Veterinary Epidemiology PDF

Veterinary epidemiology Third edition Veterinary epidemiology THIRD EDITION Michael Thrusfield Veterinary Clinical Studies Royal (Dick) School of Veterinary Studies University of Edinburgh Blackwell Science First edition © 1986 by Butterworth & Co. (Publishers) Ltd Second edition © 1995 by B...

Veterinary epidemiology Third edition Veterinary epidemiology THIRD EDITION Michael Thrusfield Veterinary Clinical Studies Royal (Dick) School of Veterinary Studies University of Edinburgh Blackwell Science 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 Editorial offices: Blackwell Science Ltd,9600 Garsington Road,Oxford OX4 2DQ,UK Tel: +44 (0) 1865 776868 Blackwell Publishing Professional,2121 State Avenue,Ames,Iowa 50014-8300,USA Tel: +1 515292 0140 Blackwell Science Asia Pty,550 Swanston Street,Carlton,Victoria 3053,Australia Tel: +61 (0)3 8359 1011 The right of the Author to be identified as the Author of this Work has been asserted in accordance with the Copyright,Designs and Patents Act 1988. All rights reserved. N o 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 the UK Copyright,Designs and Patents Act 1988,without the prior permission of the publisher. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names,service marks,trademarks or registered trademarks of their respective owners. The Publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. First published 1986 by Butterworth & Co. (Publishers) Ltd Second edition published 1995 by Blackwell Science Ltd Reissued in paperback with updates 1997 Reprinted 1999,2000,2001,2003 (twice) Third edition published 2005 Reissued in paperback with updates 2007 2 2008 ISBN: 978-1-405-15627-1 Library of Congress Cataloging-in-Publication Data (from the hardback third edition) Thrusfield,M.V. Veterinary epidemiology / Michael Thrusfield. - 3rd ed. p. cm. Includes bibliographical references and index. ISBN-13: 978-0-632-06397-0 (hardback: alk. paper) ISBN-1O: 0-632-06397-1 (hardback: alk. paper) 1. Veterinary epidemiology. I. Title. SF780.9.T48 2005 636.089'44 -dc22 2005 004 105 A catalogue record for this title is available from the British Library Set in 10/12pt Palatino by Graphicraft Ltd,Hong Kong Printed and bound in Great Britain by TJ International Ltd,Padstow,Cornwall The publisher's policy is to use permanent paper from mills that operate a sustainable forestry policy,and which has been manufactured from pulp processed using acid-free and elementary chlorine-free practices. Furthermore,the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. For further information on Blackwell Publishing, visit our website: www.blackwellpublishing.com To Marjory and Harriet, and in memory of David Contents From the preface to the first edition xii Evans' rules 37 From the preface to the second edition xiii Variables 38 Preface to the third edition XIV Types of association 38 Confounding 40 The development of veterinary medicine 1 Causal models 40 Formulating a causal hypothesis 42 Historical perspective 1 Methods of deriving a hypothesis 43 Domestication of animals and early methods of Principles for establishing cause: Hill's criteria 44 healing 1 Changing concepts of the cause of disease 2 4 Describing disease occurrence 46 Impetus for change 4 Quantification in medicine 10 Some basic terms 46 Contemporary veterinary medicine 11 Basic concepts of disease quantification 49 Current perspectives 11 The structure of animal populations 50 The fifth period 16 Contiguous populations 50 Recent trends 16 Separated populations 52 Measures of disease occurrence 53 2 The scope of epidemiology 22 Prevalence 53 Incidence 53 Definition of epidemiology 22 The relationship between prevalence and The uses of epidemiology 23 incidence rate 56 Types of epidemiological investigation 25 Application of prevalence and incidence values 57 Epidemiological subdisciplines 26 Mortality 57 Components of epidemiology 28 Survival 58 Qualitative investigations 28 Example of calculation of prevalence, incidence, Quantitative investigations 28 mortality, case fatality and survival 60 Epidemiology's locale 32 Ratios, proportions and rates 61 The interplay between epidemiology and other Displaying morbidity and mortality values and sciences 32 demographic data 65 The relationship between epidemiology and other Mapping 65 diagnostic disciplines 32 Geographic base maps 67 Epidemiology within the veterinary profession 33 Geographical information systems 69 3 Causality 34 5 Determinants of disease 75 Philosophical background 34 Classification of determinants 75 Causal inference 35 Host determinants 78 Methods of acceptance of hypotheses 36 Genotype 78 Koch's postulates 37 Age 79  i Contents Sex 79 Trends in the temporal distribution of disease 144 Species and breed 80 Short-term trends 144 Other host determinants 81 Cyclical trends 144 Agent determinants 82 Long-term (secular) trends 145 Virulence and pathogenicity 82 True and false changes in morbidity and Gradient of infection 85 mortality 146 Outcome of infection 86 Detecting temporal trends: time series Microbial colonization of hosts 87 analysis 146 Environmental determinants 88 Trends in the spatial and temporal Location 88 distribution of disease 150 Climate 88 Spatial trends in disease occurrence 150 Husbandry 90 Space-time clustering 151 Stress 91 Interaction 92 9 The nature of data 152 Biological interaction 93 Classification of data 152 Statistical interaction 94 Scales (levels) of measurement 153 The cause of cancer 95 Composite measurement scales 155 Data elements 156 6 The transmission and maintenance of Nomenclature and classification of disease 156 infection 98 Diagnostic criteria 157 Horizontal transmission 98 Sensitivity and specificity 158 Types of host and vector 98 Accuracy, refinement, precision, reliability Factors associated with the spread of infection 100 and validity 159 Routes of infection 103 Bias 160 Methods of transmission 105 Representation of data: coding 161 Long-distance transmission of infection 106 Code structure 162 Vertical transmission 110 Numeric codes 162 Types and methods of vertical transmission 110 Alpha codes 163 Immunological status and vertical transmission 110 Alphanumeric codes 164 Transovarial and trans-stadial transmission in Symbols 165 arthropods 110 Choosing a code 165 Maintenance of infection 111 Error detection 166 Hazards to infectious agents 111 Maintenance strategies 112 10 Surveillance 168 Some basic definitions and principles 168 7 The ecology of disease 116 Definition of surveillance 168 Basic ecological concepts 116 Goals of surveillance 169 The distribution of populations 116 Types of surveillance 169 Regulation of population size 120 Some general considerations 171 The niche 123 Sources of data 173 Some examples of niches relating to disease 124 Mechanisms of surveillance 179 The relationships between different types of Surveillance networks 179 animals and plants 126 Surveillance in developing countries: Ecosystems 130 