Burden of Osteoporosis among U.S. Adults: Estimates from Medicare Claims Data

Purpose The objectives were to estimate the recent burden of osteoporosis (OP) using Medicare claims data and to assess variation in OP prevalence by age, gender, race, and geographic region. Methods The Medicare Chronic Condition Data Warehouse provided 1999–2004 claims data on an enhanced 5% probability sample of Medicare beneficiaries. The analysis included beneficiaries covered for 12 months in each year by Medicare part A and B and not enrolled in a health maintenance organization. We developed four case definitions for OP using increasingly rigorous criteria for assessing the patterns of claims for OP. We also developed two estimates of the burden of OP. The first was the proportion of beneficiaries that had at least one claim for OP in each year. The second was the prevalence of OP in 2004, calculated as the proportion of beneficiaries with at least one claim for OP during the entire study period of 1999 to 2004. Results using the most conservative case definition are presented. Sensitivity analyses based on variations in case definitions, insurance coverage requirements and other factors were conducted. Results From 1999 to 2004, the total number of eligible beneficiaries ranged from 1,712,614 to 1,855,200 persons, a substantial proportion of whom was 65–74 years (36.9%–38.3%), female (61.2%–63.0%), white (86.0%–86.9%) or black (8.9%–9.4%). Using the most rigorous OP definition, the annual proportion of beneficiaries with at least one OP claim ranged from 5.2% in 1999 to 6.6% in 2004. Overall OP prevalence was 14.9% in 2004. OP prevalence was 15.7 % among whites and 6.7% among African Americans; was higher in women (22.1%) compared to men (3.4%); and increased with age, from 5.4% among those under 65 to 22.6% among those 85+ years of age. Regional variation in prevalence was small. Conclusion These results indicate sizeable differences in the burden of OP according to race, gender, and age. Results are discussed with respect to the varying case definitions and the methodological limitations of using claims data for epidemiologic studies.