The Americleft Project: Plaster Dental Casts Versus Digital Images for GOSLON Yardstick Ratings When Used in Intercenter Comparisons.

The purpose of this investigation was to determine reliability and validity of GOSLON Yardstick ratings using plaster casts versus photo galleries of digital images in actual intercenter comparisons. The dental arch relationships of 112 patients with complete unilateral cleft lip and palate from 3 North American cleft/craniofacial centers were rated in 2 separate studies. In the first, plaster casts were used. For a later intercenter comparison, the same dental casts were scanned, digital bases added, and two-dimensional photographic galleries (6 views) were created for each set of casts. Three raters experienced with the GOSLON Yardstick carried out 2 separate ratings of the plaster casts in the first study, then of the photographic gallery of scanned digital images of the same casts in the second study. Inter- and intrarater reliabilities were calculated using the Weighted Kappa statistic. Average scores for each patient were calculated and compared between methods with correlation statistics and a Bland-Altman plot. Kruskal-Wallis test was used to compare results between centers using both media. Reliability using both methods was very good and comparable between methods. Mean weighted Kappas were: inter-rater = 0.815 (plaster) versus 0.891 (photo); and intrarater = 0.866 (plaster) versus 0.891 (photo). There was a highly significant correlation (r = 0.920). Mean difference between centers was 0.033 of a GOSLON category. The level of significance of the differences found between centers with both methods was identical, confirming the interchangeability of both media presentations.