Ebook Veterinary epidemiology (4/E): Part 1

Veterinary Epidemiology


Veterinary Epidemiology
Fourth Edition

Michael Thrusfield
Veterinary Clinical Sciences
Royal (Dick) School of Veterinary Studies
University of Edinburgh

With Robert Christley
Epidemiology & Population Health
Institute of Infection & Global Health, and
Institute of Veterinary Science
University of Liverpool

And Helen Brown, Peter J. Diggle, Nigel French, Keith Howe, Louise Kelly, Annette O’Connor,
Jan Sargeant and Hannah Wood


This edition first published 2018 © 2018 by John Wiley & Sons Ltd
Edition History
First edition © 1986 by Butterworth & Co. (Publishers) Ltd
Second edition © 1995 by Blackwell Science Ltd
Third edition © 2005, 2007 by Blackwell Science Ltd, a Blackwell Publishing company
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse
material from this title is available at />The right of Michael Thrusfield and Robert Christley to be identified as the authors of the editorial material in this work has been asserted in
accordance with law.
Registered Offices

John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Office
9600 Garsington Road, Oxford, OX4 2DQ, UK
For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears in standard print versions
of this book may not be available in other formats.
Limit of Liability/Disclaimer of Warranty
The contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and
should not be relied upon as recommending or promoting scientific method, diagnosis, or treatment by physicians for any particular patient. In
view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the
use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or
instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for
added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no
representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties,
including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or
extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or
product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors
endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with
the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be
suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this
work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be
liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
Library of Congress Cataloging-in-Publication Data
Names: Thrusfield, M. V., author. | Christley, Robert, 1968- author.
Title: Veterinary epidemiology / by Michael Thrusfield, Veterinary Clinical
Sciences, Royal (Dick) School of Veterinary Studies, University of
Edinburgh ; with Robert Christley and [8 others].
Description: Fourth edition. | Hoboken, NJ : Wiley, 2018. | Includes
bibliographical references and index. |

Identifiers: LCCN 2017051658 (print) | LCCN 2017053201 (ebook) | ISBN
9781118280263 (pdf) | ISBN 9781118280270 (epub) | ISBN 9781118280287
(paperback)
Subjects: LCSH: Veterinary epidemiology. | MESH: Epidemiologic
Methods–veterinary
Classification: LCC SF780.9 (ebook) | LCC SF780.9 .T48 2018 (print) | NLM SF
780.9 | DDC 636.089/44–dc23
LC record available at />Cover images: (Top: from left to right) © holbox/Shutterstock; © Paul Looyen/Shutterstock; © Lorado/Gettyimages;
© Lisa Van Dyke/Gettyimages; (Map) © yukipon/Gettyimages; (Bottom: from left to right) © Palenque/Gettyimages;
© l i g h t p o e t/Shutterstock; © Seiji/Shutterstock; © claire norman/Shutterstock
Cover design by Wiley
Set in 10/12 pt Warnock by SPi Global, Pondicherry, India
10 9 8 7 6 5 4 3 2 1


In memory of George


vii

Contents
Contributors xviii
From the preface to the first edition xix
From the preface to the second edition xx
From the preface to the third edition xxi
Preface to the fourth edition xxii
About the companion website xxiv
1

The development of veterinary medicine

Michael Thrusfield

1

Historical perspective 1
Domestication of animals and early methods of healing
Changing concepts of the cause of disease 2
Impetus for change 5
Quantification in medicine 10
Contemporary veterinary medicine 12
Current perspectives 12
The fifth period 19
Recent trends 20
Further reading 25
2

The scope of epidemiology
Michael Thrusfield

1

28

Definition of epidemiology 28
The uses of epidemiology 29
Types of epidemiological investigation 32
Epidemiological subdisciplines 33
Components of epidemiology 35
Qualitative investigations 35
Quantitative investigations 36

Epidemiology’s locale 39
The interplay between epidemiology and other sciences 39
The relationship between epidemiology and other diagnostic disciplines
Epidemiology within the veterinary profession 40
Further reading 41
3

Causality 42
Michael Thrusfield

Philosophical background 42
Causal inference 43
Methods of acceptance of hypotheses

44

40


viii

Contents

Koch’s postulates 45
Evans’ rules 45
Variables 46
Types of association 46
Non-statistical association 46
Statistical association 46
Confounding 49

Causal models 50
Formulating a causal hypothesis 53
Methods of deriving a hypothesis 53
Principles for establishing cause: Hill’s criteria
Further reading 56
4

Describing disease occurrence
Michael Thrusfield

55

58

Some basic terms 58
Basic concepts of disease quantification 61
The structure of animal populations 62
Contiguous populations 62
Separated populations 65
Measures of disease occurrence 67
Prevalence 67
Incidence 67
The relationship between prevalence and incidence rate 70
Application of prevalence and incidence values 72
Mortality 72
Survival 73
Example of calculation of prevalence, incidence, mortality, case fatality and survival
Ratios, proportions and rates 76
Mapping 80
Geographic base maps 80

Further reading 84
5

Determinants of disease
Michael Thrusfield

86

Classification of determinants 86
Host determinants 89
Genotype 89
Age 90
Sex 91
Species and breed 92
Behaviour 93
Other host determinants 93
Agent determinants 94
Virulence and pathogenicity 94
Gradient of infection 97
Outcome of infection 98
Microbial colonization of hosts 100
Environmental determinants 101
Location 101
Climate 101
Husbandry 104

