Objective: To compare a disease specific measure of functional status, the Health Assessment Questionnaire (HAQ) with a generic measure of functional status, the Groningen Activity Restriction Scale (GARS) in a cross cultural and longitudinal setting. Besides the comparison of psychometric properties, sex differences were also studied.
Methods: In a sample of 634 patients with rheumatoid arthritis (RA) (290 from The Netherlands, 116 from France, 228 from Norway), the psychometric properties (Cronbach's alpha and rho) of the GARS and the HAQ were compared among countries. The sensitivity for change between T1 and T2 was tested by means of the standardized response mean for a one year followup; furthermore, the relative efficiency of both instruments was computed. A t test was used to trace sex differences.
Results: The psychometric properties of both the HAQ and the GARS were very good. Both form unidimensional scales of hierarchically ordered items; alpha and rho are > 0.91 for both instruments. For patients who improved between T1 and T2, measured by the Ritchie Articular Index, the standardized response mean of the HAQ and the GARS were moderate, whereas for patients who worsened, the standardized response means were small. The results based on the relative efficiency were ambiguous. Furthermore, the mean score for women on the HAQ was 0.25 higher in contrast to scores for men. The GARS did not yield different scores for women and men.
Conclusion: The internal consistency of the GARS and the HAQ are comparable. The HAQ has yielded different results for women and men. Therefore, when the HAQ is used, sex differences should be taken into account. With respect to sensitivity to change, the GARS and the HAQ obtained much better results for patients who improved in contrast to patients who worsened. The GARS was found to be more sensitive to change for patients who improved.