Seminar - Reliable estimation of adjusted relative risks using the log binomial model

22 November 2012

Presenter:Professor Ian Marschner, Macquarie University

It has been said that the only excuse for an odds ratio is a case-control study. Nonetheless, although relative risks are usually easier to interpret than odds ratios, logistic regression remains the standard analysis method for prospective studies with dichotomous outcomes. One reason for this is that the log-binomial model for adjusted relative risks is prone to numerical instability and convergence problems as it does not respect the constraint that the fitted risks must lie between 0 and 1. In this talk I will present a method for fitting the log-binomial model that overcomes these problems, using a variant of the EM algorithm. As well as reliable model fitting, the method also conveniently accommodates semi-parametric generalizations which provide more flexible adjustment of relative risks. Illustrative applications from studies in acute myocardial infarction will be discussed.

About the speaker

Ian Marschner is Professor of Statistics at Macquarie University where his research focuses on the development of new statistical methodology with applications in health and medicine, particularly randomized clinical trials. He also holds the title of Professor of Biostatistics at the NHMRC Clinical Trials Centre, and he formerly held appointments as Director of Biometrics at Pfizer and Associate Professor of Biostatistics at Harvard University.

Hosted by The George Institute for International Health

Time: 11am

Location: Level 10, King George V Building, Royal Prince Alfred Hospital

Contact: Serigne Lo

Phone: 02 9657 0329

Email: 2b36030f15263c2a0c5c5d3a0b3e1f37443501792000161f5047