75


Contents

Stress 105
Interaction 106
Biological interaction 108
Statistical interaction 109
The cause of cancer 110
Further reading 112
6

The transmission and maintenance of infection
Michael Thrusfield

115

Horizontal transmission 115
Types of host and vector 115
Factors associated with the spread of infection 118
Routes of infection 121
Methods of transmission 123
Long-distance transmission of infection 125
Vertical transmission 129
Types and methods of vertical transmission 129
Immunological status and vertical transmission 129
Transovarial and trans-stadial transmission in arthropods
Maintenance of infection 131
Hazards to infectious agents 131
Maintenance strategies 132
Transboundary diseases 135
Further reading 136
7

The ecology of disease
Michael Thrusfield

130

138

Basic ecological concepts 139
The distribution of populations 139
Regulation of population size 142
The niche 148
Some examples of niches relating to disease 150
The relationships between different types of animals and plants 152
Ecosystems 155
Types of ecosystem 156
Landscape epidemiology 158
Nidality 159
Objectives of landscape epidemiology 161
Landscape characteristics determining disease distribution 164
Further reading 165
8

Patterns of disease
Michael Thrusfield

168

Epidemic curves 168
Kendall’s Threshold Theorem 168

Basic reproductive number (R0) 169
Dissemination rate 172
Common-source and propagating epidemics 172
The Reed–Frost model 173
Kendall’s waves 175
Trends in the temporal distribution of disease 177
Short-term trends 177

ix


x

Contents

Cyclical trends 178
Long-term (secular) trends 179
True and false changes in morbidity and mortality 180
Detecting temporal trends: time series analysis 180
Trends in the spatial and temporal distribution of disease
Spatial trends in disease occurrence 186
Space–time clustering 186
Further reading 187
9

Comparative epidemiology
Michael Thrusfield

189

Types of biological model 189
Cancer 191
Monitoring environmental carcinogens 191
Identifying causes 192
Comparing ages 193
Some other diseases 196
Diseases with a major genetic component 196
Some non-infectious diseases 197
Diseases associated with environmental pollution
Reasoning in comparative studies 199
Further reading 199
10

The nature of data
Michael Thrusfield

198

201

Classification of data 201
Scales (levels) of measurement 201
Composite measurement scales 204
Data elements 205
Nomenclature and classification of disease 205
Diagnostic criteria 207
Sensitivity and specificity 208
Accuracy, refinement, precision, reliability and validity
Bias 210
Representation of data: coding 210

Code structure 211
Numeric codes 212
Alpha codes 213
Alphanumeric codes 214
Symbols 215
Choosing a code 215
Error detection 216
Further reading 217
11

186

Data collection and management
Michael Thrusfield

219

Data collection 219
Questionnaires 219
Quality control of data 228
Data storage 229
Database models 229
Non-computerized recording techniques

231

209


Contents

Computerized recording techniques 232
Veterinary recording schemes 232
Scales of recording 232
Veterinary information systems 234
Some examples of veterinary databases and information systems
Geographical information systems 244
Further reading 248
12

Presenting numerical data 251
Michael Thrusfield and Robert Christley

Some basic definitions 251
Some descriptive statistics 252
Measures of position 253
Measures of spread 254
Statistical distributions 254
The Normal distribution 254
The binomial distribution 255
The Poisson distribution 255
Other distributions 256
Transformations 256
Normal approximations to the binomial and Poisson distributions
Estimation of confidence intervals 257
The mean 257
The median 258
A proportion 258
The Poisson distribution 259
Some epidemiological parameters 260

Other parameters 261
Bootstrap estimates 261
Displaying numerical data 262
Displaying qualitative data 262
Displaying quantitative data 263
Monitoring performance: control charts 266
Further reading 269
13

Surveys 270
Michael Thrusfield and Helen Brown

Sampling: some basic concepts 270
Types of sampling 272
Non-probability sampling methods 272
Probability sampling methods 272
What sample size should be selected? 275
Estimation of disease prevalence 275
Detecting the presence of disease 284
The cost of surveys 290
Calculation of confidence intervals 290
Further reading 294
14

237

Demonstrating association
Michael Thrusfield

296

Some basic principles 296
The principle of a significance test
The null hypothesis 297

296

257

xi


xii

Contents

Errors of inference 297
Multiple significance testing 298
One- and two-tailed tests 298
Independent and related samples 299
Parametric and non-parametric techniques 299
Hypothesis testing versus estimation 300
Sample-size determination 300
Statistical versus clinical (biological) significance 300
Interval and ratio data: comparing means 302
Hypothesis testing 302
Calculation of confidence intervals 303
What sample size should be selected? 304
Ordinal data: comparing medians 304
Hypothesis testing 304

Calculation of confidence intervals 308
What sample size should be selected? 309
Nominal data: comparing proportions 309
Hypothesis testing 310
Calculation of confidence intervals 313
What sample size should be selected? 314
χ 2 test for trend 314
Correlation 316
Multivariate analysis 317
Statistical packages 318
Further reading 318
15

