Theory of Individual Health Histories and Dependent Competing Risks

• Activity Number: 466
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Risk Analysis
• Author(s): Igor Akushevich and Konstantin Arbeev*+ and Svetlana Ukraintseva and Anatoliy Yashin
• Companies: Duke University and Duke University and Duke University and Duke University
• Keywords: cause-specific mortality risks ; longitudinal data ; risk of disease onset ; observational plans ; age trajectories ; stochastic differential equations

Abstract

Data on individual health histories, age trajectories of physiological variables and cause-specific mortality allow for studying joint evolution of health and physiological states and their effects on mortality risk. Generally cause-specific mortality risks are mutually dependent. Here we present a competing risk model with a weaker assumption of conditional independence of cause-specific time to death, given a stochastic process with two mutually dependent continuous and jumping components. The jumping component describes fast changes in health status and the continuous one describes slower individual physiological aging. Formulation of the exact model based on the Kolmogorov equations for conditional density, the Gaussian approximation, and the parameter estimation in different observational plans in longitudinal measurements are discussed and properties of the model estimation are investigated in simulation study. Application to the Framingham Heart Study resulted in evaluation of a system of indices representing aging related changes and indicators of health. The results suggest the importance of avoiding the marginal independence assumption when appropriate data are available.