A recent 6-month follow-up of methadone maintenance (MM) suggested that addicts supported by public assistance before seeking MM have a globally poor prognosis compared to addicts supported by employment or illegal activities. We therefore examined 2.5-year outcome among addicts prognostically stratified by three major sources of pretreatment income: employment (n = 48), welfare (n = 46), and criminal activities (n = 57). Significant improvements in substance abuse, family, legal, and psychological problems were reported by all three groups. Furthermore, the welfare group showed the most improvement in medical status and on the Social Adjustment, Maudsley neuroticism, and Beck Depression scales. Thus, the welfare patients did not appear to have a globally poorer prognosis, although they showed the least improvement in employment and legal problems. To examine whether welfare patients on MM had no more than "expected" improvement in legal and employment problems, patients treated with MM (n = 83) were compared to those getting only detoxification (DT) (n = 40). Welfare patients showed no more improvement in these two areas from MM treatment than from DT alone, while at follow-up the "employed" and "criminal" groups had substantially less illegal income and unemployment with MM than with DT alone. Thus, welfare patients may do relatively poorly in some areas when treated on MM.