Observational studies
Michael Thrusfield

319

Types of observational study 319
Cohort, case-control and cross-sectional studies
Measures of association 321
Relative risk 321
Odds ratio 323
Attributable risk 325
Attributable proportion 327
Interaction 328
The additive model 328
Bias 330
Controlling bias 332
What sample size should be selected? 335

Calculating the power of a study 336
Calculating upper confidence limits 337
Further reading 338
16

Design considerations for observational studies
Robert Christley and Nigel French

319

339

Descriptive observational studies 339
Analytical observational studies 340
Design of cohort studies 340
Design of case-control studies 346
Design of cross-sectional analytical studies 352
Overview of other study designs 354
Further reading 359


Contents

17

Clinical trials 361
Michael Thrusfield

Definition of a clinical trial 361
Design, conduct and analysis 364

The trial protocol 364
The primary hypothesis 364
The experimental unit 367
The experimental population 368
Admission and exclusion criteria 368
Blinding 369
Randomization 369
Trial designs 370
What sample size should be selected? 372
Losses to follow-up 373
Compliance 373
Terminating a trial 374
Interpretation of results 374
Meta-analysis 375
Goals of meta-analysis 376
Components of meta-analysis 377
Sources of data 377
Data analysis 378
Further reading 380
18

Validity in epidemiological studies
Robert Christley and Nigel French

383

Types of epidemiological error 383
Accuracy, precision and validity in epidemiological studies 384
Background factors 385
Interpretation bias 385

Selection bias 386
Examples of selection biases 387
Information bias 390
Examples of information biases 390
Statistical interaction and effect-measure modification 392
Confounding 392
Criteria for confounding 393
Confounding and causal diagrams 394
Controlling confounding 394
Errors in analysis 395
Communication bias 395
Further reading 396
19

Systematic reviews 397
Annette O’Connor, Jan Sargeant and Hannah Wood

Evidence synthesis 397
Overview of systematic reviews 397
Differences between systematic reviews and narrative reviews 398
Questions that are suitable for systematic reviews 398
Types of review questions suitable for systematic reviews 399
Extensive search of the literature 399
Assessment of risk of bias in a systematic review 400

xiii


xiv

Contents

Steps of a systematic review 400
Step 1: Define the review question and the approach to conduct of the review (i.e., create a protocol)
Step 2: Comprehensive search for studies 403
Step 3: Select relevant studies from the search results 406
Step 4: Collect data from relevant studies 407
Step 5: Assess the risk of bias in relevant studies 409
Step 6: Synthesize the results 412
Step 7: Presenting the results 416
Step 8: Interpret the results and discussion 419
Further reading 419
20

Diagnostic testing
Michael Thrusfield

421

Serological epidemiology 421
Assaying antibodies 421
Methods of expressing amounts of antibody 421
Quantal assay 423
Serological estimations and comparisons in populations 424
Antibody prevalence 424
Rate of seroconversion 425
Comparison of antibody levels 426
Interpreting serological tests 427
Refinement 427
Accuracy 429

Evaluation and interpretation of diagnostic tests 430
Sensitivity and specificity 430
Youden’s index 433
Diagnostic odds ratio 434
Predictive value 434
Likelihood ratios 436
ROC curves 441
Aggregate-level testing 443
Multiple testing 444
Diagnostic tests in import risk assessment 446
Guidelines for validating diagnostic tests 447
Validating diagnostic tests when there is no gold standard 448
Agreement between tests 450
Practical application of diagnostic tests 456
Further reading 456
21

Surveillance 457
Michael Thrusfield

Some basic definitions and principles 457
Definition of surveillance 457
Goals of surveillance 458
Types of surveillance 459
Some general considerations 461
Sources of data 464
Mechanisms of surveillance 471
Surveillance networks 475
Surveillance in less-economically-developed countries: participatory epidemiology
Principles of participatory epidemiology 477

Techniques of data collection 478
Strengths and weaknesses of participatory epidemiology 481

475

402


Contents

Some examples of participatory epidemiology 483
Companion-animal surveillance 483
Wildlife surveillance 485
Aquatic-animal surveillance 485
Assessing the performance of surveillance systems 486
Improving the performance of surveillance: risk-based surveillance
Further reading 488
22

Statistical modelling 492
Robert Christley and Peter J. Diggle

Simple linear regression models 492
Key assumptions of linear regression models 495
Modelling more than one input variable 499
Handling categorical input variables 500
Non-linear modelling of quantitative input variables 502
Additive models 502
Categorization of the input variable 502
Transformation of the input and/or output variable 504

Piece-wise regression 504
Modelling interactions 505
Model selection 506
Modelling binary outcomes 509
Generalized linear models 511
The multiple logistic regression model 511
Model selection for logistic regression models 512
Diagnostic checking of logistic regression models 513
Generalized additive models 514
Modelling clustered data 514
Further reading 519
23

Mathematical modelling
Michael Thrusfield

520

Types of model 521
Modelling approaches 521
Deterministic differential calculus modelling 521
Stochastic differential calculus modelling 525
Empirical simulation modelling 526
Process simulation modelling 527
Monte Carlo simulation modelling 528
Matrix population modelling 530
Network population modelling 532
Contact-network modelling 533
Systems modelling 534
The rational basis of modelling for active disease control

Available knowledge, and the functions of models 534
From theory to fact 535
Model building 536
Further reading 538
24

