Workshop on Analysis of Survival and Event History Data

May 16-19, 2011

Program

Monday, May 16, 2011

08:30 - 09:00
Registration
Room(s) 5345

08:30 - 09:00
Coffee & Croissants
Room(s) 6245

Session I - An Overview of current issues in life history studies - 1
Meeting room(s) : 6214

This day will be devoted to the discussion of some particular large cohort studies including the Womens Health Initiative and the Ontario Health Study. Additional talks will survey methodological issues involving the management of kidney transplant registries and genetic epidemiology for chronic disease processes.


09:00 - 09:10
Welcoming addresses
09:10 - 10:00
Ross Prentice
(Fred Hutchinson Cancer, Research Center)
Opportunities and challenges arising in the women’s health initiative clinical trial and cohort study
Abstract
10:00 - 10:50
John D. Kalbfleisch (University of Michigan, School of Public Health)
Robert Wolfe (University of Michigan, School of Public Health)
Monitoring outcomes of medical facilities and procedures
Abstract
10:50 - 11:10
Coffee break
Room(s) 6245
11:10 - 12:00
Lyle J. Palmer
(Ontario Institute for Cancer Research)
The Ontario health study: analytic issues related to large-scale longitudinal cohort studies
Abstract
12:00 - 12:15
General Discussion
12:15 - 13:45
Lunch break

Session I - An Overview of current issues in life history studies - 2
Meeting room(s) : 6214

13:45 - 14:35
Danyu Lin
(University of North Carolina)
Survival analysis with incomplete genetic data
Abstract
14:35 - 15:00
General Discussion

15:00 - 15:30
Coffee break
Room(s) 6245

Session - Chaire Aisenstadt Chair
Meeting room(s) : 6214

15:30 - 17:00
James Robins
(Harvard School of Public Health)
Multiply robust higher order U-statistics estimators for continuous time dependent censoring
Abstract

17:00
Reception
Room(s) 6245

 

Tuesday, May 17, 2011

08:30 - 09:00
Coffee & Croissants
Room(s) 6245

Session II- Joint models for longitudinal and event history data
Meeting room(s) : 6214

This session will be devoted to the discussion of approaches to the construction of joint models for longitudinal data and lifetime event data. The importance of adopting modeling frameworks which address the needs of particular applications will be emphasized. Applications involving cancer biomarkers and markers of chronic disease processes such as diabetes will be considered.

09:00 - 09:45
Peter J. Diggle
(Lancaster University)
Monitoring progression towards end-stage renal failure
Abstract
09:45 - 10:30
Joseph W. Hogan
(Brown University)
Causal Inference for Survival and Recurrent Events with Censored Exposure: Application to Treatment of HIV/TB Coinfection in Western Kenya
Abstract
10:30 - 11:00
Coffee break
Room(s) 6245
11:00 - 11:45
Jeremy M.G. Taylor
(University of Michigan)
Joint modelling of longitudinal and survival data to estimate treatment effects
Abstract
11:45 - 12:00
General Discussion

12:00 - 13:30
Lunch break

Session III - Incomplete or mismeasured covariate data
Meeting room(s) : 6214

Longitudinal and life history studies routinely involve incomplete response or covariate data. In many cases the response or covariate values are difficult to observe directly and more easily determined surrogate or auxillary variables are available instead. These represent the true response or covariate measured with error. Suitable methods for addressing these missing or mismeasured data problems will be reviewed including modelling strategies, imputation, and inverse weighted estimating equations.

13:30 - 14:15
Roderick J. Little
(University of Michigan)
On repeated measures models with incomplete covariates that are not missing at random
Abstract
14:15 - 15:00
Joseph Ibrahim
(University of North Carolina)
Diagnostic measures and goodness of fit statistics for Cox regression model with missing covariates
Abstract
15:00 - 15:30
Coffee break
Room(s) 6245
15:30 - 16:15
Donna Spiegelman
(Harvard University)
Measurement error correction for survival data analysis with time-varying covariates
Abstract
16:15 - 16:30
General Discussion

 

Wednesday, May 18, 2011

08:30 - 09:00
Coffee & Croissants
Room(s) 6245

Session IV - Predictive models for lifetime events
Meeting room(s) : 6214

There is an increasing interest in the development of statistical methods for the prediction of outcomes in individual subjects arising from the desire to advance the field of personalized medicine. Such models and methods can be used both to classify individuals into risk groups and also to guide treatment decisions based on patient characteristics at diagnosis and preliminary response to treatment. Criteria for assessing the predictive ability of models will be considered along with suitable strategies for model validation and inference.

