Premium
Evaluating periodontal disease misclassification mechanisms under partial‐mouth recording protocols
Author(s) -
Heaton Brenda,
Sharma Praveen,
Garcia Raul I.,
Dietrich Thomas
Publication year - 2018
Publication title -
journal of clinical periodontology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 3.456
H-Index - 151
eISSN - 1600-051X
pISSN - 0303-6979
DOI - 10.1111/jcpe.12874
Subject(s) - medicine , periodontitis , periodontology , quadrant (abdomen) , dentistry , periodontal disease , cohort , protocol (science) , pathology , alternative medicine
Aim To evaluate the assumptions underlying the use of partial‐mouth recording ( PMR ) protocols and the associated mechanisms of potential misclassification of periodontal disease. Methods Using data from 640 participants in the VA Dental Longitudinal Study, we compared tooth‐specific and site‐specific clinical measures and calculated sensitivity and specificity of different PMR protocols by applying the Centers for Disease Control and Prevention in collaboration with the American Academy of Periodontology definitions for periodontitis as the full‐mouth reference standard. Additionally, we evaluated alternative case definitions for PMR protocols that accounted for the reduction in numbers of teeth under observation. Results In this cohort, periodontitis presented as a generalized condition in that measures of clinical severity did not differ meaningfully according to site measured, oral quadrant or jaw. Sensitivity of disease classification under PMR protocols was a function of the number of teeth and sites under observation and the case definition applied. Sensitivity increased when case definitions were modified to account for the smaller number of teeth under observation with PMR protocols. However, specificity was reduced. Conclusions Misclassification of periodontal disease by PMR protocols is not random, even if sites under observation are randomly selected. PMR protocols can be selected/modified to maximize sensitivity, but they do so at the expense of bias in mean measures of severity.