Risk analysis 540
Michael Thrusfield and Louise Kelly

Definition of risk 540
Risk analysis and the ‘precautionary principle’ 543
Risk analysis in veterinary medicine 543

534

486

xv


xvi

Contents

Components of risk analysis 545
Hazard identification 546
Risk assessment 546
Risk management 548
Risk communication 551
Qualitative or quantitative assessment? 551

Semi-quantitative risk assessment 551
Qualitative risk analysis 552
Framework for qualitative risk assessment 552
Qualitative risk assessment during epidemics 554
Quantitative risk analysis 556
Framework for quantitative risk assessment 556
What level of risk is acceptable? 560
Further reading 563
25

Economics and veterinary epidemiology
Keith Howe and Michael Thrusfield

565

General economic concepts 565
Production functions 565
Disease and animal production functions 566
Value and money 567
Money and prices 567
Opportunity cost 568
Technical and economic efficiency 568
Positive and normative economics 569
Levels of aggregation 569
Disease contained at farm level 569
Disease not contained at farm level 570
Zoonotic disease 570
Disease at international level 571
Evaluating disease-control policies 575
Components of disease costs 576

Optimum control strategies 577
Partial budgets 579
Social cost–benefit analysis (CBA) 579
Summary of methods 582
Further study 582
Further reading 584
26

Health schemes 586
Michael Thrusfield

Private health and productivity schemes 586
Structure of private health and productivity schemes
Dairy health and productivity schemes 588
Pig health and productivity schemes 591
Sheep health and productivity schemes 592
Beef health and productivity schemes 594
National schemes 597
Accredited/attested herds 597
Health schemes 598
Companion-animal schemes 599
Further reading 603

586


Contents

27

The control and eradication of disease 604
Michael Thrusfield

Definition of ‘control’ and ‘eradication’ 604
Strategies of control and eradication 605
Important factors in control and eradication programmes 616
Outbreak investigation 623
Cause known: foot-and-mouth disease 623
Cause unknown: chronic copper poisoning 625
The epidemiological approach to investigation of outbreaks 626
Veterinary medicine in the 21st century 628
Livestock medicine 628
Companion-animal medicine 629
Further reading 630
General reading 633
Appendices 635
Appendix I: Glossary of terms 636
Appendix II: Basic mathematical notation and terms 641
Appendix III: Some computer software 643
Appendix IV: Veterinary epidemiology on the Internet 648
Appendix V: Student’s t-distribution 650
Appendix VI: Multipliers used in the construction of confidence intervals based on the Normal distribution,
for selected levels of confidence 651
Appendix VII: Values of exact 95% confidence limits for proportions 652
Appendix VIII: Values from the Poisson distribution for calculating 90%, 95% and 99% confidence intervals
for observed numbers from 0 to 100 658
Appendix IX: The χ2 distribution 660
Appendix X: Technique for selecting a simple random sample 661
Appendix XI: Sample sizes 663
Appendix XII: The probability of detecting a small number of cases in a population 669

Appendix XIII: The probability of failure to detect cases in a population 671
Appendix XIV: Sample sizes required for detecting disease with probability, p1, and threshold number of
positives 672
Appendix XV: Probabilities associated with the upper tail of the Normal distribution 676
Appendix XVI: Lower- and upper-tail probabilities for Wx, the Wilcoxon–Mann–Whitney rank-sum
statistic 678
Appendix XVII: Critical values of T + for the Wilcoxon signed ranks test 683
Appendix XVIII: Values of K for calculating 95% confidence intervals for the difference between
population medians for two independent samples 685
Appendix XIX: Values of K∗ for calculating 95% confidence intervals for the difference between
population medians for two related samples 688
Appendix XX: Common logarithms (log10) of factorials of the integers 1–999 689
Appendix XXI: The correlation coefficient 691
Appendix XXII: The variance-ratio (F) distribution 692
References 694
Index 841

xvii


xviii

Contributors

Michael Thrusfield
Veterinary Clinical Sciences
Royal (Dick) School of Veterinary
Studies
University of Edinburgh
Edinburgh, UK

Robert Christley
Institute of Infection and Global Health
Institute of Veterinary Science
University of Liverpool
Liverpool, UK
Helen Brown
The Roslin Institute
University of Edinburgh
Edinburgh, UK
Peter J. Diggle
CHICAS
Lancaster Medical School
Lancaster University
Lancaster, UK
Nigel French
Institute of Veterinary Animal and
Biomedical Sciences
Massey University
Palmerston North, New Zealand
Keith Howe
Centre for Rural Policy Research
University of Exeter
Exeter, UK

Louise Kelly
University of Strathclyde
Glasgow, UK
Annette O’Connor
College of Veterinary Medicine
Iowa State University

