Abteilung Biostatistik

Survival Analysis

Course information

Class in biostatistics master program, Fall semester 2011

Instructor: Jan Beyersmann, Abteilung Medizinische Biometrie und Statistik, Institut für Medizinische Biometrie und Medizinische Informatik, Universitätsklinikum Freiburg


Time and venue

Institute for Social and Preventive Medicine, Division of Biostatistics, Hirschengraben 84, Room F05.


Program
The main topic of the lecture “Survival Analysis” is the analysis of event time data. Medical examples are survival times of patients or so- called “disease free” survival. Observation of such data is typically incomplete. For example, only the minimal survival time is known for patients who still live at the time of the statistical analysis (right censoring). The methods of survival analysis are therefore based on the hazard rate, an instantaneous rate of events. The common nonparametric methods for the estimation of the cumulative hazard (Nelson-Aalen), for the comparison of hazard rates (logrank tests) and for the modeling of the hazard rate depending on explaining variables (Cox model) will be discussed. The estimators of survival probabilities (e.g. Kaplan-Meier) are non-random functions of the hazard estimators. An outlook on the generalisation of the methods to varying entry times (left truncation) and more complex event structures (competing risks, multistate models) will be given at the end of the course.

Prerequisites: Good knowledge of the basic principles of probability theory and statistics, experience with the statistical programming language R.

Literature: J. Klein, M. Moeschberger: Survival Analysis. 2nd Edition. Springer, 2003

Exam

The exam takes place on January 16, 2012, 13:00-15:00 in Y27H25. Allowed support materials are lecture notes and a non-programmable pocket calculator.

Lecture notes, exercises and code

Download page (closed 08/02/2012, contact Eva Furrer if you need a document)

top
Mobile Ansicht | Klassische Ansicht