The proposed disease controlling antirheumatic therapy (DC-ART) definition requires that the therapy change the course of rheumatoid arthritis (RA) for at least 1 year, evidenced by (1) sustained improvement in physical function, (2) decreased inflammatory synovitis, and (3) slowing or prevention of structural joint damage. Selected studies are reviewed. All studies were at least 1 year in duration, but most did not include all 3 of the DC-ART requirements. In these studies, patients treated with placebo generally had no improvement in inflammatory synovitis and progressive structural joint damage, judged by serial joint radiographs. A minority of studies significantly favored one or another of the available agents (gold injections, D-penicillamine, auranofin, antimalarials, azathioprine, sulfasalazine, methotrexate), but the evidence for any one agent is not convincing. For future DC-ART clinical trials patients with early RA should be studied. A hybrid study design may be useful, combining an initial double blind randomized controlled clinical trial with continuing longterm observation of all withdrawals using specified clinical, radiographic, and self report assessments at regular intervals, and an intent-to-treat analysis comparing longterm response rates of the original control and experimental therapy groups. Responsive subgroups should be sought, their characteristics identified, and their responsiveness confirmed in additional trials limited to the identified subgroup.