Workshops

Two full day (9am - 5pm) pre-conference workshops will be held at the Australian National University in Canberra on Saturday 5 February 2005.
Workshop registration fees

	Before 29 Oct 2004	After 29 Oct 2004
Participant	$250	$300
Student participant	$100	$150

Registration form         Accommodation information (from 2003)         Canberra-Thredbo travel information

IMPORTANT CORRECTION
The secure fax number for workshop registration is +61 2 6216 7112.

Please make sure this is the number you use.

Joint modelling of survival and longitudinal data	Design and Analysis of Microarray Studies with GenStat
Robin Henderson Lancaster University, UK	David Baird AgResearch, NZ
Assumed knowledge	Assumed knowledge
Participants should 1. have a basic knowledge of survival analysis including the Cox proportional hazards model 2. be familiar with the material in Analysis of Longitudinal Data (Diggle, Liang and Zeger, 1994 or Diggle, Heagerty, Liang and Zeger, 2002) on GEEs, GLMMs and techniques for incorporating lagged variables in models	Participants should 1. have a basic understanding of microarray technology 2. be familiar with the Windows operating system Familiarity with GenStat is not assumed, but would be an advantage.
Workshop contents	Workshop contents (draft)
Session 1: Biomarkers for Survival Cox PH in R.Measurement error and frailty. SIMEX. Regression calibration. Time-dependent covariates. Intermittently observed time-dependent covariates: biomarkers. A joint model. Session 2: Dropout from Longitudinal Studies Longitudinal data and standard models. Dropout. Effects of dropout. Diagnostics. Models: pattern mixture, selection, random effect joint models. Assumptions and sensitivity. Session 3: Further Topics from Joint Modelling 1. Semiparametric estimation for the PH model with biomarkers: Tsiatis and Davidian approach. 2. Dropout in paired comparisons	Session 1: Designs for microarray studies. Session 2: Reading in microarray data. Visualization of microarray data. Session 3: Normalization. Estimation of effects. Error estimation using Empirical Bayes. Session 4: Mixture models for estimating False Discovery Rate and Power. Clustering of Slides, Probes and Targets.
Location	Location
Moran G007	Moran G008