participatory epidemiology 179 Biotope 130 Techniques of data collection 184 Types of ecosystem 130 Strengths and weaknesses of participatory Landscape epidemiology 132 epidemiology 186 Nidality 132 Some examples of participatory epidemiology 186 Objectives of landscape epidemiology 133 11 Data collection and management 188 8 Patterns of disease 137 Data collection 188 Epidemic curves 137 Questionnaires 188 The Reed-Frost model 140 Quality control of data 195 Kendall's waves 142 Data storage 196  Contents Database models 196 Parametric and non-parametric techniques 249 Non-computerized recording techniques 197 Hypothesis testing versus estimation 249 Computerized recording techniques 198 Sample size determination 250 Data management 201 Statistical versus clinical (biological) Changing approaches to computing 201 significance 250 The Internet 203 Interval and ratio data: comparing means 252 Veterinary recording schemes 204 Hypothesis testing 252 Scales of recording 204 Calculation of confidence intervals 252 Veterinary information systems 205 What sample size should be selected? 253 Some examples of veterinary databases and Ordinal data: comparing medians 254 information systems 207 Hypothesis testing 254 Calculation of confidence intervals 257 1 2 Presenting numerical data 214 What sample size should be selected? 258 Nominal data: comparing proportions 258 Some basic definitions 214 Hypothesis testing 258 Some descriptive statistics 215 Calculation of confidence intervals 261 Measures of position 216 What sample size should be selected? 261 Measures of spread 216 X2 test for trend 262 Statistical distributions 217 Correlation 263 The Normal distribution 217 Multivariate analysis 264 The binomial distribution 218 Statistical packages 265 The Poisson distribution 218 Other distributions 218 1 5 Observational studies 266 Transformations 219 Normal approximations to the binomial and Types of observational study 266 Poisson distributions 219 Cohort, case-control and cross-sectional Estimation of confidence intervals 220 studies 266 The mean 220 Ecological studies 269 The median 221 Measures of association 269 A proportion 221 Relative risk 269 The Poisson distribution 221 Odds ratio 270 Some epidemiological parameters 222 Attributable risk 272 Other parameters 223 Attributable proportion 273 Bootstrap estimates 223 Interaction 274 Displaying numerical data 224 The additive model 275 Monitoring performance: control charts 224 Bias 276 Controlling bias 278 13 Surveys 228 What sample size should be selected? 281 Calculating the power of a study 282 Sampling: some basic concepts 228 Calculating upper confidence limits 283 Types of sampling 229 Multivariate techniques 284 Non-probability sampling methods 230 The logistic model 284 Probability sampling methods 230 What sample size should be selected? 232 16 Clinical trials 289 Estimation of disease prevalence 232 Definition of a clinical trial 289 Detecting the presence of disease 238 Design, conduct and analysis 291 The cost of surveys 242 The trial protocol 291 Calculation of confidence intervals 242 The primary hypothesis 291 The experimental unit 293 14 Demonstrating association 247 The experimental popUlation 294 Some basic principles 247 Admission and exclusion criteria 294 The principle of a significance test 247 Blinding 294 The null hypothesis 248 Randomization 295 Errors of inference 248 Trial designs 296 One- and two-tailed tests 248 What sample size should be selected? 297 Independent and related samples 249 Losses to 'follow-up' 298 '< Contents Compliance 298 Process simulation modelling 346 Terminating a trial 299 Monte Carlo simulation modelling 347 Interpretation of results 299 Matrix population modelling 349 Meta-analysis 300 Network population modelling 351 Goals of meta-analysis 300 Systems modelling 352 Components of meta-analysis 301 The rational basis of modelling for active Sources of data 301 disease control 352 Data analysis 302 Available knowledge, and the functions of models 352 17 Diagnostic testing 305 From theory to fact 353 Model-building 354 Serological epidemiology 305 Assaying antibodies 306 357 20 The economics of animal disease Methods of expressing amounts of antibody 306 Quantal assay 307 Popular misconceptions 357 Serological estimations and comparisons in Economic concepts and principles 358 populations 308 Disease as an economic process 359 Antibody prevalence 308 Assessing the economic costs of disease 361 Rate of seroconversion 309 Optimum control strategies 361 Comparison of antibody levels 309 Cost-benefit analysis of disease control 363 Interpreting serological tests 311 Partial farm budgets 363 Refinement 311 Social cost-benefit analysis 364 Accuracy 312 Evaluation and interpretation of 21 Health schemes 368 diagnostic tests 313 Private health and productivity schemes 368 Sensitivity and specificity 313 Structure of private health and productivity Predictive value 316 schemes 368 Likelihood ratios 318 Dairy health and productivity schemes 370 ROC curves 321 Pig health and productivity schemes 372 Aggregate-level testing 323 Sheep health and productivity schemes 373 Multiple testing 323 Beefhealth and productivity schemes 375 Diagnostic tests in import risk assessment 325 National schemes 378 Guidelines for validating diagnostic tests 327 Accredited/attested herds 378 Agreement between tests 327 Health schemes 378 Practical application of diagnostic tests 329 Companion-animal schemes 379 18 Comparative epidemiology 331 22 The control and eradication of disease 384 Types of biological model 331 Definition of 'control' and 'eradication' 384 Cancer 332 Strategies of control and eradication 385 Monitoring environmental carcinogens 332 Important factors in control and eradication Identifying causes 333 programmes 393 Comparing ages 334 Outbreak investigation 398 Some other diseases 336 Cause known: foot-and-mouth disease 398 Diseases with a major genetic component 336 Cause unknown: chronic copper poisoning 400 Some non-infectious diseases 337 Veterinary medicine in the 21st century 401 Diseases associated with environmental Livestock medicine 401 pollution 337 Companion-animal medicine 402 Reasoning in comparative studies 338 General reading 404 19 Modelling 340 Appendices 406 Types of model 341 Modelling approaches 341 Appendix I: Glossary of terms 407 Deterministic differential calculus modelling 341 Appendix II: Basic mathematical notation and Stochastic differential calculus modelling 344 terms 412 Empirical simulation modelling 345 Appendix III: Some computer software 414  Contents ,\1 Appendix IV: Veterinary epidemiology on Appendix XV: Probabilities associated with the Internet 418 the upper tail of the Normal distribution 448 Appendix V: Student's t-distribution 420 Appendix XVI: Lower- and upper-tail Appendix VI: Multipliers used in the probabilities for Wx, the Wilcoxon-Mann­ construction of confidence intervals Whitney rank-sum statistic 449 based on the Normal distribution, Appendix