Time Trends of Mortality and Morbidity for U.S. Older Adults

Igor Akushevich, Duke University
Arseniy Yashkin, Duke University
Julia Kravchenko, Duke University
Frank Sloan, Duke University
Anatoliy I. Yashin, Duke University

The tremendous research potential of Medicare claims data for analyzing current and forecasting future health patterns of the U.S. older adults remains largely unexplored. This study evaluated the age-adjusted mortality trends, as well as the time trends for incidence and prevalence of diabetes mellitus, a number of cardio- and cerebro-vascular diseases and common cancers. The identified patterns of change in disease incidence and prevalence were then used to explain the observed mortality trends. Finally, forecasting methodologies based on the Lee-Carter approach were used to extend these trends into the future. Data came from three major databases with available Medicare claims data: the Health and Retirement Study, the Surveillance, Epidemiology, and End Results Registry data, and the U.S. 5% Medicare data files as well as from the Human Mortality Database (the latter was used for comparisons). The patterns observed in all four datasets were in agreement with each other.