Hip fractures are a leading cause of morbidity and associated health care costs in older people in the United States. As the general disease burden of the aging population and related cost pressures on the United States health care system mount in the years ahead, efforts are needed to reduce the morbidity and cost burden of hip fractures. While interventions to reduce the incidence of hip fracture have had modest success, costs of treating and rehabilitating hip fractures must also be reduced to lower the economic burden. Previous studies of hip fracture costs have been based on claims data only, have focused on estimation of mean costs across specific populations or on health care system structural characteristics that influence post-acute care use after hip fracture, and have not examined factors that may explain the large individual variation in costs due to hip fracture. In this proposal and in response to PA-09-265 issued by the National Institute on Aging, we propose use a database linking the landmark NIH-funded prospective cohort study of the epidemiology of fractures, the Study of Osteoporotic Fractures (SOF), with Medicare claims data. Unlike previous fracture outcome studies that have at most retrospective assessment of very limited pre-fracture characteristics, the SOF cohort has comprehensive, multiple measurements of pre-fracture characteristics for each study participant and has confirmed 1254 incident hip fractures since January 1, 1992 among 10,336 women. The linked SOF-Medicare claims dataset offers a unique opportunity to quantify with accuracy and precision the health care costs attributable to hip fracture as well as provide robust estimates of the associations of individual patient pre- fracture characteristics and health care system structural variables with health care costs and utilization attributable to hip fracture.
The specific aims of our proposal are to: 1) Estimate the associations of pre-fracture trajectories of physical performance, positive affect, depression, cognitive function and body weight with health care costs, acute hospital length of stay, skilled nursing facility (SNF) length of stay, and inpatient rehabilitation facility (IRF) length of stay attributable to hip fracture;2) Estimate the associations of health care system structural variables (characteristics of local acute care hospitals, SNF's, IRF's, home health care agencies, and health care provider supply) with hip fracture attributable costs and acute hospital, SNF, and IRF lengths of stay after hip fracture;and 3) Estimate the long-term trajectory of health care utilization and costs attributable to hip fracture. Quantification of the health care costs attributable to hip fracture and identification of the associations of individual patient pre-fracture predictors and health care system structural variables that independently predict health care use and costs will lay the foundation for future intervention studies and policy planning to modify these factors such that the overall societal cost burden of hip fractures can be reduced.
We propose to use a database linking the Study of Osteoporotic Fractures prospective cohort dataset with Medicare claims data to accurately and precisely quantify the health care costs attributable to hip fracture. We will then estimate the associations of individual patient trajectories of pre-fracture characteristics and health care system structural variables with health care costs and utilization attributable to hip fracture. Quantification of health care costs attributable to hip fracture and identification of the associations of individual patient pre- fracture predictors and health care system structural variables that independently predict utilization and costs will lay the foundation for future intervention studies and policy planning to reduce the overall societal burden of hip fractures.
