Survival data, or more generally, time-to-event data (where the “event” can be death, disease, recovery, relapse or another outcome), is frequently encountered in epidemiologic studies. Censoring is a problem characteristic of most survival data and requires special data analytic techniques.
This online medical course will give an introduction to survival analysis and cover many of the types of survival data and analysis techniques regularly encountered in epidemiologic research. The necessary statistical theory will be presented, but the course will focus on practical examples, with an emphasis on matching data analysis to the research question at hand. Lab sessions will give students the opportunity to apply the theory to real datasets using the free statistical software R.
By the end of the course, you should be able to:
• recognize or describe the type of problem addressed by a survival analysis
• define and recognize censored data
• define and interpret a survivor function and a hazard function, and describe their relation
• recognize the computer printout from a Cox proportional hazards model, a stratified Cox model, and a Cox model extended for time-dependent covariates
• state the meaning of the proportional hazards assumption and know how to check this assumption
• recognize which survival analysis technique is appropriate for a given research question and dataset
• interpret the computer printout for survival models, including hazard ratios, hypothesis testing, and confidence intervals
• Introduction to Survival Data and Analysis
Students are introduced to the characteristics of survival data, censoring (left, right, interval) & truncation. The Kaplan-Meier curves and log-rank test are introduced, followed by the Cox proportional hazards model.
• Checking the Cox Model
Parametric Models Methods for checking the assumptions of the Cox model are introduced. Students get to know stratified Cox models. In addition to the semi-parametric Cox PH model, students are introduced to fully parametric models for survival data.
• Advanced Cox regression, more on censoring and truncation
Students get more into-depth information on analyzing data with censoring and truncation. The importance of taking time-dependent covariates into account by incorporating them as time-dependent variables in the model is explained.
• Competing risks and informative censoring
Methods for handling competing risks and informative censoring are addressed.
To successfully complete this course, you need to actively participate in the discussion forums and complete the learning unit assignments, including:
• Individual and group assignments
• A final assignment: this involves the completion of a daily quiz. The course is closed by the presentation of a case study by the student. The submission deadline and the resit deadline will be announced when they become available. You are allowed to redo the final assignment once.
It might be that, due to a force majeure situation, you cannot be present during the first exam moment.
• MSc Epidemiology Postgraduate students must then, preferably prior to the first exam option, ask the academic counsellor for permission to be absent. Please note that the academic counsellor can ask for some form of proof of your absence (e.g. in case of illness) to establish if you are applicable for authorized absence. Jaco de Fockert-Koefoed, MSc is the academic counsellor you need to turn to through email@example.com.
• All other participants should contact the MSc Epidemiology Office instead, through MSc-Epidemiology@umcutrecht.nl.
In short, as of now, it is no longer possible to skip the first deadline option and -automatically- submit the assignment at the second (re-sit) deadline. Unauthorized missing the first deadline results in no longer being able to finish the course that college year.
To enroll in this course, you need:
• A BSc degree
• At least one course in basic statistical methods, up to and including simple and multiple linear regression, such as: Classical Methods in Data Analysis, Introduction to Biostatistics for Researchers, or their equivalent
• Basic programming experience in R, e.g. the ability to read in data and run a simple linear model