The randomized controlled clinical trial is the "gold standard" to evaluate therapeutic interventions, but is more effectively applied to studies of the short-term treatment of acute diseases than to the long-term treatment of chronic diseases. Clinical observations often provide more accurate outcome data in rheumatoid arthritis (RA) than clinical trials. Limitations of clinical trials to depict long-term outcomes in RA include a relatively short observation period, patient selection resulting from exclusion criteria, inflexible dosage schedules and concomitant drug therapies, evidence that some markers of inflammatory activity are suboptimal surrogate indicators of long-term articular damage, the fact that statistically significant results are not necessarily clinically important, the influence of the design on the results-despite a control group, ignoring of individual variation in reporting results, the non-standardized interpretation of side effects which introduces bias, distortion of the "placebo effect", and lack of capacity to detect rare side effects. The clinical trial represents only an initial step in the evaluation of a therapy for a chronic disease. Awareness of these limitations should lead to the improved design of clinical trials and clinical studies to improve the long-term outcome for people with RA.