Psychometric Properties of the Oral Health Impact Profile and New Methodological Approach

Abstract

Objective: Evaluate the validity, reliability, and factorial invariance of the complete Portuguese version of the Oral Health Impact Profile (OHIP) and its short version (OHIP-14). Methods: A total of 1,162 adults enrolled in the Faculty of Dentistry of Araraquara/UNESP participated in the study; 73.1% were women; and the mean age was 40.7 + 16.3 yr. We conducted a confirmatory factor analysis, where chi(2)/df, comparative fit index, goodness of fit index, and root mean square error of approximation were used as indices of goodness of fit. The convergent validity was judged from the average variance extracted and the composite reliability, and the internal consistency was estimated by Cronbach standardized alpha. The stability of the models was evaluated by multigroup analysis in independent samples (test and validation) and between users and nonusers of dental prosthesis. Results: We found best-fitting models for the OHIP-14 and among dental prosthesis users. The convergent validity was below adequate values for the factors functional limitation and physical pain for the complete version and for the factors functional limitation and psychological discomfort for the OHIP-14. Values of composite reliability and internal consistency were below adequate in the OHIP-14 for the factors functional limitation and psychological discomfort. We detected strong invariance between test and validation samples of the full version and weak invariance for OHIP-14. The models for users and nonusers of dental prosthesis were not invariant for both versions. Conclusion: The reduced version of the OHIP was parsimonious, reliable, and valid to capture the construct impact of oral health on quality of life, which was more pronounced in prosthesis users.