• Start date:
  • 6 November 2017
  • Duration:
  • 5 days
  • Working as:
  • PhD

Survival analysis is the study of the distribution of life times, i.e. the times from an initiating event (birth, diagnosis, start of treatment) to some terminal event (relapse, death). This type of data analysis is most prominently (but not only) used in the biomedical sciences. A special feature of survival data is that it takes time to observe the event of interest. As a result for a number of subjects the event is not observed, but instead it is known that it has not taken place yet. This phenomenon is called censoring and it requires special statistical methods.

During the course different types of censored data will be introduced and techniques for estimating the survival function by employing both parametric and non-parametric methods will be illustrated. Multiplicative hazards regression models , testing and inference techniques will be studied in great details. Special aspects as time-dependent covariates effects, stratification, time and prediction will be introduced. Techniques to be used to assess the validity of the hazard regression model will be discussed. Alternative to Cox model will be discussed and predictive models will be introduced. The last part of the course focus on more advanced models like competing risks and multi-states.

A competing risks model is concerned with failure time data where each subject may experience one of the K different type of terminal events. Multi-states are employed when some intermediate events may occur before the final event of interest and one is interested in the effects of the occurrence of those intermediate events on the final events. Also for these more complex models estimation and prediction techniques will be discussed. The course ends with a discussion about sample size calculations.

Practicals
When participants don't have a workstation in the LUMC, they are required to bring their own laptop for practicals. To use the wireless network, the mac-address of the laptop should be registered with the LUMC and SPSS should be installed. LUMC registered laptops are required for the computer practicals. The other software used during the course is freely available and instructions will be given how to install the software. A certificate will be awarded to participants who attended the course in full.

Requirements
Basic knowledge of statistics (e.g. the Boerhaave course Basic methods and reasoning in Biostatistics) and of regression models (e.g. the Boerhaave course Basic methods and reasoning in Biostatistics).

Teaching environment
Lectures, self-study assignments and practical sessions with SPSS and R

Exams
None (to obtain a proof of participation, lectures should have been attended and special practical exercises should have been submitted). Students who wants to obtain ECTS for participation and the submitted practical exercises please click on the checkbox during your registration for the course.

Course material
Slides of the lectures

Target group
Master and PhD students in the bio-medical sciences

Organizing committee

  • Ms. Marta Fiocco, LUMC, Medische Statistiek
  6 November teacher M. Fiocco
08:30 Registration (with coffee or tea)
09:00 Introduction to survival analysis, examples
09:30 Censoring
10:15 Definition of important concepts: hazards, survival curves  
10:45 Coffee break 
11:15 Parametric distributions  
11:45 The Kaplan-Meier estimate, Median survival
12:15 Lunch break
13:30  Practical exercises 
15:30 Plenary discussion of practical exercises
   
  7 November teacher M. Fiocco
09:00 Recall concepts from day I 
09:15 Calculation of follow-up, median follow-up 
09:30 The Nelson-Aalen estimate of the cumulative hazard 
10:00 The life table method
10:30 Coffee break
11:00 Alternative observation schemes; left truncation and censoring, interval censoring
11:15 Estimation in parametric distributions 
11:45 Comparison of survival curves: the log-rank test
12:30 Lunch break 
13:30 Practical exercises
16:30 Plenary discussion of practical exercises
   
  8 November 2017 teacher M. Fiocco
08:45 Registration (no coffee or tea available)  
09:00 Regression methods
09:15 Cox’s proportional hazards model
10:30 Special aspects Time-dependent covariates
10:45 Coffee break
11:00 Special aspects Time-dependent covariate effects. Stratified Cox regression  
11:30 Predicted survival for specific covariates  
12:30 Lunch break
13:30 Practical exercises
16:30 Plenary discussion of practical excercises
   
  9 November teacher M. Fiocco
08:45 Registration (no coffee or tea available)
09:00 Cox’s proportional hazards model Checking model assumptions, goodness-of-fit, explained variation
09:45 Alternative regression methods Accelerated failure time model 
10:45 Coffee break 
11:15 Alternative regression methods Poisson regression  
12:00 Lunch break
13:30 Practical exercises 
15:30 Plenary discussion of practical exercises 
   
  10 november teacher H. Putter
08:45 Registration (no coffee or tea available)
09:00 Competing risks and multi-state models 
10:00 Clustered data
10:45 Coffee break
11:00 Sample size calculations
12:00 Interim analysis 
12:30 Lunch break
13:30 Practical exercises
16:30 Plenary discussion of practical exercises 
M. Fiocco
Leids Universitair Medisch Centrum (LUMC)
H. Putter
Leids Universitair Medisch Centrum (LUMC)
LUMC
Albinusdreef 2, 2333 ZA Leiden

Price

Participation fee for employees LUMC  € 150,00

Participation fee outside LUMC  € 950,00

Bachelor/master students of the Leiden University are free of charge**