Periodontal Disease: A Two-way Random Effects Analysis

• Activity Number: 312
• Type: Topic Contributed
• Date/Time: Tuesday, August 9, 2005 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Lei Qian*+ and Vladimir Spolsky and Robert Weiss
• Companies: University of California, Los Angeles and University of California, Los Angeles and University of California, Los Angeles
• Keywords: tooth disease ; probing depth ; Russell's Index ; Bayesian analysis ; WINBug

Abstract

Numerous screening systems have been developed to detect and assess periodontal disease, a major cause of tooth loss in adults. Researchers have found inconsistencies in these criteria, especially when studying the association of periodontal disease with other diseases. We used a Bayesian two-way random effects logistic regression to examine the relationship of two sets of major disease criteria, Russell's Periodontal Index and three probing periodontal measurements used by NIDCR: Marginal Bleeding, Calculus, and Probe Depth. A sample of 22 adults with up to 32 teeth was used to illustrate the approach. There were two crossed random effects, subjects, and teeth. The results showed that among the three probing measurements, Marginal Bleeding and Calculus were significantly related to Periodontal Index, but Probing Depth was not. Some modifications of the model were suggested based on the results. Finally, a possible way to fix the inconsistency between Periodontal Index and Probing Depth was discussed.