Estimating the long-term functional burden of osteoporosis-related fractures

Summary: Although fractures are associated with short-term reductions in functional status, there is limited information on longer-term burden of fracture. This study documents evidence of an association between fractures and significant declines and functional health and activities that persist but attenuate beyond two years. Introduction: Although fractures are associated with short-term reductions in functional status and may have other short-term effects on healthcare utilization (hospitalization and follow-up care), there is limited information on long-term burden of fracture beyond 12 to 24 months post-fracture. Analysis of the long-term health burden can inform policymakers, health care practitioners, and payers. Methods: We acquired a data set containing the 1992–2012 Health and Retirement Survey data linked to the same individuals’ Medicare claims. Fracture cases (n = 745) were matched to non-fracture controls using propensity scores matching. A regression-adjusted difference-in-difference (DD) approach was used to compare the change in functional status measures from baseline to two post-fracture periods for fracture cases relative to the change over the same time periods for matched controls. Self-reported measures of functional status were examined: limitations to activities of daily living (ADLs), limitations to instrumental activities of daily living (IADLs), a mobility index, a gross motor skills index, a fine motor skills index, and self-reported general health status. Results: Fracture cases reported increases in limitations to ADLs, difficulties with mobility, difficulties with gross motor skills, and difficulties with fine motor skills in each HRS collection period (the survey is administered every 2 years) following the fracture or index date (thus up to two years later) than matched controls (all p values < 0.05). The magnitude of these effects diminished in the second post-fracture wave (two to four years after fracture/index date), but they were still statistically significant. Conclusions: Results suggest that fractures are associated with significant declines in some measures of functional activities up to two years following the fracture. The effects persist beyond two years but are smaller in magnitude.