Skip to main content
Unit of study_

PUBH5217: Biostatistics: Statistical Modelling

In this unit, you will learn how to analyse health data using statistical models. In particular, how to fit and interpret the results of different statistical models which are commonly used in medicine and health research: linear models, logistic models, and survival models. This unit is ideal for those who wish to further develop their research skills and/or improve their literacy in reading and critiquing journal articles in medicine and health. The focus of the unit is very applied and not mathematical. Students gain hands on experience in fitting statistical models in real data. You will learn how to clean data, build an appropriate model, and interpret results. This unit serves as a prerequisite for PUBH5218 Advanced Statistical Modelling.

Code PUBH5217
Academic unit Public Health
Credit points 6
PUBH5018 or FMHU5002
(PUBH5211 or PUBH5212 or PUBH5213)

At the completion of this unit, you should be able to:

  • LO1. read a data file into statistical software
  • LO2. manipulate and edit data set in statistical software
  • LO3. conduct appropriate exploratory data analysis when the outcome variable is continuous, binary, or time-to-event using statistical software
  • LO4. fit a multiple linear regression model using statistical software, where the explanatory variables may be continuous and/or categorical
  • LO5. interpret the results from a multiple linear regression model
  • LO6. identify the difference between a potential confounder and actual confounder
  • LO7. conduct appropriate model building strategies for building a regression model
  • LO8. use statistical software to conduct appropriate model checking for linear regression models
  • LO9. use logistic regression to assess the association between a binary outcome and multiple covariates
  • LO10. use Poisson regression to assess the association between a count outcome and multiple covariates
  • LO11. write statistical software to analyse categorical data, and correctly interpret the produced output
  • LO12. identify when it is appropriate to use survival analysis methods, including the logrank method and Cox proportional hazards model
  • LO13. define censoring, the survivor function, and the hazard function
  • LO14. produce survival curves in a plot using statistical software and correctly interpret such a plot
  • LO15. perform logrank tests and fit Cox proportional hazards models to analyse survival data
  • LO16. assess the proportional hazards assumption, and describe what to do if this assumption does not hold
  • LO17. write statistical software to analyse survival data, and correctly interpret the produced output.