Dentists' variability in restorative decisions, microscopic and radiographic caries depth.

Restorative and dental caries depth decisions were recorded for 5168 un restored approximal tooth surfaces by 17 dentists who worked in the school dental clinics of the North York (Ontario) Public Health Department. Each dentist examined 15 pairs of experimental bitewing radiographs for which true caries depth had previously been determined by microscopy of the sectioned teeth following production of the radiographs. The dentists independently recorded their restorative decisions and radiographic caries depth perceptions. The relationship between the variation in the dentists' restorative decisions and their perceptions of caries depth based on a re-reading of the bitewings on the one hand, and true caries depth on the other was also examined. The percentages of total variability in each dentist's restorative decisions attributable to radiographic and to microscopic caries depth were estimated using regression analyses. Large variations were found among the 17 dentists' distributions of overall restorative and depth decisions. The relationship between microscopic caries depth and the dentists' restorative decisions was, understandably, less strong than that of the dentists radiographic perceptions of caries depth and restorative decisions. Relative to true caries depth, high numbers of false positive and false negative restorative decisions were made. Overall, 50% of the variability in the dentists' restorative decisions was explained by the perceptions of radiographic caries depth; however, among individual dentists, the range was from 29% for one dentist to 69% for another. A much lower percentage of the overall restorative variation was explained by microscopic depth, 18%. Like the finding of the only two previous European studies that quantified the role of radiographs on clinical decisions, this study demonstrated that dentists' perceptions of dental caries depth using bitewing radiographs play a major but variable role in their restorative decisions for approximal tooth surfaces.