Errors in measuring parameters associated with periodontal health and disease.

The intra-examiner reliability for the Plaque Index (PII), Calculus Index (CI), Gingival Index (GI), and periodontal loss of attachment (LA) was evaluated in 2 different physical and clinical environments. Examiner reliability for each index was consistently better when clinical evaluations were performed in an office setting on subjects who had low levels of periodontal disease than under more difficult field conditions evaluating subjects with higher disease levels. In an office environment intra-examiner reliabilities were very good for subject-based averages (intraclass r's were 0.85, 0.83, 0.81, and 0.78 for GI, PII, LA, and CI, respectively); they were moderate to good for site-based scores (kappas were 0.63, 0.54, and 0.47 for CI, GI, and PII). Under field conditions reliabilities ranged from excellent to fair for subject-based averages (intraclass r's of 0.98, 0.86, 0.76, and 0.33 for LA, CI, GI, and PII, respectively); for site-based scores they ranged from moderate to fair (kappas of 0.50, 0.49, and 0.21 for GI, CI, and PII, respectively). Small but significant systematic errors were detected under field conditions for the PII and CI, and for the PII in the office environment as well. The results observed suggest that the examination process can affect the clinical status of the surface or site when replicate assessments of plaque and calculus are conducted within a 24-hour period. Our results also confirm that measurement error standard deviation for loss of attachment evaluated with a periodontal probe can be maintained at 0.13 mm for whole-mouth means and 0.54 mm for individual sites, even when examining subjects with substantial periodontal disease under field conditions.