Leader: Associate Professor G. Heller

Offered

Alfred Hospital First semester 2008 (Off-campus)

Synopsis

Biostatistical applications of survival analysis with emphasis on underlying theoretical and computational issues, practical interpretation and communication of results. Case studies, students will explore the various methods for handling survival data. Kaplan-Meier curve definition and its extension, survival prospects using logrank test and confidence intervals for relative risks, graphical displays and assessing underlying assumptions. Mantel-Haenszel method's connection to survival analysis. Cox proportional hazards model for handling continuous covariates. Various extensions of this model, including time-dependent covariates, multiple outcomes and censored linear regression model.

Objectives

On completion of this unit students should be able to:

1. understand the major theoretical and computational issues underlying survival analysis;
2. develop appropriate survival analysis strategies based on unit matter considerations, including choice of models, control for confounding and appropriate parameterisation;
3. be proficient at using at least two different statistical software packages (eg Strata, Excel) to perform survival analysis;
4. Understand the construction, use and interpretation of appropriate graphs for showing results and checking statistical assumptions;
5. express the results of statistical analyses of censored data in language suitable for
1. communication to medical investigators and
2. publication in biomedical or epidemiological journals; and
6. appreciate the role of newer techniques including parametric non-modelling, floating odds ratios and competing risks.

Assessment

Written assignments 100%.

Prerequisites

MPH1040, EPM5002, EPM5003 and EPM5004