This study evaluated the validity of self-reported periodontitis measures among 2404 Japanese adults aged 40 to 75 years. A self-administered questionnaire survey and a clinical periodontal examination were conducted from 2013 through 2016. The self-reported periodontitis questions included 3 sociodemographic, 3 health, and 5 periodontal health-related items. Based on the clinical case definition of periodontitis, 26.5% of participants were found to be periodontally healthy, 2.7% had mild periodontitis, 55.2% moderate periodontitis, and 15.6% severe periodontitis. No single self-reported question demonstrated satisfactory validity in predicting the presence or absence of periodontitis. The predictive ability in mild and/or moderate periodontitis was poor even after combining multiple sociodemographic, health, and periodontal health-related questions. In severe periodontitis, the model including age, sex, education level, smoking status, diabetes history, body mass index, informed by a dentist, gingival bleeding, calculus deposit, and tooth mobility, presented moderate predictive performance (C-statistic: 0.676, sensitivity: 65.2%, and specificity: 61.1%). An age-stratified analysis on severe periodontitis showed that sensitivity was higher, and specificity was lower in older age group (60-75 years) than younger age group (40-59 years). Further refinement of questions in the self-report is required to increase the accuracy of the prediction of clinical periodontitis.
Keywords: periodontal disease; self-report; sensitivity; specificity; validity.