Factorial structure and measurement invariance of the Chinese version of the Oral Health Impact Profile-14 among clinical populations and non-clinical populations: an evidence for public oral investigations

Abstract

Objective: Oral health-related quality of life (OHRQoL) is a multidimensional concept that is commonly used to examine the impact of oral health status on quality of life. The purpose of this study was to examine the optimal factor model of the Chinese version of the Oral Health Impact Profile (OHIP-14) questionnaire in clinical populations, measurement invariance across clinical status and gender cohorts. This would ensure equal validity of the Chinese version of OHIP-14 in different populations and further support public oral investigations.

Methods: The Chinese version of OHIP-14 was used to investigate 490 dental patients and 919 college students. Confirmatory factor analysis (CFA), item analysis and reliability, measurement invariance, and the t-test were used for data analyses.

Results: We found that the 7-factor structure had the best-fit index in the sample (CFI = 0.970, TLI = 0.952; SRMR = 0.029, RMSEA = 0.052(0.040,0.063)). The reliability of the scales was satisfactory (Cronbach's α = 0.942). The error variance invariance fitted the data adequately in measurement invariance, indicating that measurement invariance is acceptable both across the clinical and non-clinical populations (∆CFI=-0.017, ∆RMSEA = 0.010) and across genders in the clinical population (∆CFI = 0.000, ∆RMSEA=-0.003). T-test for scores showed that the clinical populations scored significantly higher than the non-clinical populations, as did the overall score (t = 7.046, p < 0.001, d = 0.396), in terms of functional limitation (t = 2.178, p = 0.030, d = 0.125), physical pain (t = 7.880, p < 0.001,d = 0.436), psychological discomfort (t = 8.993, p < 0.001, d = 0.514), physical disability (t = 6.343, p < 0.001, d = 0.358), psychological disability (t = 5.592, p < 0.001, d = 0.315), social disability (t = 5.301, p < 0.001,d = 0.304), social handicap (t = 4.452, p < 0.001, d = 0.253), and that in the non-clinical populations, females scored significantly higher than males, as did in terms of physical pain (t = 3.055, p = 0.002, d = 0.280), psychological discomfort (t = 2.478, p = 0.014, d = 0.222), and psychological disability (t = 2.067, p = 0.039, d = 0.188).

Conclusion: This study found that the Chinese version of OHIP-14 has measurement invariance between the clinical and non-clinical populations and across genders in the clinical populations, and can be widely used in OHRQoL assessment for public oral investigations.

Keywords: Confirmatory factor analysis; Measurement invariance; Oral Health Impact Profile-14; Oral health-related quality of life; Oral investigations.