The validity and reliability of an automated method of scoring dental arch relationships in unilateral cleft lip and palate using the modified Huddart-Bodenham scoring system

To evaluate an automated software tool for the assessment of dental arch relationships using the modified Huddart and Bodenham (MHB) index. Cohort of 43 models of subjects aged 9-21 with UCLP and the ten GOSLON reference models sets. The 53 sets of plaster models were scored using the MHB index and scanned with a benchtop scanner. The digital models were MHB scored visually using a commercial software program and landmarked for automatic scoring using a software plug-in. Scoring/landmarking was undertaken by three observers and repeated after 1 month. Intra- and inter-observer reproducibility were tested using Cronbach's alpha and intraclass correlation coefficients (ICC) (threshold > 0.9). Bland-Altman plots demonstrated inter-observer agreement for each model format. Random and systematic error with digital landmark identification error were determined using the x, y, and z co-ordinates for 28 models digitized twice 1 month apart using Cronbach's alpha and a t-test, respectively. Intra-operator landmark identification was excellent (Cronbach's alpha = 0.933) with no differences between sessions (P > 0.05). Intra-observer reproducibility was excellent for all examiners (Cronbach's alpha and ICC 0.986-0.988). Inter-observer reproducibility was highest for the software plug-in (0.991), followed by plaster (0.989) and OrthoAnalyzer (0.979) and Bland-Altman plots confirmed no systematic bias and greater consistency of scores with the automated software. The automated MHB software tool is valid, reproducible, and the most objective method of assessing maxillary arch constriction for patients with UCLP. The authors declare no conflict of interest or financial relationship with any organization or software used within the study.