Rheumatology trials: selection of adequate response variables and their evaluation

Response variables in rheumatology trials should describe not only the disease manifestations, but should also measure the changes of impairment, disability and handicap. One should use the following criteria, at least: analogue pain scale, score of patient's overall well-being, functional score (A.R.A., Ritchie or Keitel) and Arthritis Impact Measurement Scale (A.I.M.S.). The statistical evaluation of these time-dependent measurements should be done with nonparametric methods using all information over time.