The Reliability and Validity of Short Form-12 Health Survey Version 2 for Chinese Older Adults

Abstract

We assessed the information regarding the psychometric properties of the Short Form-12 Health Survey Version 2 (SF-12v2) among older adults in China. A cross-sectional study was conducted on a stratified representative sample of older adults (≥60 years) residing in community and nursing home settings in 2017-18. Reliability was estimated using the internal consistency method. Validity was assessed using convergent and discriminant validity checks, factor analyses (including both exploratory and confirmatory factor analyses [EFA and CFA]), and "known groups" construct validity. The final sample comprised 1000 older adults (451 community-dwelling and 549 institutional). Cronbach's α was 0.81 for the Physical Component Summary (PCS) and 0.83 for the Mental Component Summary (MCS), showing satisfactory internal consistency for both. Most items were strongly correlated with their represented component (Spearman's correlation coefficient: 0.62-0.87), although the correlation of SF items with PCS was a bit stronger than that with MCS. A two-factor structure (physical and mental health) indicated by EFA jointly accounted for 68.50% of the variance and presented adequate goodness-of-fit indices (GFI=0.98, AGFI=0.92, RMSEA=0.08, 90% Cl RMSEA=0.06 to 0.11, NFI=0.98, and CFI=0.98) in CFA. Known-groups comparison showed that SF-12v2 summary scores did well in differentiating subgroups of older adults by age, marital status, and self-reported health problems (P≤0.05). SF-12v2 is a reliable and valid health-related quality of life instrument for Chinese older adults that works equally well with older adults under institutional care and community-based home care models.