Comparison of reliability of categorical and continuous scales for radiographic assessments of bone infill following secondary alveolar bone grafting

Abstract

Objectives: To investigate examiner reliability of scoring intraoral radiographs of bone infill following secondary bone grafting using a categorical scale (modified Kindelan Index) and a 10-cm visual analog scale (VAS). To investigate the level of clinical experience on the reliability of these assessments. Design: Retrospective study involving 10 clinicians of varying clinical experience who were blind to the patient’s identity. Setting: All grafting was carried out within the New Zealand public hospital service. Patients: Fifty-eight grafts (48 UCLP and 5 BCLP) were assessed with 14 duplicated radiographs randomly added to the sample, making a total of 72 radiographs for assessment. Main Outcome Measures: Weighted kappa was used for intra- and interrater examiner reliability for the categorical scale and correlations for the VAS. Results: Relatively poor intrarater examiner agreement for categorical scoring (median 0.46) was found. Clinicians with greater experience generally recording higher intrarater weighted kappa. The intrarater correlations for the continuous scale (median 0.89) suggest better consistency regardless of the level of experience. The interrater relationships were generally low, with an average of weighted kappa of 0.25 while the overall average VAS correlation was 0.49. For the majority of scorers, there was a relationship between the VAS and the Kindelan Index, with lower VAS scores related to worse Kindelan scores. Conclusions: The use of a VAS produced better intra- and interexaminer reliability than the categorical Kindelan Index and was more consistent despite the level of clinical experience. Further evaluation of the validity of the VAS for the assessment of bone infill is warranted.