XVII: Critical values of P for the for selected levels of confidence 421 Wilcoxon signed ranks test 454 Appendix VII: Values of exact 95% confidence Appendix XVIII: Values of K for calculating limits for proportions 422 95% confidence intervals for the difference Appendix VIII: Values from the Poisson between population medians for two distribution for calculating 90%, 95% independent samples 456 and 99% confidence intervals for Appendix XIX: Values of K* for calculating observed numbers from 0 to 100 429 95% confidence intervals for the difference Appendix IX: The X2-distribution 431 between population medians for two Appendix X: Technique for selecting a simple related samples 459 random sample 432 Appendix XX: Common logarithms (lOglO) of Appendix XI: Sample sizes 434 factorials of the integers 1-999 460 Appendix XII: The probability of detecting Appendix XXI: The correlation coefficient 462 a small number of cases in a population 441 Appendix XXII: Some veterinary observational Appendix XIII: The probability of failure to studies 463 detect cases in a population 443 Appendix XXIII: The variance-ratio (F) Appendix XIV: Sample sizes required for distribution 480 detecting disease with probability, PIt and Appendix XXIV: Risk analysis 482 threshold number of positives (in brackets) (probability of incorrectly concluding that References 503 a healthy population is diseased [in square Index 593 bracketsD 444 From the preface to the first edition The common aim of the many disciplines that com­ frequently with multifactorial causes. Fourthly, eco­ prise veterinary medicine is an increase in the health nomic evaluation has become important: the economic of animal populations, notably of domestic livestock advantages of disease control, which are obvious with and companion animals. This goal has traditionally the major animal plagues such as rinderpest, can be been achieved by individual diagnosis and treatment: difficult to identify when overt disease and dramatic procedures that evolved contemporaneously in veterin­ changes in levels of performance are not involved. ary and human medicine, when infectious diseases, These four changes in the approach to, and appre­ which had predominantly single causes and clearly ciation of, disease have added momentum to the identifiable signs, were commonplace. emergence of veterinary epidemiology as a discipline Four major changes in the veterinarian's appreci­ concerned with the measurement of the amount of ation of and approach to disease problems have occurred disease and its economic effects, the identification and over the past 20 years. First, despite traditional control quantification of the effects of factors associated with techniques, for example slaughter and vaccination, disease, and the assessment of the effects of prevention some diseases remain at refractory levels and now and treatment of disease in groups of animals. require continuous scrutiny to detect changing levels A knowledge of elementary statistics is essential of occurrence associated with ecological and manage­ for an understanding of the full range of epidemiolo­ ment factors. An example is the detection of 'pockets' gical techniques. Hitherto, most epidemiology books of bovine tuberculosis in England in areas where either have assumed a knowledge of statistics or infection of badgers is recorded. Secondly, the control have avoided a description of the mathematical manip­ of infectious disease has freed animals from major ulations that are commonly used in epidemiology. causes of death, thereby facilitating the emergence of However, the extent of statistical teaching varies widely non-infectious diseases as major problems: examples between veterinary schools. Two chapters therefore are the cardiac, dermal and renal diseases of dogs. are included as an introduction to basic statistics, Many of these diseases have a poorly understood, and are intended to make this book statistically self­ often complex (i.e., multifactorial) cause. Thirdly, the sufficient (though not comprehensive). Similarly, a intensification of animal industries has highlighted chapter includes an introduction to computers, which new 'diseases of production', often manifested as are now used widely in the recording and analysis of poor performance, rather than clinical disease, and epidemiological data. - From the preface to the second edition Since publication of the first edition, veterinary medi­ tion systems, and many companion animal diseases cine has faced several new problems, and has been are similarly complex. The solving of these problems forced to evaluate established ones more critically. and fulfilling of these tasks rely heavily on epidemio­ Bovine spongiform encephalopathy emerged as a logical principles and techniques. serious problem in the United Kingdom. Rinderpest is All chapters of the first edition have been revised. still the subject of a global eradication campaign. There Chapter 11 has been modified to take account of the is an increasing demand for comprehensive, high­ increasing popularity of microcomputers, and the quality, technical and economic information on animal rapid development of veterinary information systems. disease and productivity at the national and inter­ New chapters on clinical trials and comparative epi­ national level; and information systems, such as the demiology have been added in response to sugges­ United States' National Animal Health Monitoring tions from colleagues. More statistical methods are System, have been designed to suit these require­ included in Chapters 12-15 and 17. The goal of this edi­ ments. The moves towards an open market, both in the tion nevertheless remains the same as that of the first: European Union and internationally following the to provide an introduction to veterinary epidemiology Uruguay round of the General Agreement on Tariffs for veterinary undergraduates, postgraduates who and Trade, highlight the need for information on have received limited or no training in epidemiology, animal disease status in trading nations. Multifactorial and practising veterinarians and members of other diseases continue to predominate in intensive produc- disciplines with an interest in the subject. Preface to the third edition The ten years since publication of the second edition described in Chapter 22. Andrew Catley, Matthias have witnessed further expansion in the application of Greiner, Sam Mansley, Andrew Paterson, Nick Taylor veterinary epidemiology. Quantitative methods have and David Whitaker criticized various parts of the text. increased dramatically, both in their development and lowe a particular debt of thanks to George Gettinby use. Evidence-based clinical veterinary medicine is now and Keith Howe, whose scholarship and guidance widely appreciated, and relies heavily on the results of have been enthusiastically offered to me during the epidemiological analyses, including observational writing of all three