09:00 - 09:45
Tianxi Cai
(Harvard School of Public Health)
Risk prediction with complex studies
Abstract
09:45 - 10:30
Richard Simon
(National Cancer Institute)
Use of resampling methods for the evaluation of high-dimensional survival risk models
Abstract
10:30 - 11:00
Coffee break
Room(s) 6245
11:00 - 11:45
Patrick Heagerty
(University of Washington)
Non-parametric estimation of a time-dependent predictive accuracy curve
Abstract
11:45 - 12:00
General Discussion

12:00 - 13:30
Lunch break

Session V - Statistical models and methods for multivariate and clustered lifetime data
Meeting room(s) : 6214

Strategies for the construction of multivariate models will be considered including those based on random effect or frailty models, copula functions and multistate models. Merits of these different approaches will be considered for assessing the effect of cluster-level and within-cluster covariates. Particular scientific studies including twin and family studies will be used as the basis for discussion.

13:30 - 14:15
Rebecca Betensky
(Harvard School of Public Health)
Analysis of failure time in the presence of dependent truncation
Abstract
14:15 - 15:00
Li Hsu
(Fred Hutchinson Cancer Research Center)
On estimating the marginal hazard function using the family history information of case-control data
Abstract
15:00 - 15:30
Coffee break
Room(s) 6245
15:30 - 16:15
Yi Li
(Dana-Farber Cancer Institute)
A new class of estimating equation-based dantzig selectors for marginal regression models based on correlated outcomes
Abstract
16:15 - 16:30
General Discussion

 

Thursday, May 19, 2011

08:30 - 09:00
Coffee & Croissants
Room(s) 6245

Session VI - Response-biased sampling
Meeting room(s) : 6214

Epidemiological studies often involve a response-biased design for reasons of efficiency in cost and time. Such studies have received increased attention in recent years with the development of nested case-control and case-cohort designs. Strategies for the efficient selection of patients and robust analysis will be considered in this session for a variety of settings including disease history studies and genetic epidemiology studies.

09:00 - 09:45
Ornulf Borgan
(University of Oslo)
Nested case-control and case-cohort designs: an overview and some future challenges
Abstract
09:45 - 10:30
Nilanjan Chatterjee
(National Cancer Institute)
Estimating penetrance of rare genetic mutation using ascertained families
Abstract
10:30 - 11:00
Coffee break
Room(s) 6245
11:00 - 11:45
Alice S. Whittemore
(Stanford University School of Medicine)
Two-stage sampling designs for validating personal risk models
Abstract
11:45 - 12:00
General Discussion

12:00 - 13:30
Lunch break

Session VII - Issues in the analysis of complex life history processes
Meeting room(s) : 6214

This session will involve the study of approaches to modeling complex life history data with particular attention given to time-varying covariate processes, multiple types of life time events, and multistate models. The utility of latent variable (parameter-driven) models and data-driven models will be debated.

13:30 - 14:15
Stephen Cole
(University of North Carolina at Chapel Hill)
Marginal structural models for case-cohort study designs to estimate the effect of antiretroviral therapy initiation on incident AIDS or death
Abstract
14:15 - 15:00
Somnath Datta
(University of Louisville)
Nonparametric inference for multistate models
Abstract
15:00 - 15:30
Coffee break
Room(s) 6245
15:30 - 16:15
Douglas E. Schaubel
(University of Michigan)
Partly conditional estimation of the effect of a time-dependent factor in the presence of dependent censoring
Abstract
16:15 - 16:30
General Discussion