Evaluating the Factor Structure of the Preliminary Version of EuroQol Health and Well-Being Instrument in China: A Replication of the Confirmatory Factor Analysis

Objectives: 

During the development of EuroQol Health and Well-Being instrument (EQ-HWB), the Chinese online sample showed a weaker model fit for the confirmatory factor model based on the UK sample, even after adjustments. This may be because of cultural differences affecting item interpretation or demographic disparities, particularly in age and education. To address this, the study aimed to replicate the Chinese-adjusted model, a version of the UK model modified based on the Chinese online sample, using a more diverse Chinese face-to-face sample to assess potential improvements in model fit. 

Methods: 

We conducted face-to-face interviews to recruit respondents with diverse age and education backgrounds. The Chinese-adjusted model, a bifactor confirmatory model with 9 factors and 2 measurement factors, was replicated in 2 stages. First, the Chinese online sample validated the model. Then, the face-to-face sample assessed improvements in model fit. The confirmatory factor analysis model fit criteria were as follows: comparative fit index and Tucker-Lewis index greater than 0.95 and root mean square error of approximation less than 0.06. 

Results: 

We recruited 553 respondents, including 96 with lung cancer, 125 with diabetes, 102 with depression, 113 with schizophrenia, 16 with other diseases, and 101 healthy individuals. In the face-to-face sample, the goodness-of-fit indices for the Chinese-adjusted model were a comparative fit index of 0.943, Tucker-Lewis index of 0.939, and root mean square error of approximation of 0.052 (90% CI 0.050-0.054), indicating a significantly better fit than in the original study. 

Conclusions: 

The misfit of the UK EQ-HWB model to Chinese data in a previous study had seemed to be because of limited variance. Using a more diverse data set, our study supported the EQ-HWB factor structure's comparability at an international level.