editions of this book, and whose studies, clinical trials and the quantitative interpreta­ longstanding friendship is particularly cherished. Sally tion of diagnostic tests. The successful employment of Macaulay typed several tables, and Rhona Muirhead epidemiological techniques has witnessed continued again produced numerous figures. Susanna Baxter, progress in rinderpest eradication, whereas the epi­ Samantha Jackson and Emma Lonie of Blackwell demic of foot-and-mouth disease in Europe (notably in Publishing, and Mary Sayers, assisted with conversion the UK) in 2001 presented fresh challenges. of the manuscript to published text. Finally, I am grate­ All chapters from the second edition have been ful to the Literary Executors of the late Sir Ronald A. revised. Numerous topics (e.g., causality, now expli­ Fisher FRS and the late Dr Frank Yates FRS, and to the citly considered in Chapter 3) have been expanded in Longman Group Limited, London, for permission to response to questions raised by undergraduates, post­ reprint Tables III, IV, V and VII from their book, graduates and professional colleagues, with the aim of Statistical Tables for Biological, Agricultural and Medical consolidating the more enduring principles and con­ Research (6th edition, 1974). cepts of epidemiology. Surveillance is now addressed specifically in Chapter 10. Chapter 17 has been en­ Author's note on the paperback reissue larged in response to increased interest in diagnostic­ test validation and performance. The opportunity also The publication of this paperback edition has provided has been taken to correct several typographical errors the opportunity to correct typographical errors. The that were present in the second edition. The text con­ 'Further reading' section of several chapters has also tinues to be an introduction to veterinary epidemio­ been expanded. The opportunity has also been taken logy, directed at all with an interest in the subject. to include an appendix on risk analysis (Appendix My gratitude is again due to colleagues whose XXIV) in response to requests from readers. I am comments during the writing of this third edition have grateful to Stuart MacDiarmid and aforementioned been invaluable. Michael Campbell and Janet Wittes colleagues for their perceptive comments on this held lively email discussions in relation to sample-size addition. calculation for the comparison of two sets of ordinal data, and an appropriate method is now included in Michael Thrusfield Chapter 14. Bruce Gummow provided additional September information on the chronic copper poisoning incident 2006  The development of veterinary medicine Veterinary epidemiology is concerned with disease in owner is better educated, and the value of individual animal populations. Its evolution has spanned several animals relative to veterinary fees has decreased. centuries and has been central to the successful control Therefore, contemporary large animal practitioners, if of animal disease. This introductory chapter traces they are to meet modern requirements, must support the development of veterinary medicine in general herd health programmes designed to increase produc­ (including relevant aspects of human medicine), tion by preventing disease, rather than just dispensing showing that it has been inseparably linked to that of traditional treatment to clinically sick animals. veterinary epidemiology. In the developing countries, the infectious diseases Although man's association with animals began in still cause considerable loss of animal life and pro­ prehistoric times, the development of scientific veter­ duction. Traditional control techniques, based on inary medicine is comparatively recent. A milestone identification of recognizable signs and pathological in this growth was the establishment of the first per­ changes, cannot reduce the level of some diseases to manent veterinary school at Lyons, France, in 1762. an acceptable degree. Different techniques, based on Early developments were governed largely by eco­ the study of patterns of disease in groups of animals, nomic rather than humanitarian motives, associated are needed. with the importance of domestic stock as a source Similarly, contemporary companion-animal practi­ of food and as working animals; and there are still tioners, like their medical counterparts, are becoming important economic reasons for concern about disease increasingly involved with chronic and refractory dis­ in animal populations. Later, with the advent of the eases which can be understood better by an investiga­ industrial revolution and the invention of the internal tion of the diseases' characteristics in populations. combustion engine, the importance of draft animals This chapter outlines the changing techniques of declined in the developed countries. Although dogs veterinary medicine by tracing man's attempts at and cats have been companion animals for several controlling disease in animals, and introduces some thousand years, it is only recently that they and other current animal disease problems that can be solved by pets have increased in importance as components of an epidemiological approach. 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 Historical perspective or defects. Apart from routine immunization and pro­ phylactic treatment of internal parasites, restricted Domestication of animals and early methods attention had been given to herd health and com­ of healing prehensive preventive medicine, which give proper consideration to both infectious and non-infectious The importance of animal healers has been acknow­ diseases. ledged since animals were initially domesticated, Currently, the nature of traditional clinical practice when they were already likely to have been chronically is changing in all developed countries. The stock affected by various infections (McNeill, 1977). The , The development of veterinary medicine 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 sub­ Saharan 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 intro­ duced to Africa by European slave traders. The Spanish brought turkeys to Europe from North America. The early Egyptian healers combined religious and Fig. 1.1 A generalized map to show the spread of farming from the medical roles by being priest-healers, often associated Near East to Europe in years BC, (From Dyer, 1990.) 