Relationships between clinician variability and radiographic guidelines.

Abstract

This study evaluated the replicability of clinical measurements under careful calibration of multiple dentists and how the replicability can relate to their use as selection criteria in guidelines for prescribing dental radiographs. For 48 consenting patients, three dentists clinically examined each patient and recorded the presence of all clinical findings using standardized selection criteria. The examinations were performed independently of each other, but with periodic conferences of the dentists to clarify general measurement criteria. The degree of agreement among the dentists is described by the interrater agreement kappa for several standard clinical indications such as rating of caries, periodontal disease, and tooth mobility. Almost perfect agreement was obtained for tooth status, restoration size, and restoration material. Moderate agreement resulted for measures of caries, defective restoration presence, and gingival recession presence. Only fair agreement was obtained for other periodontal disease measures. The relationship between extent of agreement and guidelines' results was examined for the FDA Guidelines. The differences among the dentists' clinical measurements resulted in considerable differences among the radiographs that were selected by the FDA Guidelines' criteria. Even so, the missed disease rates for 490 patients in a larger study of the FDA Guidelines' efficacy were very low and did not vary greatly among the three dentists. We conclude that guidelines' criteria can be quite robust to variation from dentists' clinical measurement differences, as seen from the FDA Guidelines applied under the idealized setting where the dentists are periodically recalibrated through group discussions of the clinical measurements' definitions and interpretations.