A study of examiner errors associated with measurement of denture plaque

The purpose of the present study was to compare the reproducibility and reliability of different methods of measuring the distribution of plaque on the fitting surface of maxillary dentures. Data from a clinical trial were used to determine the reproducibility of (1) the original Budtz-Jörgensen index, (2) a proposed morphometric method for facilitating the measurements by this index, and (3) the index of Schubert and Schubert. The recordings from previous tests of reproducibility of an additive index were used to estimate reliability by means of a test-retest procedure and the internal consistency method. The morphometric point-estimator scoring method gave the highest intra- and inter-examiner reproducibility for the three-graded scale (Scott's pi, 0.90 and 0.81) and for the five-graded scale (0.86 and 0.80). Reproducibility was markedly reduced when the distribution of denture plaque was assessed by visual examination (Scott's pi, 0.81 and 0.72), especially when using the five-graded scale (0.73 and 0.51). The PH index exhibited an intra-examiner reliability of 0.83 with the three-graded scale and 0.78 with the five-graded scale. The corresponding inter-examiner figures were 0.76 and 0.70. The test-retest method showed that in the areas M1 and M2 the mean amount of plaque was significantly reduced at the second examination, whereas no significant change occurred in areas T1 and T2. The reliability coefficients obtained by the additive index were greater than 0.88. Error variance was 8% and 12% of the total variance for this index but only 3% with the morphometric test method. The internal consistency reliability gave coefficients that were higher than those obtained by the test-retest method.