The study group consisted of 73 employed alcoholics, who were directed to inpatient treatment. They were evaluated (structured interview, GGT, and MCV) before the treatment period and 8 months after the termination of the treatment. Patients with diseases or medications influencing the laboratory tests were not included in the study group. According to the structured interview 72.6% of the patients were drinking in a controlled manner (less than 40 g/d ethanol) during the 2-month period prior to the follow-up evaluation. According to the laboratory data 52.1% of the patients had been drinking in a controlled manner (i.e. both GGT and MCV values were normal). With the interview and laboratory data combined 39.7% of the patients had been drinking in a controlled manner. Thus the drinking outcome was strongly influenced by the selection of the variables used to measure the outcome. It is recommended that no studies should be carried out about the drinking outcome of alcoholism treatment without the use of laboratory markers as a part of the evaluation.