Ames, Iowa, USA
Jan Sargeant
Ontario Veterinary College
University of Guelph
Guelph, Canada
Hannah Wood
York Health Economics Consortium
York, UK


xix

From the preface to the first edition
The common aim of the many disciplines that comprise veterinary medicine is an increase in the health
of animal populations, notably of domestic livestock
and companion animals. This goal has traditionally
been achieved by individual diagnosis and treatment:
procedures that evolved contemporaneously in veterinary and human medicine, when infectious diseases,
which had predominantly single causes and clearly
identifiable signs, were commonplace.
Four major changes in the veterinarian’s appreciation of and approach to disease problems have
occurred over the past 20 years. First, despite traditional control techniques, for example slaughter and
vaccination, some diseases remain at refractory levels
and now require continuous scrutiny to detect changing levels of occurrence associated with ecological
and management factors. An example is the detection
of ‘pockets’ of bovine tuberculosis in England in areas
where infection of badgers is recorded. Secondly, the
control of infectious disease has freed animals from
major causes of death, thereby facilitating the emergence of non-infectious diseases as major problems:
examples are the cardiac, dermal and renal diseases

of dogs. Many of these diseases have a poorly understood, often complex (i.e., multifactorial) cause.
Thirdly, the intensification of animal industries has
highlighted new ‘diseases of production’, often manifested as poor performance, rather than clinical

disease, and frequently with multifactorial causes.
Fourthly, economic evaluation has become important:
the economic advantages of disease control, which are
obvious with the major animal plagues such as rinderpest, can be difficult to identify when overt disease and
dramatic changes in levels of performance are not
involved. These four changes in the approach to,
and appreciation of, disease have added momentum
to the emergence of veterinary epidemiology as a discipline concerned with the measurement of the
amount of disease and its economic effects, the identification and quantification of the effects of factors
associated with disease, and the assessment of the
effects of prevention and treatment of disease in
groups of animals.
A knowledge of elementary statistics is essential for
an understanding of the full range of epidemiological
techniques. Hitherto, most epidemiology books either
have assumed a knowledge of statistics or have avoided
a description of the mathematical manipulations that
are commonly used in epidemiology. However, the
extent of statistical teaching varies widely between veterinary schools. Two chapters therefore are included
as an introduction to basic statistics, and are intended
to make this book statistically self-sufficient (though
not comprehensive). Similarly, a chapter includes an
introduction to computers, which are now used widely
in the recording and analysis of epidemiological data.

xx

From the preface to the second edition
Since publication of the first edition, veterinary medicine has faced several new problems, and has been
forced to evaluate established ones more critically.
Bovine spongiform encephalopathy emerged as a serious problem in the United Kingdom. Rinderpest is still
the subject of a global eradication campaign. There is
an increasing demand for comprehensive, high-quality
technical and economic information on animal disease
and productivity at the national and international
level; and information systems, such as the United
States’ National Animal Health Monitoring System,
have been designed to suit these requirements. The
moves towards an open market, both in the European
Union and internationally following the Uruguay
round of the General Agreement on Tariffs and Trade,
highlight the need for information on animal disease
status in trading nations. Multifactorial diseases continue to predominate in intensive production systems,

and many companion animal diseases are similarly
complex. The solving of these problems and fulfilling
of these tasks rely heavily on epidemiological principles and techniques.
All chapters of the first edition have been revised.
Chapter 11 has been modified to take account of the
increasing popularity of microcomputers, and the
rapid development of veterinary information systems.
New chapters on clinical trials and comparative epidemiology have been added in response to suggestions
from colleagues. More statistical methods are included
in Chapters 12–15 and 17. The goal of this edition nevertheless remains the same as that of the first: to provide an introduction to veterinary epidemiology for
veterinary undergraduates, postgraduates who have

received limited or no training in epidemiology, and
practising veterinarians and members of other disciplines with an interest in the subject.


xxi

From the preface to the third edition
The ten years since publication of the second edition
have witnessed further expansion in the application
of veterinary epidemiology. Quantitative methods
have increased dramatically, both in their development and use. Evidence-based clinical veterinary medicine is now widely appreciated, and relies heavily of
the results of epidemiological analyses, including
observational studies, clinical trials and the quantitative interpretation of diagnostic tests. The successful
employment of epidemiological techniques has witnessed continued progress in rinderpest eradication,
whereas the epidemic of foot-and-mouth disease in
Europe (notably in the UK) in 2001 presented fresh
challenges.

All chapters from the second edition have been
revised. Numerous topics (e.g., causality, now exclusively
considered in Chapter 3) have been expanded in
response to questions raised by undergraduates, postgraduates and professional colleagues, with the aim of
consolidating the more enduring principles and concepts of epidemiology. Surveillance is now addressed
specifically in Chapter 10. Chapter 17 has been enlarged
in response to increased interest in diagnostic-test validation and performance. The opportunity also has been
taken to correct several typographical errors that were
present in the second edition. The text continues to be
an introduction to veterinary epidemiology, directed at
all with an interest in the subject.

xxii

Preface to the fourth edition
Since the publication of the third edition, veterinary
epidemiology has continued to contribute to the successful control of animal disease; notably, in the global
eradication of rinderpest. However, some diseases
remain refractory, both at the international level
(e.g., rabies) and nationally (e.g., bovine tuberculosis
in England). Additionally, new microbes, such as
Schmallenberg virus, pose fresh challenges. Evidence-based veterinary medicine, which relies heavily
on epidemiological techniques, has continued to
expand in both livestock and companion-animal practice. Similarly, epidemiology is central to the practice
of ‘One Health’, which recognizes the link between
human, animal and ecosystem health, and which
recently has gained increased attention.
All chapters of the third edition have been revised.
Although the subject coverage remains essentially
the same as earlier editions, the order of some chapters
has been changed. The first part of the book now concerns mainly qualitative material. This is followed by
largely quantitative topics. The last part addresses
the application of epidemiology to disease control.
Risk analysis, which was included as an appendix to
the paperback re-issue of the third edition, is now
incorporated into the main body of the text. An
updated version of the bibliography of observational
studies given in Appendix XXII of the third edition
has been excluded from this edition on the grounds
that, with the substantial increase in the number of
such studies, it is unlikely to be comprehensive, whilst

