OBJECTIVE: To determine the validity of 15 standardized instruments frequently used to measure the outcome of chronic arthritis treatment.

METHODS: Analyses were performed on data collected at a rehabilitation programme (n=216). The outcome measures evaluated were health-related quality of life, global health, pain, physical function and aerobic capacity. The instrument items were linked to the International Classification of Functioning, Disability and Health (ICF) (content validity), construct validity was analysed based on predetermined hypothesis (Spearman's correlations, rs), and responsiveness (after 18 days and 12 months) by the standardized response mean.

RESULTS: Most instruments covered the ICF component body function and/or activity-participation, only a few covered the environmental component. The short Euroqol-5 Dimensions performed as well as the longer health-related quality of life instruments in covering the ICF and in responsiveness. The health-related quality of life instruments did not measure similar constructs as hypothesized, neither did pain measures. The Bath Ankylosing Spondylitis indices covered several components of the ICF often exhibiting a large responsiveness. Aerobic capacity had the largest responsiveness of all measures.

CONCLUSION: Many instruments are not highly correlated, although at face value they appear to measure the same construct, information also applying to content validity and responsiveness. Results from this study can assist in choosing outcome measures in the clinic and in research.