Prediction of 3-year outcome of treated alcoholics by an empirically derived multivariate typology.

Abstract

This study examined whether an empirically derived multivariate typology can predict outcome in treated alcoholics. Two hundred fifty-nine hospitalized male alcoholics were divided into two subtypes by means of cluster analysis of clinical factors. Type A showed the features of sporadic late-onset alcoholism, whereas type B manifested the characteristics of familial early-onset alcoholism. The prospective 3-year outcomes in the two subtypes were compared. The follow-up rate was 83%. The mortality and abstinence rates for type A patients (N = 98) were 15.3% and 32.7%, respectively, and for type B patients (N = 116), 23.3% and 18.1%, respectively. The age-corrected relative mortality risk and relative abstinence rate for type A patients compared with type B patients were 0.59 and 1.60, respectively. Familial early-onset alcoholism has a poor outcome in Japan, and this multivariate typology has high predictive validity.