Jennifer A Horney, Shannon C Grabich and Nathanae Rosenheim
The growing Medicare population remains disproportionately vulnerable to health effects from natural disasters. While case studies have characterized impacts of individual disasters, comparative data using common measurements over time and across locations are not available. County-level Medicare claims for beneficiaries in the 48 contiguous U.S. States for 2008-2012 were merged with Federal Emergency Management Agency data for each U.S. county from 2007-2012 to create a balanced panel dataset. To estimate average annual within-county change in the percent of Medicare beneficiaries with health conditions, exposure to a disaster declaration and a hazard factor variable were compared using fixed-effects regression models adjusted for Medicaid eligibility and beneficiary age, gender, and race. Using the disaster declaration exposure we found significant negative (diabetes, high cholesterol) changes for counties with declared disasters. Using the hazard exposure factor variable resulted in significant positive (heart attack, diabetes, arthritis) and negative (diabetes, high cholesterol) changes in disaster exposed counties. Associations between disaster exposure and health outcomes among Medicare beneficiaries are mixed and should be explored further. This can inform the development of innovative and timely pre-disaster interventions, as well as contribute to enhanced disaster resilience among beneficiaries and the Medicare system.