also adding considerably to the size of the book. Some
old references have been retained, and some added, to
give the reader a perspective on the historical development of knowledge and ideas. New, more advanced,
chapters on the design of observational studies, validity in epidemiological studies, systematic reviews, and
statistical modelling have been included, should the
reader wish to pursue these topics in depth. However,
the goal of this edition remains the same as that of the
first three: to provide an introduction to veterinary
epidemiology for veterinary undergraduates, postgraduates who have received little or no training in veterinary epidemiology, and practising veterinarians and

members of other disciplines with an interest in the
subject.
My gratitude is again due to colleagues whose
scholarship and comments during the writing of this
fourth edition have been invaluable. Ian Handel was
very generous with his time in engaging in discussions on test validation without a gold standard,
and in providing data for the Bland–Altman analysis
in Chapter 20. David Argyle supplied comments on
the aetiology of cancer, discussed in Chapter 5. Alex
Donaldson criticized aspects of disease control in the
last chapter. Helen Brown, Louise Kelly and Keith
Howe are recognized for their contributions to Chapters 13, 24 and 25, respectively. I am also indebted to
Robert Christley, who collaborated with Nigel French
in writing Chapters 16 and 18 and with Peter Diggle
in writing Chapter 22, and who supplied additions to
some of the other chapters. Annette O’Connor, Jan
Sargeant and Hannah Wood wrote Chapter 19, and
Marshall Dozier passed valuable comments on this
chapter.
Eoghan Clarkson and Brian Mather skilfully converted rough diagrams into reproducible figures.

Pat Fairbairn, Doreen Graham, Bogusia McRoberts
and Lisa Norman, librarians at the Royal (Dick) School
of Veterinary Studies, University of Edinburgh, efficiently obtained many of the references cited in
the text.
Justinia Wood at Wiley continued her earlier
involvement with previous editions of the book, and,
latterly, Erica Judisch and Purvi Patel took over her
role. Jane Andrew, Nick Morgan, James Schultz and
Kathy Syplywczak assisted with conversion of the
manuscript to published text.
Finally, I am grateful to the Literary Executors of the
late Sir Ronald A. Fisher FRS and the late Dr Frank
Yates FRS, and to the Longman Group Limited, London, for permission to reprint Tables III, IV, V and VII
from their book, Statistical Tables for Biological, Agricultural and Medical Research (6th edition, 1974).
Michael Thrusfield
Edinburgh


Preface to the fourth edition

Over the years since publication of the third edition,
veterinary science, veterinary epidemiology and animal health more generally have faced new challenges
and opportunities. The eradication of rinderpest has
seen the successful elimination of only the second disease and the first animal disease. However, the continued emergence of new animal diseases, the increased
recognition of the role of zoonotic diseases in the
emergence of diseases of people, and increasing appreciation of the impacts of global change, including climate change, global movement of people, animals and
products has led to greater appreciation of the need to
understand these phenomena and the changes in
health with which they are associated. Simultaneously,
new methods have been developed and older methods

‘rediscovered’. From techniques for using ‘big data’ to
participatory approaches to engaging with communities, veterinary epidemiology continues to develop
across a broad front. It is an exciting time.
This fourth edition sees incorporation of additional
chapters on epidemiological study design, validity of
epidemiological studies, statistical modelling and systematic reviews. These new chapters provide up to
date information on these key issues, complementing,
it is hoped, the material in the revised and updated
chapters from the third edition.
I am indebted to many people who have contributed to development of this book. In particular,
Michael Thrusfield has provided essential guidance
and scrutiny. At Wiley, Justinia Wood was a source
of support and advice that has been seamlessly continued by Erica Judisch. I am indebted to Nigel
French and Peter Diggle for agreeing to co-author

chapters with me, and to Annette O’Connor, Jan Sargeant and Hannah Wood who have authored the
chapter on systematic reviews. Numerous people
have provided guidance and critique on sections of
the text, including, but not limited to, Gina Pinchbeck, Dan O’Neill, Dave Brodbelt, Dai Grove White,
Cathy McGowan, Alan Radford, Navneet Dhand and
Peter Cripps.
Many colleagues at the University of Liverpool have
provided advice, including Sarah O’Brien, Matthew
Baylis, Malcolm Bennett and Jo Turner. I have been
(and continue to be) lucky to work with many talented postgraduate students who have challenged
me to think again about things I might otherwise have
taken for granted: David Sutton, Nick Malikides, Joel
Hotchkiss, Emily Brook, Maz Behnke, Carri Westgarth, Chris Jewell, Susie Robinson, Jane Murray,
Marnie Brennan, Lutfi Al Tunesi, Bryony Parsons,
Richard Smith, John Mehers, Andy Stringer, Claire

