Survival Analysis in Epidemiology: Dr. N. Birkett

Survival Analysis in Epidemiology (EPI 5344)

Schedule of Classes

This table provides the current schedule of classes for EPI 5344. For session-specific readings, you should consult the separate material given here. The last column of the table provides links to any files which relate to the session (e.g. Power Point slides, if any).

# Date Topic

Page Ref in Allison,2nd edition)

PPTX Files PPF Files
1 March 6 Introduction to Survival methods. Incidence proportion vs. incidence rate. The hazard function. Censoring. Risk set. Actuarial and Kaplan Maier methods. Confidence intervals.
p 1-38, 49-59

A: Concepts
B: Kaplan-Meier
C: SAS code

A: Concepts
B: Kaplan-Meier
C: SAS code

2

March 13

Estimation of the hazard function. Smoothing methods. Concept of proportional hazards. Non-regression approaches to comparison of groups. Test for trend.

p 29-70

A: Hazards
B: S(t) comps
C: Trends
D: Smoothing
E: SAS code

A: Hazards
B: S(t) comps
C: Trends
D: Smoothing
E: SAS code

3 March 20 Introduction to Maximum Likelihood methods. Parameter estimation. 95% confidence interval estimation. Parametric regression models, focusing on exponential and piecewise exponential models.
p 71-80, 87-95, 98-103, 112-116

A: MLE
B: Parametric models

A: MLE
B: Parametric models

4 March 27 COX Models I. Introduction of the Cox model. Estimation of parameters. Ties. Types of predictor variables and modeling approaches. Interpretation of coefficients. HR variability by 'predictor' variables vs. 'over time'.
p 125-153

A: COX intro
B: Parameters
C: SAS code

A: COX Intro
B: Paremeters
C: SAS code

5 April 3

COX Models II. Application of Cox models to confounding and effect modification. Dose response effects and non-linearity issues. Time dependent covariates. Numerical problems.

p 153-172

A: Confounding
B: Dose response
C: Time varying
D: Numerical issues

A: Confounding
B: Dose response
C: Time varying
D: Numerical issues

6 April 10 COX Models III. Non-proportional hazard Cox models. Testing of proportional hazards assumption. Estimating survival curves from Cox models. Interval censoring, left truncation. Approaches for discrete-time models
p 172-193

A: S(t) estimation
B: PH testing
C: Various

A: S(t) estimation
B: PH testing
C: Various

Assignments

The details of the course evaluation methods and assignments are discussed here. The deadline dates for each assignment are:

Assignment #1:March 20, 2012

Assignment #2: April 27, 2012

About Us | Site Map| Contact Us | Disclaimer | ©2012 Nicholas Birkett