Survival Analysis Applied to Biostatistics
Objectives
Particularities of survival data (censoring, truncation, etc.)
Risk and survival functions, and other representations
Non-parametric estimation of survival functions – Non-parametric comparisons of survival distributions
Semi-parametric models using the Cox model – Case of time-dependent variables.
Verification of Cox model assumptions: proportional hazards assumption and log-linearity assumption
Advanced survival models: introduction of the notion of competitive risk
Use of propensity scores in a Cox model
Course outline
Define the relationship between survival function, distribution function, instantaneous hazard function and density.
Identify the presence of censoring and truncation in a data set.
Describe and interpret survival data using the Kaplan-Meier estimator.
Formulate research questions involving survival data as regression problems.
Solve classical multi-sample comparison problems of survival distributions.
Implement a duration model using statistical software (R).
Check the goodness of fit of the model to survival data.
Interpret inference results on survival data using multivariate semiparametric models.
Incorporate time-dependent covariates into the proportional hazards model.
Identify deviations from key model assumptions
Prerequisites
Probability, inferential statistics, SAS, R (1A)
Linear regression, GLM, Duration model, Martingales, Resampling methods (2A)