Comparisons between US norm-based two-component and Japanese norm-based three-component SF-36 summary scores in systemic lupus erythematosus patients.

Abstract

We compared the US norm-based two-component vs. Japanese norm-based three-component summary scores of the Medical Outcomes Study Short Form-36 (SF-36) in patients with systemic lupus erythematosus (SLE). One hundred fourteen Japanese SLE patients were studied. SF-36 physical component summary (PCS) and mental component summary (MCS) scores were computed by the US norm-based two-component (US2) and Japanese norm-based three-component (JP3) models, respectively, and compared. Their association with demographics and disease characteristics was also analysed. The US2-PCS scores were significantly higher than the JP3-PCS scores (p < .001); however, the US2-MCS and JP3-MCS scores were not significantly different (p = .16). Bland-Altman analyses demonstrated that the US2-PCS scores were generally higher than the JP3-PCS scores, and their difference was larger in the subjects with lower PCS scores. However, the multiple linear regression analyses for the PCS and MCS scores computed by the different models demonstrated mostly equivalent standardized regression coefficients with the variables. Although the agreement between the US norm-based two-component and Japanese norm-based three-component models of the SF-36 was insufficient, their scores demonstrated similar associations with other variables. The application of the US original version could be acceptable in certain studies depending on the research question.