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 dog, naturally a hunter, was probably the first animal Indian Sanskrit texts from the Vedic period (1800- to be domesticated over 14 000 years ago when it 1200 Be). became the companion of early hunters. Sheep and goats were domesticated by 9000 BC in the fertile Nile valley and were the basis of early pastoral cultures. Changing concepts of the cause of disease A few of these societies have lasted, for example the Jews, but many were superseded by cattle cultures; in Concepts of the cause of disease have changed and some the pig increased in importance (Murray, 1968). evolved1. A method of treatment used by early An Egyptian cattle culture evolved from 4000 BC, Egyptians was incantation. This was partly ritual, but and farming spread from the Near East into Europe also reflected their belief in supernatural spirits as a (Figure 1.1). There is archaeological evidence of cattle possible cause of disease. Approaches to treatment shrines in Anatolia dating back to 6000 BC (Mellaart, and prevention are the direct result of theories of 1967). This record illustrates that animals had a cause. There have been five main theories up to the religious as well as an economic significance in early middle of the last century2. One theory was often civilizations. The aurochs was central to the religion of superseded by another, but traces of each can still be the Sumerians, who migrated throughout Asia, North seen in different parts of the world. Africa and Europe in the third millennium BC taking their animals and beliefs with them. India is the largest Demons cattle culture that remains. Cattle cultures also persist in north-east Africa, the result of interaction between Early man attributed disease to supernatural powers, the Ancient Egyptians and early Nilotic tribes. Cattle the product of animism which imbued all moving still play important roles in these cultures: they are food, things with a spirit. In this 'spirit-world', disease could companionship, and status and religious symbols be produced by witches3, superhuman entities and to the Suk (Beech, 1911) and Dinka tribes (Lienhardt, spirits of the dead (Valensin, 1972). Treatment there­ 1961) of the Sudan. fore included: placation, for example by sacrifice; The first extensive colonization of the Eurasian steppe and semi-arid areas occurred in the third mil­ lennium Be. The horse provided the key to successful 1 Causality is outlined in this chapter specifically in the context of exploitation of the area north of the Black Sea, the disease. A more general discussion is presented in Chapter 3. 2 Theories of the cause of disease also have similarities with theories Caucasus, and the Taurus and Zagros mountains of the origin of species, and both have rationalistic and theological (Barraclough, 1984), and a Eurasian horse culture, dimensions (Bullock, 1992). associated with warrior tribes, emerged (Simpson, 3 A witch was originally defined as 'one who by commerce with the 1951). Some of these tribes overran the older cattle Devil has a full intention of attaining his (or her) own ends' (Bodin, 1580). Witchcraft became widespread in Europe between the 12th and 18th cultures. The horse is represented in Iranian, Greek centuries. In the depositions of witch trials, there are many examples and Celtic pantheons. It has become a symbol of veter­ of the supposed induction of disease and death in man and domestic inary medicine in the form of a centaur, one of which, animals by witches (L'Estrange Ewen, 1933).  Historical perspective exorcism (forcible expulsion); evasion, for instance Metaphysical medicine scattering millet seeds to avoid vampires (Summers, 1961); and transference, often to human and animal The next development did not assume the existence 'scapegoats'4, probably the best known single example of a supreme being, either demonic or divine, but of which is the Gadarene swine (the Bible: Mark 5, assumed the presence of occult forces beyond the i-xiii). The techniques included: ritual ceremonies; physical universe. This 'metaphysical' medicine em­ material objects that could be suspended (amulets bodied a theory of natural laws but excluded scientific and periapts), carried (talismans), hung in a building principles such as observation and the repeatability (fetishes and icons) or displayed in the community of phenomena. The moon, stars and planets were (totems); the use of special people such as witch considered to affect health (Whittaker, 1960), these doctors; and incantations. The meaning of the Indian concepts being obvious predecessors of astrology. word 'brahmin' originally was 'healer' because the Several outbreaks of rinderpest in Dark Age Europe brahmin were a class of healers. In the neolithic period were ascribed to earthquakes, floods and comets. (4200-2100 BC), trepanning (the removal of a bone Treatment frequently included particularly foul medi­ disc from the skull) may have been practised to release cines and practices that persisted for many centuries. demons from sick people. A recommended 17th century cure for broken wind During the 19th century, many European peasants in horses comprised toads, swallows and moles still believed that diseases of cattle were caused by evil roasted alive and mixed with shoe soles. Divination, spirits, which could be kept at bay by fire (Frazer, practised by the Babylonians using sheep livers, 1890), and the African Nuer tribe occasionally still and the 'Doctrine of Signatures' which suggested uses incantations during ritual sacrifice when cattle a similarity between the disease and its cure - for epidemics occur (Evans-Pritchard, 1956)5. Moreover, example using toads to treat warts - were notable sacrifice was practised in England as late as the 19th metaphysical developments. century (Baker, 1974). The universe of natural law Divine wrath A major intellectual revolution began in Greece in The demonic theory involved many spirits; the next the sixth century BC in which the universe was development, monotheistic in origin, argued that dis­ rationalized without either demonic or metaphysical ease was the product of a displeased supreme being: influences. The Greeks thought that disease was the disease was punishment. This belief is prominent in result of derangement of four humours of the body, the Old Testament, for example the animal plague of which were associated with four properties (heat, Egypt (the Bible: Exodus 9, iii) and is also evident in moisture, dryness and cold) and with four elements Persian and Aztec writings. The only effective treat­ (air, earth, water and fire) (Figure 1.2). Diseases were ment of disease induced in this way was placation considered to be caused by external forces, includ­ because exorcism and evasion would not be effective ing climatic and geological changes that affected the against a supreme being. Traces of this belief have persisted until recent times. The English veterinary surgeon, William Youatt, writing in 1835, supported the practice of burning crosses on the heads of cattle to cure and prevent disease. In 1865, Queen Victoria, 1\ CHARACTERISTIC believing that the current British rinderpest (cattle "'" Moisture Humour Dryness Yellow bile plague) outbreak was the result of divine displeasure, Blood = Humour = -....,'" ordered that a prayer should be used in each church in Associated = Air Associated = Fire element element u England while the epidemic continued. J: Source = Heart Source = Liver en 0: Excess _Sanguine Excess - Choleric w temperament (bilious) u temperament « 4 The scapegoat had the dual purpose of averting and magically trans­ 0: « Humour = Phlegm Humour = Black bile ferring guilt and evil, both generally and at a specific time of crisis, such J: u Associated = Water Associated Earth as plague or failure of crops. It takes its name from the Hebrew rites of the element element Day of Atonement when a goat was driven into the wilderness after "C Source = Pituitary the High Priest had ritually confessed the sins of the people and trans­ (5 Source = Spleen U gland ferred them to the goat. The custom occurs universally from Ancient Excess - Phlegmatic Excess _ Melancholic Babylonian times to modern times, where human sacrificial scapegoats temperament temperament have been known in some tribal societies (Cooper, 1990). 