Modelling variability of quality of life scores: a study of questionnaire version and bilingualism

Modelling variability of quality of life scores can not only improve our understanding of the characteristics of the measurement tools, but also shed light on sample size requirements. Although the English and Chinese versions of the Short Form 36 Health Survey (SF-36) are commonly considered equivalent, they have not been compared in terms of variability. Furthermore, bilingual and monolingual persons may differ in cognition and responses to questionnaires.

In a community-based survey of quality of life in Singapore, a society where both English and Chinese are widely used and bilingualism is prevalent, 2590 respondents answered either version of the SF-36. We studied the impact of questionnaire version and bilingualism on the variability of SF-36 scores by regression modelling, with adjustment for covariates.

The Chinese version had smaller variances in the physical functioning (PF) and the physical component summary scores than the English version. The variance ratios (VRs) were respectively 0.32 and 0.60 (each p < 0.01), controlling for covariates. Bilingualism was not associated with variability in SF-36 scores except PF (VR = 0.78; p < 0.05).

As a result of a smaller variance, using the Chinese version of SF-36 among bilingual Chinese people may require a smaller sample size than using the English version.