Psychometric testing of the SF-36 Taiwan version on older stroke patients

Aim and objective. To assess the psychometric properties of the SF-36 Taiwan version on a sample of older stroke patients in Taiwan. Background. The Medical Outcomes Study Short Form-36 (SF-36) has proven to be a valid and reliable instrument in evaluating outcomes among stroke patients in western countries; however, less is known regarding its value for Asian stroke patients. Design. A descriptive correlational study design was used to explore the reliability and validity of the SF-36 Taiwan version. Method. Older stroke patients (n = 87) from a medical centre in northern Taiwan were interviewed at the end of the first and sixth months after hospital discharge. Results. Items on each subscale of the SF-36 Taiwan version had similar standard deviations, supporting the scaling assumption of equal-item variance in measuring the same concept. Correlations between items and their subscale were generally ≥0·70 with a few being 0·50 or 0·60. Cronbach's alpha coefficients were >0·70 for almost all subscales, supporting internal consistency. At both the first and sixth months after discharge, participants had lower scores, especially on the physical- and social function-related scales, than the norm for older people. At both times, patients with physical dependence had lower scores across subscales than those without physical dependence, supporting construct validity. Conclusions. The SF-36 Taiwan version demonstrated good reliability and validity when applied to stroke patients at either a transitional or stable time point after hospital discharge. However, the SF-36 Taiwan version has a potential to underestimate changes in certain domains due to floor or ceiling effects. Relevance to clinical practice. Health care providers who deal with Taiwanese/Chinese stroke patients can use the SF-36 Taiwan version to assess health outcomes at either a transitional or a stable time point after hospital discharge.