Comparison of Estimates of Edentulism Among US Older Adults Between the Behavioral Risk Factor Surveillance System and the National Health and Nutrition Examination Survey

BACKGROUND:   Reducing edentulism (complete tooth loss) among 65-74 year olds is a Healthy People 2020 (HP2020) objective OH-4.2. The National Health and Nutrition Examination Survey (NHANES) is the official data source for this national target with estimates based on clinical examination by standardized dentists. NHANES, however, is not designed to provide state-level estimates. The Behavioral Risk Factor Surveillance System (BRFSS)  is the primary data source for state-level estimates for self-reported edentulism and also provides national estimates. Complete tooth loss among adults, 65 years and older, from BRFSS is a National Oral Health Surveillance System indicator. This study compared estimated prevalence of edentulism (EPE) among US adults, 65 years and older between BRFSS and NHANES, overall and by selected characteristics, for two time periods 1999-2004 and 2005-2010.

METHODS:  We pooled data for each survey to create national estimates for two time periods: 1999-2004 (BRFSS 1999, 2002 and 2004; NHANES 1999-2000, 2001-2002, 2003-2004) and 2005-2010 (BRFSS 2006, 2008, and 2010; NHANES 2005-2006, 2007-2008, 2009-2010). We used BRFSS data from years in which the tooth loss question was asked in all 50 states and DC. We used the 2000 standard population to calculate age-adjusted EPE overall and stratified by age, race/ethnicity, education level, and family poverty status from BRFSS and NHANES. We calculated absolute differences (AD) in estimates between time periods within each survey, and between BRFSS and NHANES within each time period. We report statistically significant differences based on the z-test (p≤0.05).

RESULTS: Both surveys showed decreases in EPE overall and among subgroups from 1999-2004 to 2005-2010, although decreases in NHANES estimates were not statistically significant for some subgroups. Overall EPEs from BRFSS were lower than those from NHANES by 4.9% in 1999-2004 (22.7% vs. 27.6%) and 4.0% in 2005-2010 (18.0% vs. 22.0%). Some groups also showed lower EPE from BRFSS than NHANES in both time periods: 75+ year olds (AD=-6.2%; AD=-5.7%), non-Hispanic whites (NHW) (AD=-4.3%; AD=-4.2%) and near-poor [100-199% Federal Poverty Level (FPL)] (AD=-6.7%; AD=-3.9%); others in one time period only: 65-74 year olds in 1999-2004 (AD=-3.7%), and  non-poor (≥200% FPL) (AD=-4.0%) in 2005-2010.

CONCLUSIONS:  NHANES and BRFSS both showed generally consistent declines in EPE from 1999-2004 to 2005-2010, overall and in selected subgroups. No estimates of edentulism were higher in BRFSS than in NHANES (p≤0.05). When monitoring progress in reducing edentulism, the lower estimates from BRFSS relative to NHANES estimates, overall and among some subgroups, may be considered.