MATH6033 Epidemiological Methods
Epidemiology is the study of patterns and causes of disease in populations. This course introduces different methods of carrying out, and analysing, epidemiological studies. Pertinent issues such as appropriate study design, data quality, analysis, and interpretation and presentation of results, will be discussed. A computing tutorial will be used to introduce the statistical package Stata. A reference list is provided – this is intended to be a resource for further reading and future use rather than a compulsory reading list.
Aims and Objectives
There are no formal prerequisites but the module draws on some of the material introduced in the first semester core modules, and concurrent modules in the second semester.
Having successfully completed this module you will be able to:
- Students should be able to distinguish between different types of study design
- Students should be able to describe the Bradford-Hill criteria for assessing causality
- Students should be aware of the effects that chance, confounding and bias may have on epidemiological studies
- Students should be able to anticipate particular problems that may arise in epidemiological data
- Students should be able to apply appropriate statistical analysis techniques to analyse epidemiological data
Lecture 1 – Introduction to Epidemiology Definition of epidemiology Basic epidemiological concepts Epidemiological study design Establishing causality Lecture 2 – Bias, Confounding and Interactions Types of bias The concept of confounding Interactions in epidemiological research Lecture 3 – Ratio Measures of Association Defining risks Prevalence and incidence Measures of association When not to use an odds ratio Lecture 4 – Cohort Studies 1 Design issues Follow-up of subjects Official Statistics Analysis using the person-years approach Lexis diagrams Standardised Mortality Ratios (SMRs) Lecture 5 – Cohort Studies 2 Poisson generalised linear models to test for trend in SMRs Cox Proportional Hazards model Lecture 6 – Case-control studies 1 Design of case-control studies Design issues Choice of cases and controls Matching Collection of exposure information Basic analysis of unmatched studies Test for homogeneity of odds ratios Mantel-Haenszel methods for stratified analysis 2xk tables Test for trend Basic analysis of matched studies Lecture 7 – Case-control studies 2 Logistic regression - unconditional and conditional Example of analysis using Stata Lecture 8 – Design of Epidemiological Studies A practical session to design a study to assess whether use of mobile phones increases the risk of brain cancer. Lecture 9 – Ecological Studies and Meta-analysis Design of ecological studies Main types of ecological comparison Analysis of ecological studies by mapping, correlation and clustering Definition of systematic review and meta-analysis Design of a systematic review Conducting a meta-analysis Fixed effects, random-effects and forest plots Publication bias Heterogeneity Lecture 10 – Transforming Data Why transform data? Assessing normality Commonly used transformations and interpretation Fisher-Yates normal scores Box-Cox transformation Lecture 11 – Hints and Tips Data cleaning Recording analyses Missing data Categorisation of continuous variables Subgroup analyses Significance testing Presentation of results Lecture 12 – Case Study A practical session discussing an investigation into a cluster of leukaemias near the Sellafield nuclear site in West Cumbria
Learning and Teaching
|Total study time||75|
Repeat type: Internal & External