We have proposed a set of changes in the design, analysis, and reporting of RA clinical trials which emphasizes the standardization of trials. We have suggested limiting the number of outcome measures used to those most sensitive to change and using uniform measurements of the same outcome measures in all clinical trials. We have also suggested that investigators focus on enumerating which patients improve in a trial, rather than on the mean level of improvement in treated patients versus the control patients. Finally, we have suggested the reporting of confidence intervals rather than simply stating P values. With the inauguration of these changes and more attention to other issues relevant to the quality of clinical trials, we hope that drug treatments of rheumatoid arthritis can be more easily compared with respect to effectiveness.