Background: Quality of life is gaining relevance as a criterion of success in therapeutic outcome studies. In order to record quality of life, disease-specific instruments are being used as well as generic instruments. Generic instruments offer the possibility to compare outcome among different indications; however for this it is necessary to prove the differential factorial validity of the instrument's structure.
Aim of the study: Using structural equation modeling the SF-36 was administered to a sample of patients with rheumatic diseases. The sensitivity to change of the measure was calculated after six months. Furthermore, age- and sex-specific scale values were calculated and compared with norm data.
Methods: SF-36 data of 436 patients as well as sex and age from two scientific rehabilitation research projects were pooled for this secondary analysis.
Results: The structure of the questionnaire proves to be acceptable and comparable with international results. The confirmatory analysis supports the best fit for a model with crossloadings assuming correlated main dimensions. Sensitivity to change is low in general. The best effect size is found for the subscale pain. The comparison with norm data shows that patients are impaired particularly within their physical health (regarding the SF subscales) both at the beginning of the treatment and after six months. In all age groups men have better conditions than women.
Conclusions: The generic SF-36 show sufficient factorial validity. From this point of view there is no objection to using the SF-36 in patients with rheumatic diseases. Indeed the use of the SF summary scales as a representation of (uncorrelated) physical and mental health must be questioned critically.