Agreement among dentists' recommendations for restorative treatment.

Differences among dentists in their decisions to recommend treatment have been studied principally at a "macro" level by examinations of variations in dentists' treatment recommendations for individual patients, or for patients aggregated to the practice level. If reasons for these differences are to be understood, however, a "micro" level examination of differences in dentists' decisions to recommend treatment for individual teeth is necessary. Extent of agreement to recommend treatment among dentists was explored for 1187 teeth in 43 patients. Each tooth was examined by a mean of 6.6 general practitioners. A modified reliability index and distributions of raw levels of agreement were used for quantification of agreement rates across an incomplete cross-classification for practitioners and patients. Overall inter-dentist reliability in recommending individual teeth for treatment was moderate, p = 0.62. Among unrestored teeth with no clinically evident caries, as judged by an independent experienced evident caries, as judged by an independent experienced examiner, reliability in treatment recommendations due to caries was fair, p = 0.51. Reliability was lowest for recommendations concerning previously restored teeth, p = 0.43. Among all teeth receiving at least one recommendation for treatment, only 22% received a unanimous recommendation. Over one-half of all instances of lack of agreement resulted from a single dentist's recommendation differing from those of all other dentists examining the tooth. The results suggest that much of the variation in dentists' practice profiles is due to basic differences in decisions to recommend treatment for individual teeth with specific conditions, and that the presence of previous restorations seems to magnify these differences.