RESUMEN
To perform a validation of the Geriatric Oral Health Assessment Index (GOHAI) for complete denture wearers and present a proposal for estimation of perceived oral health. This is a cross-sectional study with non-probabilistic sampling. A total of 211 subjects with a mean age of 62·5 (s.d. = 11·4) years participated, being 169 female. The GOHAI was applied in a personal interview. The construct/convergent/discriminant validity was tested using structural equation modelling. Confirmatory factor analysis was used to verify the fit of three proposals of the GOHAI: three-factor, one-factor and second-order hierarchical models. The stability of the models was evaluated in independent samples. The three-factor model presented an inadequate fit, and items 3, 4 and 9 were removed. The new structure presented an acceptable fit and strong invariance in independent samples. The convergent, discriminant validity and internal consistency were below adequate. The one-factor model presented an adequate fit to the sample. Convergent validity was compromised. A strong invariance of the one-factor model was observed. To calculate the overall scores of the GOHAI factors (three-factor model) or of the oral health perception (one-factor model), a matrix of regression weights for each item in the model was presented as a suggestion. We found an adequate fit of the both structures of the GOHAI for denture wearers, but the three-factor structure was more parsimonious. We suggested considering the weights of the regression model to calculate the overall score of perceived oral health or of its factors in different samples.
