Factors associated with the accuracy of self-reported osteoporosis in the community

We examined the agreement between self-reported osteoporosis and bone mineral density (BMD) results through dual-energy x-ray absorptiometry (DXA) using data from a national representative sample taken from the US communities. Six-year data from the continuous National Health and Nutrition Examination Survey 2005-2006, 2007-2008, and 2009-2010 were merged. Participants included adults 50 years of age or older whose data appeared in both questionnaire and medical examination data files. Self-reported osteoporosis was defined by an affirmative response to a question in the osteoporosis questionnaire then compared with BMD-defined osteoporosis, defined by BMD values taken from the examination data. Agreement between self-reported osteoporosis and DXA results were low. Kappa was only 0.24 (95 % confidence interval = 0.21-0.27), and sensitivity and positive predictive value were 28.0 and 40.8 %, respectively. When stratified by gender or age group, agreement remained poor. Self-report of osteoporosis would not be suitable for accurate prevalence estimates for osteoporosis regardless of gender or age group.