Endpoint measures in rheumatoid arthritis clinical trials: group summary and individual patient analysis.

Author(s): van der Heide A, Jacobs JW, Schenk Y, Bijlsma JW

Affiliation(s): Cooperating Department of Rheumatology, University Hospital Utrecht, The Netherlands.

Publication date & source: 1994-12, J Rheumatol., 21(12):2195-9.

Publication type: Clinical Trial; Multicenter Study; Randomized Controlled Trial

OBJECTIVE. To evaluate the responsiveness of measures of disability, discomfort, and disease process in rheumatoid arthritis (RA) clinical trials, when used as group summary variables and as variables of individual patient improvement. METHODS. Disease outcome and process measures were assessed in 97 patients with RA of recent onset, who were participating in a prospective trial comparing the effectiveness of several drug treatment strategies. Measurements were done after 3 and 6 months of treatment. Group summary analysis was performed with tests of statistical significance of changes, and by calculating effect sizes (i.e., mean change in an endpoint divided by its standard deviation). Individual patient improvement was defined as improvement of > or = 33% compared to baseline, according to recommendations of the recently held Conference on Outcome Measures in Rheumatoid Arthritis Clinical Trials. RESULTS. Almost all mean group changes in endpoints were statistically significant (p < 0.001). Effect sizes and figures on individual patient improvement provided additional information: physical discomfort measures were rapidly responding measures that did not further improve after 3 months; disease process measures, joint count, erythrocyte sedimentation rate and C-reactive protein also responded quickly and kept improving up to 6 months; the disability measures were relatively unchanged at 3 months, and only the self-report questionnaire score showed considerable improvement at 6 months. CONCLUSION. Effect sizes and data on patients who showed clinical improvement in disease process or outcome measures offset the strongly significant p values of statistical tests for almost all endpoint measures. Although discomfort measures rapidly responded to therapy, disability and disease process measures may not reach optimal improvement within 6 months.