University of Newcastle upon Tyne

School of Mathematics and Statistics

Statistics Seminars 2004-2005

29 April 2005, L401, 2:15pm

Professor Ørnulf Borgan, Division of Statistics and Insurance Mathematics, University of Oslo

Abstract

I will consider situations where the outcome of main interest is a survival time, and where information on fixed as well as internal time-dependent covariates is available for the subjects under study. For such situations, a regression analysis including all covariates will give insight on the importance of the internal time-dependent covariates, but it may underestimate the effect the fixed covariates, including treatment. On the other hand, an analysis without the internal time-dependent covariates will give a correct estimate of the effect of the fixed covariates, but it will offer no information on the effects of the internal time-dependent covariates

In the talk I will outline how one, by using a generalization of classical path analysis, may reconcile these two approaches. In particular the new approach, denoted dynamic path analysis, makes it possible to obtain a detailed picture on how treatment and other fixed covariates partly have a direct effect on survival and partly have an indirect effect mediated through the internal time-dependent covariates. The path analysis is performed, at each failure time, by using ordinary linear regression to estimate the relation between the covariates, while Aalen's additive hazard model is used for regressing the survival times on covariates. The methodology will be illustrated using data on a randomized trial on survival for patients with liver cirrhosis.

The talk is based on joint work with Johan Fosen, Egil Ferkingstad, and Odd O. Aalen. 

Return to timetable for 2004-05 seminars

School of Mathematics and Statistics Homepage