Multivariable analysis is an essential tool in modern epidemiology. Particularly observational research and research about aetiology is multivariable by nature. To critically appraise the published literature and to engage in epidemiological research, a basic understanding of multivariable models and the underlying assumptions is required.
The aim of this course component is to enable participants to appraise and apply four types of multivariable analysis, i.e. linear, logistic, Poisson and Cox regression, in the domain of public health research.
This is an advanced course component: it should allow a deeper understanding of basic concepts of epidemiology and statistics, such as bias, confounding, effect modification, stratification, significance, and study design.
This course is relevant for:
- MPH students planning to use multivariable analysis methods for their thesis work
- All students preparing for a research career
- All students planning to use epidemiological literature in a critical way
- PhD candidates planning to use multivariable analysis methods in their research projects
The Multivariable Analysis (MVA) course employs a variety of teaching and learning methods to enhance your understanding and application of concepts and methods. Interactive lectures serve as a platform for introducing theoretical aspects of multivariable analysis. You will have the opportunity to engage in group exercises where you can apply the learned theory to practical scenarios, fostering a deeper comprehension of the subject matter. Quizzes will be conducted to assess your progress and reinforce their knowledge.
A key component of the course involves the analysis of datasets using the R statistical software which will allow you to gain hands-on experience in applying multivariable analysis techniques to real data. Additionally, you will critically read scientific papers that report on multivariable analysis methods, further enhancing your understanding and critical thinking skills.
Examples and exercises are selected and updated based on the expertise and research projects of the staff of the Department of Public Health at ITM.
After completion of the course the student should be able to:
- Understand four multivariable analysis methods, i.e. multiple linear regression, logistic regression, Poisson and Cox regression (survival analysis)
- Identify the measurement scale of outcome variables
- Justify when it is appropriate to use these methods
- Identify and explain confounding and effect modification
- Translate a research question into a multivariable regression model
- Recognise different approaches to model building and make a case for a specific approach
- Write, interpret and apply simple linear regression equations
- Perform multiple linear regression, logistic regression, Poisson regression and survival analysis using statistical software
- Appraise research papers that use multivariable analysis methods