You are here:

# Advanced statistical methods in veterinary epidemiology

## VPM013

This module will provide an introduction to advanced methods of statistical modelling of epidemiological data.

### Prerequisites

To take this course, you must have passed 'Statistical Methods in Veterinary Epidemiology'. You will use some mathematical formulae, but no knowledge of calculus is required.

You also need access to Microsoft Office and ArcGIS (version 8.0 or later) plus the extensions Spatial Analyst and 3D Analyst. Software costs are not included in your course fee. To purchase software, contact the Course Administrator (UK only) or ESRI (select country of residence).

### Topics covered

Section 1: Investigation of Spatial Patterns of Animal Disease

In this module (Units 1-4), you learn about geographical information systems, based on the ESRI ArcView software. Computer-based activities will help you to produce maps from spatial data. You will also learn how to undertake exploratory analyses and statistical modelling of spatial data, using DynESDA and the ESRI Spatial Analyst extension tools.

Section 2: Advanced Methods of Statistical Analysis, Part I

The second module (Units 5-8) is developed as Computer-Assisted Learning (CAL) material. The first unit demonstrates how to review and summarise the framework and process of generalised linear regression models. Units 6 and 7 are about classical analysis of matched case–control studies. In Unit 8, different sampling schemes for case–control studies and the exposure odds ratio estimates are described.

Section 3: Advanced Methods of Statistical Analysis, Part II

The final module (Units 9-14) builds on the advanced statistical methods from Module 2. Units 9-11 explore the Poisson distribution, issues with Cox regression, and strategies for obtaining adequate regression models. Units 12-14 explore how you conduct analyses of data based on correlated observations, statistical approaches to meta-analysis and systematic review, and attributable risk.

### Learning outcomes

If you complete the module successfully, you should be able to:

• understand attributes of geographic data, geo-referencing systems, and the importance of spatial autocorrelation in spatial data analysis.
• use GIS packages to display a map, plot point data, and create a kernel density surface.
• describe the structure of a fixed-effects model of disease count data and interpret the regression coefficients from a mixed-effects model.
• select, apply and interpret the results of regression methods for the analysis of case–control and cohort studies, using appropriate software.
• plan a strategy of analysis for an epidemiological data-set.
• describe the effects of correlated data on epidemiological analysis, and the use of statistical methods which take account of such correlations.
• review and summarise information from many studies using meta-analysis.
• estimate the risk attributable to an exposure in a population.

### Assessment

This module is assessed by:

• A three-hour unseen written exam (80%)
• A written assignment (20%)