IADR Abstract Archives
Comparison of Visual and Digital Assessment: Evaluating Common Wax-Up Errors
Objectives: To evaluate the correlation of visual and digital assessment in evaluating common wax-up errors.
Methods: A ½ fractional factorial design with replication was used for six treatment conditions (present, absent) upon 64 preparations. These treatment conditions corresponded to wax-up errors involving: Embrasure, Line Angle, Buccal Contour, Buccal cusp, Mesial Marginal Ridge, and Triangular Ridge. The prepared teeth were scanned and compared to a master model using Compare software at a tolerance of 0.25mm (Digital Score). These same preparations were assessed by expert faculty, who graded all 64 pieces on two separate occasions. Expert grading was done using a rubric with 23 criteria and three subscales. Each subscale was handled by a dedicated faculty member for both instances of grading. When errors were found, faculty also rated the severity of the error as mild (1), moderate (2), or severe (3). A traditional expert grade was assigned using the total number of error free criteria divided by 23 (Visual Score). A modified grade was assigned using the severity of the error to weight the point deduction (Weighted Visual Score). The hypothesis was that there would be no agreement among the three scoring methods. Inter/intra-rater agreement was examined using the intraclass correlation coefficient, spearman rank correlation coefficient, and difference analysis.
Results: Inter-rater reliability between weighted visual and digital scoring (ICC=0.58 & 0.54) was comparable to the agreement between weighted visual and visual scoring (ICC=0.59 & 0.63). The lowest agreement was between visual and digital scoring (ICC=0.11). Intra-rater reliability was excellent for visual and weighted visual scoring (ICC > 0.8). Spearman correlation provided evidence of a strong, positive correlation for both visual (Rs=0.75, P<0.0001) and weighted visual scoring (Rs=0.76, P<0.0001) with digital assessment.
Conclusions: Visual and digital assessments were valid evaluation methods. Weighted evaluation showed comparable correlation and increased agreement with digital assessment.
Methods: A ½ fractional factorial design with replication was used for six treatment conditions (present, absent) upon 64 preparations. These treatment conditions corresponded to wax-up errors involving: Embrasure, Line Angle, Buccal Contour, Buccal cusp, Mesial Marginal Ridge, and Triangular Ridge. The prepared teeth were scanned and compared to a master model using Compare software at a tolerance of 0.25mm (Digital Score). These same preparations were assessed by expert faculty, who graded all 64 pieces on two separate occasions. Expert grading was done using a rubric with 23 criteria and three subscales. Each subscale was handled by a dedicated faculty member for both instances of grading. When errors were found, faculty also rated the severity of the error as mild (1), moderate (2), or severe (3). A traditional expert grade was assigned using the total number of error free criteria divided by 23 (Visual Score). A modified grade was assigned using the severity of the error to weight the point deduction (Weighted Visual Score). The hypothesis was that there would be no agreement among the three scoring methods. Inter/intra-rater agreement was examined using the intraclass correlation coefficient, spearman rank correlation coefficient, and difference analysis.
Results: Inter-rater reliability between weighted visual and digital scoring (ICC=0.58 & 0.54) was comparable to the agreement between weighted visual and visual scoring (ICC=0.59 & 0.63). The lowest agreement was between visual and digital scoring (ICC=0.11). Intra-rater reliability was excellent for visual and weighted visual scoring (ICC > 0.8). Spearman correlation provided evidence of a strong, positive correlation for both visual (Rs=0.75, P<0.0001) and weighted visual scoring (Rs=0.76, P<0.0001) with digital assessment.
Conclusions: Visual and digital assessments were valid evaluation methods. Weighted evaluation showed comparable correlation and increased agreement with digital assessment.