminate subgroups in terms of etiological variables, presenting symptoms, and drinking patterns, none of the classification systems emerges as a strong predictor of outcome status. In addition, these typologies do not discriminate well with respect to the alcoholic's drinking patterns and presenting symptoms, except in areas closely connected with the model (e.g. alcoholics with antisocial personality (ASP) indicate more social problems related to alcohol use than do primary alcoholics). Because there is a great deal of overlap among certain subtypes identified within different typological models (e.g. alcoholics with primary ASP tended to have a positive family history and a gamma pattern of impaired control), we hypothesize that empirical clustering techniques that search for naturally occurring commonalities among alcoholics may be a better way to identify homogeneous subtypes.