5 More recently, there has been a trend towards a contemporary understanding of disease (Hutchinson, 1996). Fig. 1.2 Components of humoral pathology. 1 The development of veterinary medicine population. Local outbreaks of disease were thought others, either by contact or through the air (Lobb, to be the result of local eruptions of noxious air: 1745). The 18th century American lexicographer and miasmata (miasmas)6. The word 'malaria' literally essayist, Noah Webster, classified diseases as mias­ means 'bad air' and hints at the 19th century belief that matic (e.g., pneumonia) or contagious (e.g., smallpox), the disease was caused by stale air around swamps. representing an intermediate stage in the evolution of The concept of humoral derangement was reim­ the contagion theory (Winslow, 1934). ported into medi val Europe, via Sicily, during the The main advances in the identification of microbes Crusades, and food was imbued with the same as causes of infectious diseases occurred in the 19th properties as the humours (Tannahill, 1968). The con­ century, although the concept of a living contagious cept persists in several cultures. In indigenous Indian agent, contagium animatum, was founded in the 17th Ayurvedic human and veterinary medicine, based on century. Edward Jenner's development of a smallpox the Hindu Scriptures (Vedas), there are three humours vaccine using cowpox-infective material, and early (tridosa): vata (wind), pitta (bile) and kapha (phlegm); biological warfare conducted by American settlers derangement of vata, for example, causing asthma who gave blankets belonging to smallpox victims to and diarrhoea. This concept is also central to modern Indians as presents, implicitly recognized contagion. Mahayana Buddhist medicine. However, in Europe, Louis Pasteur's investigation of anthrax and rabies the popularity of the miasmatic theory declined at (Walden, 2003), and Robert Koch's discovery of the the beginning of the 20th century, by which time the bacteria causing tuberculosis7 and cholera (Munch, microbial theory of infectious disease was adequately 2003), firmly established microbiology and marked the supported. downfall of the miasmatic theory. The set of postulates The Greek idea of disease was susceptible to formulated by Koch to define causal agents has been scientific investigation. Careful observation and the used to identify many microbial diseases since those identification of specific causes became the hallmarks early days of bacteriology (see Chapter 3). of the fifth century Be school of medicine at Cos, and Viruses were also discovered in the late 19th cen­ were refined by Hippocrates whose text, Discourse tury, although not actually 'seen' until the invention on Airs, Waters and Places (Jones, 1923), dominated of the electron microscope in the 1930s. In 1892, medicine for many centuries. Therapy was consistent Iwanowsky demonstrated that tobacco mosaic disease with causal concepts, and included purges and altera­ could be transmitted by sap that had been filtered tions in diet. through bacteria-proof filters ( Witz, 1998). Beijerinck serially transmitted the disease using bacteria-free filtrates, and coined the term contagium vivum Contagion fluidum to describe the infectious 'living' agent. In The idea that some diseases can be transmitted from 1898-99 Loeffler and Frosch discovered the first one animal to another has its ubiquitous origins in animal virus, foot-and-mouth disease virus, and in antiquity, and ancient veterinary accounts of disease 1911 Rous reported the first virus-induced transmis­ provide strong evidence of the concept of conta­ sible tumour. giousness (Bodson, 1994). The Romans, Galen and Towards the end of the 19th century, the first Lucretius, believed that disease could be spread by arthropod carrier (a tick) of an infectious disease was airborne seeds or animacuia (not necessarily living), identified by Kilborne, Smith and Curtis, investigating which were taken in through the nose and mouth. Texas fever of cattle in the US. The Jewish Talmud describes demons as hiding 'everywhere' - in water, crumbs and air - implying transmissibility. The primitive Hindus associated sick Impetus for change rats with human plague, the first suggestion of a zoonosis. The Veronan, Fracastorius, writing in the Changing attitudes towards the cause of disease and early 16th century, argued that diseases were transmit­ the concomitant alterations in techniques of treatment ted by minute, invisible particles (Wright, 1930); and and prevention are a small part of shifts in overall Lancisi, physician to Pope Clement XI, freed Rome scientific thought. These changes have not taken place from rinderpest by using a slaughter policy to prevent gradually, but have occurred as distinct 'revolutions', infection of unaffected animals. Thomas Lobb, writing which terminate periods of stable science (Kuhn, in London in the 18th century, considered that human plague and rinderpest were caused by particles that multiplied in infected individuals and then infected 7 Anthrax and tuberculosis are diseases of great antiquity, dating back at least to Ancient Egypt, where the former bacterium was most likely the 6 This explains why urban Victorians draped thick curtains in their cause of some of the biblical plagues (Blaisdell, 1994; Willcox, 2002; windows in an attempt to keep out disease. Witowski and Parish, 2002; Sternbach, 20(3).  