Studying Health Effects Based on Medicare Data: From Empirical Estimates to Causal Inference

Igor Akushevich, Duke University
Julia Kravchenko, Duke University

The tremendous research potential of Medicare data for studying current and forecasting future health patterns of the US older adults remains largely unexplored. In this paper, the results of the series of epidemiologic, bio-demographic, and clinical analyses obtained using the Medicare Files of Service Use are presented. In particular, analysis of age patterns of disease incidence, their time trends, recovery and long-term remission after disease onsets, mutual dependence of diseases risk, and comparative analysis of effectiveness of time dependent treatments were performed. To evaluate their characteristics, the methods of empirical analyses, regression models, and causal inference were used. In summary, being a Big Data on health, Medicare datasets are rich and powerful source of information which requires application of a spectrum of analytical approaches to use the potential of these data in its full.

  See extended abstract

Presented in Poster Session 9: Children and Youth; Data and Methods