DS818: General Regression Models (Survival and Longitudinal Data)

• Have knowledge of basic mathematics at the high school level.
• Have knowledge of statistical methods coresponding to the contents of the course Basale Biostatistiske Begreber og Regression

Knowledge corresponding to Introduktion til Basale Biostatistiske Begreber og Regression is mandatory. Knowledge about Stata is preferable. Students should bring their own laptop.

• Improve competence to identify basic designs and relevant statistical analysis methods.
• Improve skills in statistical analyses, including model adaptation and model criticism.
• Provide knowledge of risk assessment when dealing with exposure and covariates.

• Translate research objectives into clear, testable statistical hypotheses.
  
• To understanding the necessity to correct for possible influencing factors besides exposure in observational or experimental studies.
• To identify a statistical analysis plan based on the substantive research problem and the data collected, and to apply the statistical methods, and acknowledge the limitations of those methods.
  
• To perform statistical analyses using logistics, ordinal and Poisson regression, as well as preform analyses related to longitudinal data and survival data. This involves effect-estimation, description of the estimation and its uncertainty,  as well as hypothesis testing and model evaluation.
• Understand and interpret the relative risks, odds ratios, and hazard ratios when comparing two populations.
• Understand why longitudinal studies require special methods when analysing longitudinal data.
  
• Understand why survival (time to event) data requires its own type of analysis techniques.
  
• Evaluate output containing statistical procedures and graphics and interpret the statistical analysis results in a the public health context.

The following main topics are contained in the course:

• The student is introduced to a number of statistical techniques applied to medical and health related examples. Topics covered are especially Poisson regression, analysis of longitudinal data and survival analysis. Other topics can be e.g. quantile regression and missing data.
• The statistical software package Stata is used for the analyses when examples are studied. Students get hands-on experience with Stata during the exercises.
• The course includes a project period where the student brings a dataset to be analyzed with tools covered in the course; suitable datasets should be of moderate size and moderately complicated from a statistical point of view (allowing the application of techniques from the course). The project period closes with a written project report and an oral project representation by each participant focusing on the dataset and the analyses.