Novel methods of balancing covariates for the assessment of dental erosion: A contribution to validation of a synthetic scoring system for erosive wear

Abstract Objectives The purpose of the study was to balance several potential erosive covariates, using traditional and novel epidemiological approaches, in order to assess the relative risks of dental erosion more precisely. Methods Traditional (univariate and logistic regression analysis) and novel techniques (propensity scores and Inverse Probability Weighting—IPW) were applied for evaluating the effect of twenty covariates on dental erosion among 502 adolescents. Results Different approaches gave different estimates of the relative risks of dental erosion. The increased consumption of carbonated soft drinks had the major erosive effect, when traditional analyses were used (unadjusted: OR = 3.475 and CI: 1.499–8.052, logistic regression: OR = 3.219 and CI: 1.373–7.547). On the other hand, IPW method indicated that the consumption of erosion drinks immediately after intense physical exercise had the highest odds ratio (OR = 1.363 and CI: 0.963–1.929), followed by the increased consumption of citrus fruit juice (OR = 1.326 and CI: 1.004–1.752). This method also demonstrated a marked improvement in balance, with the 95% CIs for each OR being considerably narrower than those reported in the initial analysis. Conclusions Standardization of the potential aetiological criteria of erosive wear is a considerably difficult process. Nevertheless, novel methods revealed that the increased consumption of carbonated soft drinks and citrus fruit juices could be included as aetiologic factors in a synthetic scoring system for erosion. Parameters which are related to salivary protective mechanisms (e.g. consumption of erosion drinks immediately after intense physical exercise) could also be a part of such an index. Further research is required in order to achieve the maximum validation of the potential erosive risk factors.