Accuracy and precision of using partial-mouth recordings to study the prevalence, extent and risk associations of untreated periodontitis.

To study the accuracy and precision of estimating the prevalence, extent and associated risks of untreated periodontitis using partial-mouth recording protocols (PRPs) A purposive sample of 431 individuals who had never been treated for periodontal disease was recruited from screening clinics at the King Saud bin Abdul-Aziz University for Health Sciences. Data were collected using questionnaires and clinical examinations. The prevalence, extent and risk associations of periodontitis were evaluated. Three PRPs were compared to full-mouth recordings (FRPs) in terms of the sensitivity, specificity, predictive values, and absolute bias. The prevalence of periodontitis was estimated with the highest accuracy and precision by examinations of the full mouth at the mesiobuccal and distolingual sites (FM)MB-DL, followed by random half-mouth (RHM) recordings. The extent of periodontitis was estimated with high precision using all the PRPs, and the absolute bias ranged from -0.6 to -2.3. The absolute bias indicated by OR for risk associations was small for the three PRPs and ranged from -0.8 to 0.8. (FM)MB-DL and RHM were the PRPs with moderate to high levels of accuracy and precision for estimating the prevalence and risk associations of periodontitis. The extent of periodontitis was estimated with high precision using all three PRPs. The results of this study showed that the magnitude and direction of bias were associated with the severity of periodontitis, the selected PRPs and the magnitude of the risk associations.