We have access to more data than ever before, but how do we find the signal in the noise? And how much trust should we put in the answers we find? We can answer these types of questions with statistical models. This course will introduce you to a family of statistical models called regression models, powerful tools for answering real-world health questions. For example, we can use clinical data to investigate how body composition correlates with lung function in children with cystic fibrosis. Or calculate a person’s likelihood of developing disease from their risk factors. If your doctor has ever calculated your cardiovascular disease risk, that was based on a regression equation.

In this course you will learn that different types of data require different types of models. You will also learn how to critically assess the models you build and how to interpret and communicate the results. We will show how code can be used to create reproducible analysis. The course will cover some theory, with a strong focus on practical application of the methods using statistical software.

Mode of study

External (Distance) and Internal (Face-to-Face) classes on campus

Key contacts

Katrina Blazek
Course Convenor
k.blazek@unsw.edu.au

A/Prof Timothy Dobbins
Course Convenor
+61 (2) 9385 3379
t.dobbins@unsw.edu.au

Who should do this course?

A prerequisite to enrol in this course is successful completion of PHCM9795 Foundations of Biostatistics or PHCM9498 Epidemiology and Statistics in Public Health. Students who have achieved a credit in the biostatistics components of these courses will be well prepared to explore the concepts covered in the Advanced Epidemiology course.

Course outcomes

Upon successful completion of this course you will be able to:

  • Identify an appropriate regression model for your research study

  • Implement statistical analysis using regression models on complex datasets with different types of variables

  • Identify and account for confounding and effect modification in epidemiological studies 

  • Interpret research findings and draw valid conclusions addressing the research question

  • Critically evaluate statistical analyses and present findings at a standard that is sufficient for submission to scientific journals

  • Produce analysis scripts using statistical software to foster reproducible and transparent research

Learning & teaching

The course focuses on developing practical experience that will assist your understanding and application of statistical techniques and in using statistical software. The focus is to provide you with the capacity to think critically about epidemiological questions and the use of regression methods to address questions in medical and public health research.

The course comprises of lectures and tutorials. In addition to the lecture materials, extra learning materials may also be posted on Moodle. The course covers common analytical techniques used in public health and medical research. These include:

  • simple and multiple linear regression for continuous outcome variables
  • logistic and log binomial regression for binary outcome variables
  • survival analysis for time-to-event data.

Students will have the choice of using Stata or R statistical software to complete this course, and resources will be provided for both packages.

Assessments

1. Assessment 1: Take-home test 

Weighting 20%

2. Assessment 2: Report

Weighting 40%

3. Assessment 3: Report

Weighting 40%

Readings & resources 

Learning resources for this course consist of:

  • course notes and recommended readings from journal articles and textbooks
  • links to all the journal articles and available ebooks from the recommended texts, embedded on Moodle