Background

Tooth wear is a multifactorial process, leading to the loss of dental hard tissues. Therefore, it is important to detect the level of tooth wear at an early stage, so monitoring can be initiated. The Tooth Wear Evaluation System (TWES) enables such a multi-stage diagnosis of tooth wear. The further developed TWES 2.0 contains a complete taxonomy of tooth wear, but its reliability has not yet been validated.

Objectives

Aim of the study was to examine in a randomized controlled trial (RCT) whether diagnoses made based on the TWES 2.0 are reproducible and whether this reproducibility is also achieved with computer-assisted diagnostics.

Methods

44 dental students received extensive training in TWES 2.0 assessment and taxonomy. The students each evaluated at least 10 (of the present 14) anonymized patient cases using gypsum models and high-resolution intraoral photographs according to TWES 2.0. One half initially evaluated on paper; the other half used dedicated software (CMDfact / CMDbrux). After half of the patient cases (5), the evaluation methods were switched (AB/BA crossover design). The diagnoses were then evaluated for agreement with the predefined sample solution.

Results

Evaluation of agreement with the sample solution according to Cohen's kappa indicated a value of 0.46 for manual (traditional) evaluation; and 0.44 for computer-assisted evaluation. Evaluation of agreement between examiners was 0,38 for manual and 0,48 for computer-assisted evaluation (Fleiss’ Kappa).

Conclusion

The results of this study proved that the taxonomy of the TWES 2.0 has acceptable reliability and can thus be used by dentists. Accordingly, the system can be learned quickly even by untrained practitioners. Comparable results are achieved with computer-assisted evaluation.