The authors reviewed the diagnostic methodology in 14 studies that examined the prevalence of coexisting psychiatric disorders in substance abusers. There was widespread variation among the studies in the timing of patient interviews, the nature of the interviews themselves, and abstinence criteria required before another psychiatric disorder could be diagnosed. These differences were reflected in some of the study results. The authors describe how variations in methodology can affect the diagnoses patients receive. They also suggest more specific abstinence criteria, based on the substances of abuse and the specific disorders being diagnosed.