Scantlebury, Alex Berriman, Puteri Nohuddin, Judy
Bettridge, Zelalem Gutu, Lisa Jameson, Kwankamon
Dittakan, Gabby Laing, Elsa Sandoval Barron, Sophie
Beale, Charlotte Rôbin, Margaux Mesle, Stephanie
Begemann, Sara Owczarczak-Garstecka, Mohamed
AL-Zeyadi, Arturo Hernandez Colina, Tamzin Furtado, John Tulloch, Catherine McLeonard and Hannah Brindle.
Finally, my heartfelt thanks and love go to Rosa and
Jake, who have, more than is reasonable, ridden this
rollercoaster with me.
Robert Christley
Liverpool

xxiii


xxiv

About the companion website
This book is accompanied by a companion website:
www.wiley.com/go/veterinaryepidemiology
The website material was produced by Robert Christley and includes:

•

Interactive multiple-choice questions


1

1

The development of veterinary medicine
Veterinary epidemiology is concerned with disease in
animal populations. Its evolution has spanned several
centuries and has been central to the successful control of many animal diseases. This introductory chapter traces the development of veterinary medicine in
general (including relevant aspects of human medicine), showing that it has been inseparably linked to
that of veterinary epidemiology.
Although man’s association with animals began in
prehistoric times, the development of scientific veterinary medicine is comparatively recent. A milestone in
this growth was the establishment of the first permanent veterinary school at Lyons, France, in 1762. Early
developments were governed largely by economic
rather than humanitarian motives, associated with
the importance of domestic stock as a source of food
and as working animals; and there are still important
economic reasons for concern about disease in animal
populations. Later, with the advent of the industrial
revolution and the invention of the internal combustion engine, the importance of draft animals declined
in the more-economically-developed countries.
Although dogs and cats have been companion animals
for several thousand years, it is only relatively recently
that they and other pets have increased in importance
as components of human society.
Until the last half of the 20th century, the emphasis
of veterinary medicine had been on the treatment of
individual animals with clearly identifiable diseases
or defects. Apart from routine immunization and prophylactic treatment of internal parasites, restricted
attention had been given to herd health and comprehensive preventive medicine, which give proper consideration to both infectious and non-infectious
diseases.
Currently, the nature of traditional clinical practice
is changing in the more-economically-developed
countries. The stock owner is better educated, and,

among livestock, the value of individual animals relative to veterinary fees has decreased. Therefore, contemporary large-animal practitioners, if they are to

meet modern requirements, must support herd health
programmes designed to increase production by preventing disease, rather than just dispensing traditional
treatment to clinically sick animals.
In the less-economically-developed countries, the
infectious diseases still cause considerable loss of animal life and production. Traditional control techniques, based on identification of recognizable signs
and pathological changes, cannot reduce the level of
some diseases to an acceptable degree. Different techniques, based on the study of patterns of disease in
groups of animals, are needed.
Similarly, contemporary companion-animal practitioners, like their medical counterparts, are becoming
increasingly involved with chronic and refractory diseases which can be understood better by an investigation of the diseases’ characteristics in populations.
This chapter outlines the changing techniques of
veterinary medicine by tracing man’s attempts at controlling disease in animals, and introduces some current animal disease problems that can be solved by
an epidemiological approach.

Historical perspective
Domestication of animals and early methods of
healing
The importance of animal healers has been acknowledged since animals were initially domesticated, when
they were already likely to have been chronically
affected by various infections (McNeill, 1977). The
dog, naturally a hunter, was probably the first animal
to be domesticated over 14 000 years ago when it
became the companion of early hunters, with evidence
of close proximity to humans as early as 31 000 years
ago (Germonpré et al., 2009); and differentiation from
its ancestor, the wolf, was likely to have occurred at
least 10 000 years ago, as hunter-gatherer societies
gradually evolved into sedentary agricultural

Veterinary Epidemiology, Fourth Edition.
© 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd.
Companion website: www.wiley.com/go/veterinaryepidemiology


2

1 The development of veterinary medicine

populations (Vilà et al., 1997). Sheep and goats were
domesticated by 9000 BC in the fertile Nile valley and
were the basis of early pastoral cultures. A few of these
societies have lasted (e.g., Pfeffer and Behera, 1997),
but many were superseded by cattle cultures; in some,
the pig increased in importance (Murray, 1968). An
Egyptian cattle culture evolved from 4000 BC, and farming spread from the Near East into Europe (Figure 1.1).
There is archaeological evidence of cattle shrines in
Anatolia dating back to 6000 BC (Mellaart, 1967). This
record illustrates that animals had religious, as well as
economic, significance in early civilizations. The
aurochs was central to the religion of the Sumerians,
who migrated throughout Asia, North Africa and
Europe in the third millennium BC taking their animals
and beliefs with them. India is the largest cattle culture
that remains. Cattle cultures also persist in north-east
Africa, the result of interaction between the Ancient
Egyptians and early Nilotic tribes. Cattle still play
important roles in these cultures: they are food, companionship, and status and religious symbols to the
Suk (Beech, 1911) and Dinka (Lienhardt, 1961) of South

