In efforts to contain costs and efficiently allocate resources, evaluation studies in alcoholism have increasingly assessed the effect of treatment on the use of health services through comparisons of treated and untreated alcoholics. The success of this approach requires that evaluators identify and adjust for differences between these two groups, thereby decreasing the likelihood that health utilization outcomes are attributed to the effects of treatment when in fact they may be related more to unidentified group differences. Using a sample of 63,873 hospitalized alcoholics, this study focused on one critical group difference, the severity of alcohol-related medical complications. Comparisons between treated alcoholics who either completed alcoholism treatment or detoxification only and untreated alcoholics with either primary medical/surgical or psychiatric diagnoses demonstrated the following: (a) untreated alcoholics with medical/surgical diagnoses were more likely to have severe alcohol-related medical complications than the other groups; (b) a positive linear relationship between health services utilization in the previous year and the severity of medical complications existed for all groups, except untreated alcoholics with psychiatric diagnoses; and (c) untreated alcoholics with psychiatric diagnoses with the most severe complications used fewer health services than any of the other three groups. Differences between treated and untreated alcoholics in both severity of medical complications and previous health utilization patterns demonstrate the need to identify and adjust for these factors in evaluation studies that examine the outcomes of alcoholism treatment.