Evidence-practice gap in treatment decisions about defective composite and amalgam restorations among Brazilian dentists

Aim: Better understanding of dentists’ decision-making about defective restorations is needed to close the evidence-practice gap (EPG). This study aimed to quantify the EPG about defective restorations and identify dentist factors associated with this EPG. Methods: 216 dentists from São Paulo State, Brazil, completed a questionnaire about three clinical case scenarios involving defective composite restorations with cementum-dentin margins (case 1) and enamel margins (case 2), and an amalgam (case 3) restoration. Dentists were asked what treatment, if any, they would recommend, including preventive treatment, polishing, re-surfacing, or repairing the restoration, or replacing the entire restoration. Replacing the entire restoration in any of these three scenarios was classified as inconsistent with the evidence, comprising an EPG. Bivariate analyses using Chi-square, ANOVA, or multiple comparison tests were performed (p<.05). Results: for defective composite restorations, 49% and 55% of dentists chose to replace the entire restoration for cases 1 and 2, respectively. Twenty-nine percent of dentists chose to replace the entire amalgam restoration. Dentists were significantly more likely to choose to replace the defective amalgam restoration than the composite restoration with a defect at the cementum-dentin margins or the enamel margins (both at p < .001). Female dentists were more likely to choose a conservative treatment than male dentists for cases 1 (p=.034) and 2 (p=.009). Dentists with a higher percentage of patients interested in individualized caries prevention were also more conservative in case 1 (p=.045). Conclusion: a substantial EPG regarding treatment decisions for defective restorations exists, especially for composite restorations. This study adds to the international evidence that an EPG exists in this clinical area and that global strategies need to be developed to close the gap.