Sudan.
The first extensive colonization of the Eurasian
steppe and semi-arid areas occurred in the third millennium BC. The horse provided the key to successful
exploitation of the area north of the Black Sea, the
Caucasus, and the Taurus and Zagros mountains
(Barraclough, 1984), and a Eurasian horse culture,
associated with warrior tribes, emerged (Simpson,
1951). Some of these tribes overran the older cattle
cultures. The horse is represented in Iranian, Greek
and Celtic pantheons. It has become a symbol of veterinary medicine in the form of a centaur, one of

which, Chiron, was considered to be the mythological
founder of Greek medicine.
There have been several movements of animals with
concomitant social and agricultural modifications
since the early changes. The camel was introduced into
Saharan Africa in the first century BC, and into the SubSaharan region around AD 400 (Spencer and Thomas,
1978; Phillipson and Reynolds, 1996), the latter already
having well established domestic cattle and goat populations (Cain, 1999; Tefera, 2004). The Spanish introduced cattle, sheep, pigs and goats to North America
in the 16th century. Haired sheep were introduced
to Africa by European slave traders. The Spanish
brought turkeys to Europe from North America.
The early Egyptian healers combined religious and
medical roles by being priest-healers, often associated
with temples. Their therapeutic techniques are
recorded in the veterinary Papyrus of Kahun (c. 1900
BC). Literary records of similar age, describing veterinary activities, are extant from other parts of the world,
such as Indian Sanskrit texts from the Vedic period
(1800–1200 BC).
Changing concepts of the cause of disease

Concepts of the cause of disease have changed and
evolved1. A method of treatment used by early
Egyptians was incantation. This was partly ritual, but
also reflected their belief in supernatural spirits as a
possible cause of disease. Approaches to treatment
and prevention are the direct result of theories of
cause. There have been five main theories up to the
middle of the last century2. One theory was often
superseded by another, but traces of each still can be
seen in different parts of the world.
Demons

3000

Early man attributed disease to supernatural powers,
the product of animism, which imbued all moving
things with a spirit. In this ‘spirit-world’, disease could
be produced by witches3, superhuman entities and
spirits of the dead (Valensin, 1972). Treatment therefore included: placation, for example by sacrifice; exorcism (forcible expulsion); evasion, for instance

00

40

5000

Figure 1.1 A generalized map to show the spread of farming
from the Near East to Europe in years BC. (Adapted from
Dyer, 1990.)

1
Causality is outlined in this chapter specifically in the context
of disease. A more general discussion is presented in Chapter 3.
2
Theories of the cause of disease also have similarities with
theories of the origin of species, and both have rationalistic and
theological dimensions (Bullock, 1992).
3
A witch was originally defined as ‘one who by commerce with
the Devil has a full intention of attaining his (or her) own ends’
(Bodin, 1580). Witchcraft became widespread in Europe between the
12th and 18th centuries. In the depositions of witch trials, there are
many examples of the supposed induction of disease and death in
man and domestic animals by witches (L’Estrange Ewen, 1933).


Historical perspective

Figure 1.2 Plague allegory: the fate
of gamblers and lechers. Giovanni
de Paolo: Allegory of the Plague,
Sienese, 1436–1437. (Reproduced
with permission of bpk/
Kunstgewerbemuseum, Staatliche
Museen zu Berlin/Saturia Linke.)

scattering millet seeds to avoid vampires (Summers,
1961); and transference, often to human and animal
‘scapegoats’4, probably the best known single example
of which is the Gadarene swine (the Bible: Mark 5,

i–xiii). The techniques included: ritual ceremonies;
material objects that could be suspended (amulets
and periapts), carried (talismans), hung in a building
(fetishes and icons) or displayed in the community
(totems); the use of special people such as witch doctors; and incantations5. The meaning of the Indian
word ‘brahmin’ originally was ‘healer’ because the
Brahmin were a class of healers. In the Neolithic
period (4200–2100 BC), trepanation (the removal of a
4
The scapegoat had the dual purpose of averting and magically
transferring guilt and evil, both generally and at a specific time of
crisis, such as plague or failure of crops. It takes its name from the
Hebrew rites of the Day of Atonement when a goat was driven into
the wilderness after the High Priest had ritually confessed the sins of
the people and transferred them to the goat. The custom occurs
universally from Ancient Babylonian times to modern times, where
human sacrificial scapegoats have been known in some tribal
societies (Cooper, 1990).
5
Historical examples of the use of amulets, talismans and ‘white
witches’ to prevent and control diseases of livestock in England and
the British Colonies are given by Baker (1974) and, as late as the 19th
century, ‘white witches’ (‘charmers’) were recommended by the
Ministry of Agriculture’s veterinarians (St Leger-Gordon, 1994).

bone disc from the skull) may have been practised to
release evil spirits from sick people (Buckland, 1882;
Wakefield and Dellinger, 1939). The practice also is
recorded in veterinary texts as early as the 16th century, but may be much older (Binois, 2015); and the
Greek physician, Galen, practised trepanation on apes

in the second century AD (Arnott et al., 2003).
During the 19th century, many European peasants
still believed that diseases of cattle were caused by evil
spirits, which could be kept at bay by fire (Frazer,
1890), and the African Nuer tribe occasionally still
uses incantations during ritual sacrifice when cattle
epidemics occur (Evans-Pritchard, 1956)6. Moreover,
sacrifice was practised in England as late as the 19th
century (Baker, 1974).
Divine wrath

The demonic theory involved many spirits; the next
development, monotheistic in origin, argued that
disease was the product of a displeased supreme
being: disease was punishment for sinful behaviour
(Figure 1.2). This belief is prominent in the Old Testament, for example, the animal plague of Egypt (the
6
More recently, there has been a trend towards a contemporary
understanding of disease (Hutchinson, 1996).

3