Examiner consistency and group balance at baseline of a caries clinical trial

Two experienced investigators (G.L. & H.H.) independently examined 629 children in grades 6-9 (ages 10-17 yr) for baseline DMFS data in a clinical trial of a caries preventive. The examiners used the same written and visual (slides) criteria for dental caries diagnosis, but did not standardize or calibrate their methods before or during the survey. Results showed overall mean DMFS scores for Examiners 1 and 2 that were remarkably similar, 8.35 and 8.16, respectively; coefficients of variation were identical, C.V. = 87%. The reliability coefficient for the two sets of data showed that only 4% of the variability in DMFS scores was due to examiner inconsistency and other measurement errors. The findings indicate that, without undergoing clinical calibration, the two experienced examiners attained a high level of agreement in scoring dental caries merely by adhering to clearly defined written and visual criteria. Only the 308 children in the 6th grade (ages 10-14 yr) participated in the study (children in grades 7-9 were a reference population). Participants were randomly assigned to one of two treatment groups. The allocation procedure produced mean DMFS scores for Groups I and II of 7.87 and 6.17 (Examiner 1) and 8.07 and 6.41 (Examiner 2), respectively. The mean scores differed by about 21% (II compared with I) for each examiner. Both differences were clinically and statistically significant (P less than 0.05). Randomized assignment had generated an imbalance of baseline DMF scores by group.