Multivariable statistical methods 

Properly addressing research questions in medical science often requires accounting for several variables (simultaneously). The aim of this module is to familiarize students with the most commonly used statistical techniques for multi-variable analysis. The content of this module builds upon the uni- and bivariate techniques of the statistics teaching in the bachelor curriculum. A particular application of the techniques in this module is the analysis and interpretation of epidemiological observational studies. Recently, Linear Mixed Models have been incorporated in the module for the analysis of longitudinal data. This module is compulsory for students with a major in Epidemiology and advisable for students that are going to work with large datasets (such as OEH and HTA-students). Students that will rather deal with small-scaled experiments (such as PB- and TOX-students) are advised to take the module Design and Analysis of Small-scaled Experiments (5AM07).

Upon completing this module the student should be able to:

1. Choose an appropriate multivariable statistical technique to analyse data gathered in the context of medical scientific research.

2. Perform multivariable regression analyses with the aid of the computer package SAS and interpret the results and report the conclusions.

3. Describe roughly the statistical background of some multivariable analysis techniques that are frequently applied in medical scientific research.

Knowledge of the module 5E002 is beneficial, but not essential.

Linear regression models including analysis of (co)variance and linear mixed models; (conditional) logistic regression; logistic discriminant-analysis; Poisson regression; matrix notation; least-squares estimation; (restricted) maximum likelihood estimation; case-control study; cohort study.

Applied Regression Analysis and Multivariable Methods. Kleinbaum DG,

Kupper LL, Muller KE. Nizam A.

4rd Edition. Brooks/Cole Publishing Company, Pacific Grove, California

(order at least 6 weeks in advance, avalable in cheap students edition!).

Students gain theoretical knowledge of multi-variable techniques via self-educational assignments, while the statistical package SAS is used for the practical training of those techniques. Discussion and interpretation of the analysis results is done in group assignments.

One of the research questions has to be elaborated to a concept research paper and a written examination concludes the course.

Maximal number of participants: 30.

This module is compulsory for students with a major in Epidemiology.