Development and psychometric properties of the Portuguese version of the Orofacial Esthetic Scale: OES-Pt

Abstract

Background: The Orofacial Esthetic Scale (OES) is an instrument to assess an individual's perception of their Orofacial Appearance (OA). However, its translation and evaluation of psychometric properties is necessary for its use in Brazilian individuals. Objectives: To develop the Portuguese version of OES (OES-Pt), estimate its psychometric properties (validity, measurement invariance and reliability) when applied to Brazilian individuals aged 18-40 years, and estimate the relationship between sociodemographic characteristics and OA. Methods: This was a cross-sectional study using a convenience sample. The sample consisted of 1,072 Brazilian individuals (70.1% female, 25.1% dental patients; mean +/- SD age: 25.7 +/- 5.7 years). After cross-cultural adaptation of OES-Pt, factorial validity was evaluated by confirmatory factor analysis. Convergent validity (average variance extracted (AVE)) and reliability (Cronbach's alpha coefficient (a) and Composite Reliability (CR)) were also estimated. Concurrent validity was assessed (Pearson's correlational analysis (r) between OES-Pt total score and item eight of the OES which refers to global assessment of OA). Measurement invariance of the factorial model (multigroup analysis using ACFI) was evaluated for independent samples (sample randomly split into two: Test Sample and Validation Sample and according to sex: male and female, age range: 18-30 and 31-40 years, and whether the individual is undergoing dental treatment or not). A Structural Equation Model estimated the relationship between sociodemographic characteristics and OA. Results: OES-Pt presented adequate fit to the sample. Convergent validity (AVE >= 0.56) and reliability (a and CR >= 0.89) were adequate. Concurrent validity was adequate (r = 0.88; p-value < 0.001). OES-Pt presented strict invariance for independent samples. Age, sex, and socioeconomic status (SES) were related to OA, indicated by standardized beta coefficients (standardized beta) of 0.036 (standard error: 0.007), 0.001 (0.094) and 0.196 (0.061), respectively on OA. These three relationships were either weak or not statistically significant. Conclusions: When measuring OA in Brazilian individuals, the OES-Pt was valid, reliable and invariant for independent samples. Age, sex and SES were weak or not statistically significantly related to OA.