The purpose of this study was to evaluate the performance of self-reported measures in predicting periodontitis in a representative US adult population, based on 2009-2010 National Health and Nutrition Examination Survey (NHANES) data. Self-reported gum health and treatment history, loose teeth, bone loss around teeth, tooth not looking right, and use of dental floss and mouthwash were obtained during in-home interviews and validated against full-mouth clinically assessed periodontitis in 3,743 US adults 30 years and older. All self-reported measures (> 95% item response rates) were associated with periodontitis, and bivariate correlations between responses to these questions were weak, indicating low redundancy. In multivariable logistic regression modeling, the combined effects of demographic measures and responses to 5 self-reported questions in predicting periodontitis of mild or greater severity were 85% sensitive and 58% specific and produced an 'area under the receiver operator characteristic curve' (AUROCC) of 0.81. Four questions were 95% sensitive and 30% specific, with an AUROCC of 0.82 in predicting prevalence of clinical attachment loss ≥ 3 mm at one or more sites. In conclusion, self-reported measures performed well in predicting periodontitis in US adults. Where preferred clinically based surveillance is unattainable, locally adapted variations of these self-reported measures may be a promising alternative for surveillance of periodontitis.
Keywords: Centers for Disease Control and Prevention (CDC); NHANES; dental health surveys; epidemiologic methods; epidemiology; health status indicators.