Bayesian Multistate Life Table Methods for Complex, High-Dimensional State Spaces

• In Event: 4464 - Regular Session on Methodology: Progress and New Approaches

Abstract

Multistate life table methods are an important tool for producing easily understood measures of population health. Most contemporary uses of these methods involve sample data, thus requiring techniques for capturing uncertainty in estimates. Over the last two decades, several methods have been developed to do so. Among these existing methods, the Bayesian approach proposed in Lynch and Brown (2005) has several unique advantages. However, this approach has been limited to estimating years to be spent in only two states, such as years to be lived "healthy" vs. "unhealthy." In this paper, we extend that method to allow for complex, high-dimensional state spaces with "partially absorbing'' states. We illustrate the new method and show its advantages using data from the Health and Retirement Survey (HRS) to investigate US regional differences in years of remaining life to be spent with diabetes, chronic conditions, and disabilities. The method works well and yields rich output for reporting and subsequent analyses.

Authors

• Emma Zang, Yale University

• Scott M. Lynch, Duke University