Objectives: To present the properties of an efficiency measure that may be used to make statistical comparisons of sensitivities of trial outcome measures. To use this measure to examine the relative sensitivities of several common rheumatoid arthritis (RA) trial outcome measures in 2 different trial settings.
Methods: Efficiency is expressed as the mean change divided by the standard deviation of change. Variability and correlations of efficiencies for typical RA trial outcomes are described.
Results: From among a variety of joint assessments and other clinical measures, pain and global measures, and health status and laboratory measures, the joint tenderness and pain measures were the most sensitive, both in a trial of nonsteroidal antiinflammatory drugs (NSAID) and in a set of second-line drug trials. Measures with rather different sensitivities in the 2 types of trials were erythrocyte sedimentation rate (low sensitivity in the NSAID trial) and physician and patient global measures (higher sensitivity in the NSAID trial).
Conclusion: The use of statistical information on efficiency estimates allows for a variety of comparisons of measures and may inform the selection of outcome measures for clinical trials.