Historical perspective 19708). Each period has its paradigm (model), which the miasmatic theory of cause evolved. Techniques serves to guide research. As time passes, anomalies of treatment required careful recognition of clinical accumulate. Initially, scientists accommodate these by signs following the Greek Coan tradition. Quarantine modifying the paradigm, but a time comes when the (derived from the Italian word meaning 'forty' - the pressures on the old framework become so great that traditional length, in days, of isolation in the Middle a crisis occurs, and there is a revolutionary shift Ages) and slaughter became preventive strategies. in paradigms. For example, in astronomy, the old These local actions, which lasted until the first century Ptolemaic model of the universe had to be modified AD, were incapable of solving major problems in the and remodified by adding new planetary epicycles horse, which was becoming an important military to account for the observed motion of the heavenly animal. This crisis resulted in the second phase: that bodies, but eventually the critical point was reached of military healers. when the old model was falling apart under the strain and was ceasing to be credible. Thus, the time was The second period: the first century AD until 1762 ripe for the dramatic shift in models called the Copernican revolution. Kuhn's thesis has also been Veterinarians specialized in equine medicine and applied to political, social and theological 'revolutions' surgery, reflecting the importance and value of horses (Macquarrie, 1978) and to the applied sciences (e.g., Richards, 1954). A major veterinary text, the (Nordenstam and Tornebohm, 1979) of which veter­ Hippiatrika, comprising letters between veterinarians, inary medicine is a pare. cavalry officers and castrators, dates from early Veterinary medicine has experienced five stable Byzantine times. The major contributor to this work periods and revolutions up to the middle of the 20th was Apsyrtus, chief hippiatros to the army of Con­ century relating to disease control (Schwabe, 1982), stantine the Great. This phase lasted until the mid-18th which stimulated the changes in the causal concepts century and was marked by a continuing interest in already described. The major problem that persisted equine matters. Several important texts were written, during these periods, precipitating crises, was large­ including Vegetius Renatus' Ars Veterinaria (published scale outbreaks of infectious disease: the classical in 1528) and Carlo Ruini's Anatomy of the Horse (pub­ animal piagues10 (Table 1.1). Military campaigns fre­ lished in 1598). Interest was taken in other species, too. quently assisted the spread of these infections (Table 1.2 The 15th century Bake of Saint Albans described dis­ and Karasszon, 1988). eases of falcons (Comben, 1969); and John Fitzherbert's Bake of Husbandrie (published in 1523) included dis­ eases of cattle and sheep. The horse, however, was The first period: until the first century AD pre-eminent. This bias survived in Europe until early The initial domestication of animals brought man into in the 20th century when equine veterinary medicine close contact with animals and therefore with their dis­ was still considered to be a more respectable occupa­ eases. The demonic theory was prevalent. However, tion than the care of other species. despite the use of control techniques consistent with Varying emphasis was placed on the miasmatic the theory, draft animals continued to die, and a crisis and metaphysical theories of cause and on humoral arose when urbanization increased the importance of pathology. The Arabians, for example, based their animals as food resources. This resulted in the develop­ medicine largely on the metaphysical theory. ment of the first stable phase of veterinary medicine. This was characterized by the emergence of veterinary The third period: 1762-1884 specialists such as the early Egyptian priest-healers and the Vedic Salihotriya who founded the first veter­ The animal plagues, especially those of cattle, became inary hospitals. Humoral pathology developed and particularly common in Europe in the mid-18th cen­ tury with the introduction of rinderpest from Asia (Scott, 1996). They provided the next major crisis H For a critical assessment of Kuhn's philosophy, see Hoyningen­ Huene (1993). involving civilian animals. The miasmatic theory per­ 9 The concept of dramatic paradigm shifts, however, may not be sisted but the miasmata were thought to originate applicable to all areas of thought and progress. The 17th century German from filth generated by man, rather than from natural philosopher, Leibnitz, argued that change (e.g., in ethics and aesthetics) sources. A third stable phase developed, characterized is gradual. 10 A plague (Greek: plege = a stroke, a blow; Latin: plangere = to strike) by improvement of farm hygiene, slaughter and traditionally is any widespread infectious disease with a high fatality rate treatment as control techniques. When rinderpest among clinically affected individuals. In veterinary medicine, the term is entered England from Holland in 1714, Thomas Bates, extended to any widespread infectious disease causing major economic surgeon to George I, advocated fumigation of build­ disruption, although the fatality rate may not be high (e.g., foot-and­ mouth disease). In human medicine, the term is now commonly ings, slaughter and burning of affected animals, and restricted to infection with the bacterium, Yersinia (Pasteurella) pestis. resting of contaminated pasture as typical tactics Table 1.1 Some dates of occurrence of animal plagues. (Most dates before 1 9 60 extracted from Smithcors, 1 9 57.) Oate Animal plagues Rinderpest Pleuropneumonia Canine Anthrax Foot-and-mouth disease Equine /II-defined diseases distemper influenza 500 Be Egypt 500 Be - time of Christ Egypt 278 Be abortion AD Rome AD 500 Rome 4th century AD (cattle) France 6th century AD (cattle) Ireland 8th century AD France 8 20, 8 50, 940-943 (cattle) Engl and 1 31 4 (cattle) AD 1 400 England 1 490*, 1 551 Italy 1 51 4 England 1 688 AD 1 700 France 1 71 0-14 Europe 1 8th century Rome 1 71 3 England 1 71 4, 1 745-46 England 1 727 France 1750 Ireland 1 728 US 1 760 England 1 733, 1 737, 1 750, Spain 1 761 1 760, 1 771 , 1 788 England 1 763 AD 1 800 England 1 841 -98 Engl and 1 839 England 1 837 England 1 8 65 England 1 870-72, North America 1 872 1 8 77-85 England 1 889-90 Africa 1 89 0 -1 900  AD 1 900 Bel gium 1 920 Engl and 1 922-25, 1 942, Czechoslovakia 1 9 57 1 9 52, 1 967-68 Britain 1 9 63 US 1 9 63 Canada 1 9 51 -52 Europe 1 965 Pol and 1969 Middle East 1 9 69-70 USSR 1 9 76 Africa 1 9 79 - 84 France, The Netherlands, India 1 9 83-85 Switzerland 1 9 80 Sweden 1 9 78-79 Turkey 1 99 1-92 Channel Isl ands, France, Isle England 1 9 89 of Wight, Italy, Spain 1 9 81 Denmark, Spain 1 9 83 Germany (W.), Greece, The Netherl ands, Portugal 1984 Italy 1 9 85, 1988, 1 99 3 India, S. Arabia 1 99 0 Nepal 1 993-94 Turkey 1 99 5 Albania, Bul garia, Greece, Macedonia, Yugoslavia 1 99 6 Taiwan 1 99 7 Bhutan 1 99 8 China 1 999 AD 2000 Greece, Japan, Mongol ia, S. Africa 2000 France, Ireland, The Netherl ands, UK 2001 * Named as 'steppe murrain' (derivative of Latin marl= to die). i\ The development of veterinary medicine Table 1. 2 Mil itary campaigns that disseminated rinderpest. problem. Public concern, highlighted by the rinderpest outbreak of 1 865 (Scott, 1997), was responsible for the Century Campaign establishment of the British State Veterinary Service in 5th Fall of Rome the same year. Similar services were founded in other 8-9th Charlemagne's conquest of Europe countries. The legislature continued to strengthen 1 2th Genghis Kahn's invasion of Eu rope the power of the veterinary services by passing Acts 1 3th Kublai Kahn's invasion of Europe and China relating to the control of animal diseases. 