Introduction to modelling in edpidemiology

Plant disease epidemics involve changes in disease intensity over time and space in a host population. If the entire host population of interest is very small, it may be possible to actually measure or observe every individual (e.g., plant) for disease intensity (and other variables of interest). However, even having such a complete collection of data for disease intensity does not automatically lead to any insight into epidemic processes or help in communicating results to others. In general, one uses models (as defined below) to summarize the essential features of the observations or measurements of interest. With models, the many individual observations are reduced to a few model terms, making it much easier to visualize and ultimately to understand the phenomena being studied, as well as to communicate that understanding to others. In fact, the populations that epidemiologists are interested in are seldom very small. For instance, even when epidemics are studied in small experimental plots, where every plant (or leaf, fruit, etc.) is actually observed, the host population of interest is generally much larger, such as all plants of the same crop cultivar in commercial fields, grown under conditions similar to those studied in the small experimental plots. In other words, researchers usually want to apply results obtained with small data sets to larger populations that are not observed. Thus, models are needed here to allow inferences to be made about these large populations based on more limited direct information.

Chapter topics include:

Introduction

Models

Methods of Model Development

Fitting of Linear Models to Data

Fitting of Nonlinear Models to Data

Applications

Maximum Likelihood