Effect sizes for interpreting changes in health status.

Health status measures are being used with increasing frequency in clinical research. Up to now the emphasis has been on the reliability and validity of these measures. Less attention has been given to the sensitivity of these measures for detecting clinical change. As health status measures are applied more frequently in the clinical setting, we need a useful way to estimate and communicate whether particular changes in health status are clinically relevant. This report considers effect sizes as a useful way to interpret changes in health status. Effect sizes are defined as the mean change found in a variable divided by the standard deviation of that variable. Effect sizes are used to translate "the before and after changes" in a "one group" situation into a standard unit of measurement that will provide a clearer understanding of health status results. The utility of effect sizes is demonstrated from four different perspectives using three health status data sets derived from arthritis populations administered the Arthritis Impact Measurement Scales (AIMS). The first perspective shows how general and instrument-specific benchmarks can be developed and how they can be used to translate the meaning of clinical change. The second perspective shows how effect sizes can be used to compare traditional clinical measures with health status measures in a standard clinical drug trial. The third application demonstrates the use of effect sizes when comparing two drugs tested in separate drug trials and shows how they can facilitate this type of comparison. Finally, our health status results show how effect sizes can supplement standard statistical testing to give a more complete and clinically relevant picture of health status change. We conclude that effect sizes are an important tool that will facilitate the use and interpretation of health status measures in clinical research in arthritis and other chronic diseases.