Multiple cause models of disease dependency

Stochastic compartment model strategies are applied to the estimation of the transition rates for primary and secondary transitions when only information on the medical conditions present at the time of death is available. It is shown that these estimates are obtainable when certain general conditions are placed upon the change of hazard rates over short time intervals. The relations of these transition parameters to various actuarial functions are presented. The consistency of the estimates of the transition parameters, their asymptotic normality and estimates of their asymptotic covariance matrix are also presented. © 1983, Taylor & Francis Group, LLC. All rights reserved.

Duke Scholars

Author Kenneth G. Manton Trinity College of Arts & Sciences