A psychometric network analysis of OHIP-14 across Australian and Brazilian populations

Abstract

Objectives: The aim of this study was to assess the structural validity of the Oral Health Impact Profile–14 (OHIP-14) instrument in Indigenous and non-Indigenous populations from Australia and Brazil using a network analysis approach. Methods: Cross-sectional data collected using OHIP-14 were obtained for Indigenous and non-Indigenous populations from Australia and Brazil. Networks were estimated using the Gaussian graphical model. Items of the OHIP-14 instrument are represented as nodes and the partial correlations between items as edges. Dimensionality was assessed using exploratory graph analysis. Structural consistency and item stability were computed using a bootstrap sampling method. Standardized node strength across each dimension was also calculated. Results: Four dimensions were identified across all samples, although the item arrangement of most dimensions presented variation. Similarities with the theoretical domains of the instrument were found. Items from the conceptually derived OHIP-14 domains formed separated clusters or blended with other items in a single dimension. Most dimensions across all samples showed an acceptable structural consistency. Item stability revealed some discrepancies among items of dimensions of both Indigenous networks. Conclusion: The psychometric network perspective adopted in this study provides validation of the OHIP-14 structure in Indigenous and non-Indigenous populations. The structural consistency and item stability analyses showed that both Indigenous networks present a higher number of cross-domain items and less defined boundaries between dimensions. These findings indicate that OHIP-14 does not measure attributes in the same extent among different cultures. Knowledge Transfer Statement This study demonstrates a new analytical framework from which to conceptualize and interpret the construct oral health–related quality of life using the Oral Health Impact Profile (OHIP-14). Network graphs facilitate knowledge translation of findings to professionals with no expertise in psychometric methods. OHIP-14 is a valuable tool to oral health research and clinical practice. Differences in conceptions of health may influence the extent that the instrument measures oral health impacts. Consequently, dimension scores do not always provide appropriate measures and should be avoided in research reports and assessments of treatment outcomes.