Discriminative ability of the short-form 36 health survey: a tale of two versions

Abstract

Background and objectives: Previous studies suggested that the Chinese version of the Short-Form 36 Health Survey (SF-36) had smaller variability in the physical functioning (PF) and physical component summary (PCS) scores than the English SF-36. This may translate into differences in discriminative ability and sample size requirement.

Materials and methods: Data were drawn from a community-based survey and a randomised crossover study of Singaporeans bilingual in the Chinese and English languages. The abilities of the two PF and PCS versions in discriminating subjects who reported chronic illness and acute disease symptoms versus those who did not were compared.

Results: In all four comparisons (i.e. two health criteria in two studies) the Chinese version of PF showed a larger effect size than the English version. In three out of four comparisons the Chinese version of PCS showed a larger effect size than the English version.

Conclusions: The Chinese version appeared more efficient in detecting a statistically significant difference between groups. Other factors being the same, the Chinese SF-36 may require a smaller sample size than the English SF-36 for the studies of physical aspects of health-related quality of life.