1 5 -1 6th Spani sh-Hapsburg conquest of Italy 1 7th War of the League of Au gsberg 1 8th War of the Spanish Succession The fourth period: 1884-1960 1 8th War of the Austrian Su ccession 18th Seven Years' War The animal plagues continued despite sanitary cam­ 18-1 9th Napoleonic Wars paigns. This crisis coincided with the inception and 19-2Oth US invasion of the Phil ippines acceptance of the microbial theory which, epitomized 1 9-20th Ital ian conquest of Eritrea 20th Worid War l by Koch's postulates, defined a specific, single cause of 20th World War II an infectious disease and therefore implied a suitable 20th Vietnam War control strategy directed against the causal agent. 20th Lebanese War This fourth stable phase of campaigns or mass 20th Sri Lankan conflict actions began in the 1880s. Treatment of disease was 20th Azerbaijan confl ict 20th Gulf War based on laboratory diagnosis involving isolation of agents and identification of lesions followed by therapy. Control of disease by prevention and, sub­ sequently, eradication involved mass testing of animals Table 1. 3 Veterinary schools founded to combat rinderpest. and immunization when an increasing number of vaccines became available12. The discovery of disease Year of City Country foundation vectors facilitated prevention by vector control. An improved understanding of infectious agents' life 1 762 Lyon France histories enabled their life-cycles to be broken by 1 766 Alfort France manipulating the environment; the draining of land 1 767 Vienna Austria to prevent fascioliasis is a good example. Bacterial 1 769 Turin Piedmont 1 773 Copen hagen Denmark diseases remained as major clinical problems until 1 777 Giessen Hesse the discovery and synthesis of antibiotics in the 20th 1 778 Hannover Hannover century, when, in human medicine, the effectiveness of control measures was starkly reflected in changes in the 'league table' of causes of deaths (Figure 1.3). (Bates, 1717-1719). Cattle owners were also com­ The veterinarian also experienced a similar increase in pensated for loss. By the mid-19th century, disinfection therapeutic power. (notably using carbolic and cresylic acids) was also Many infectious diseases were either effectively con­ being applied to control the disease (Brock, 2002). trolled or eradicated between the latter part of the 19th Half of the cattle in France were destroyed by rinder­ century and the middle of the 20th century in the pest between 1 710 and 1 714. The disease occurred developed countries using the new techniques of the irregularly until 1750, when it again became a serious microbial revolution and older techniques including problem. Little was known about the disease. This quarantine, importation restrictions, slaughter and provided impetus for the establishment of the first hygiene. In 1 892, pleuropneumonia in the US was the permanent veterinary school at Lyons in 1762, and first disease to be regionally eradicated after a cam­ others were subsequently founded to combat the dis­ paign lasting only five years. Notable British successes ease (Table 1.3) 11. included rinderpest, eradicated in 1877, pleuropneu­ The lifting of animal importation restrictions in monia in 1898, and glanders and equine parasitic England in 1 842 increased the risk of disease occurring mange in 1928. in Britain. Sheep pox entered Britain in 1847 from Germany, and pleuropneumonia became a serious 12 There had been earlier attempts to immunize animals against 11 There were other pressures, too, fostering the establishment of rinderpest, mimicking the practice of 'variolation' in humans, in which European veterinary schools. These included the need for military smallpox virus was inoculated to prevent the disease. The first such authorities to improve the effectiveness of horse-dependent armies, and rinderpest trial was undertaken in 171 1 by Bernardino Ramazzini at the the requirement for increased agricultural productivity in the face of University of Padua (Koch, 1891). Some authorities believe that the prac­ population growth (Mathijsen, 1 997). tice of immunization began in China; see Bazin (2003) for a brief history.  Historical perspective 'J /'.. r:r:·'W,:i.6f... 1:I{ank".................................... :.: (. · eatIis 1.Pne m()nla iA 1i nza;·. '1. TubercUlosis bro c tis. ·.. ' 14.4 2. Diarrhoea. and enteritis 2.TtibetculQsis · ' 11.3 3. Cholera 3.DiatrhQe an eriteritiS 8 ·1· 4. Pneumonia-Influenza... 'lHeatt disease. 8.0 bronchitis... 5. Nephdtis i.''... ·. 4.7 5. Infantile convulsions tAccidents < ' 4.5 6. Stroke 7. Strokf '..'...'....... 4.2 7. Diphtheriaaridcroup 8. Diseases of early: infancy 4.2 8. Dysentery 9. Cancer 3.7 9. Scarlet fever to. Diphtheria: 2.3 10. Nephritis 1 2..",··CI U"''''J. ·. ".................... >...... '. '4.' Pneumonia-lrifluenza,;. ·.1>t9n q........ ".... '.6 ·5. ACcld Il s ( ¢ ·r tof· i · ··.·.. ·· ".... y I ).$ i 4e.... '1. 6 ltf()tQryeh1 1 cldeAt$ 2.8... a,rll·.infancy·... 2.3.. 7..·. Dise.a:s ·. Q b... ·. '. '....... , (( ( '(·' "' "... ·..... AtteriosclerQsIS· '.'. 10.. Gi.n:hq i$.·of. Fig.1. 3 League table of causes of human mortality in the UK: (a) 1 860, (b) 1 900, (c) 1 9 70. (Data from Thrusfield, 2001. ) ! I' The development of veterinary medicine Qua ntification in m edicine A pivotal move towards comparative statistical techniques occurred when Pierre-Charles-Alexander The evolution of understanding of the cause of disease Louis developed his 'numerical method', requiring purely qualitatively was accompanied by increased systematic record keeping and rigorous analysis of interest in disease in quantitative terms. This began multiple cases (Bonett, 1973). He documented typhoid primarily as a descriptive exercise. The ancient in Paris, showing that the disease occurred predomin­ Japanese reported outbreaks of animal diseases. John antly in young adults, and that the average age of Graunt (1662) published quantitative observations on fatal cases was higher than that of survivors, suggest­ London parish registers and 'Bills of Mortality'. An ing that the younger patients had the best prognosis outbreak of rinderpest in France in the late 18th cen­ (Louis, 1 836). He subsequently demonstrated that tury was responsible for the establishment of a com­ blood-letting was of no benefit to typhoid cases, and mission on epidemics, headed by Felix Vicq d'Azyr, his calculation of average values was adopted by other Marie Antoinette's personal physician. This evolved early protagonists of clinical trials (e.g., Joseph Lister; into the Royal Society of Medicine, which pioneered see Chapter 1 6). Average values were also applied to the collection of statistical data on animal and human provide a quantitative definition of a 'normal' indi­ epidemics and the weather (Matthews, 1995). vidual; Adolphe Quetelet (1835), for example, recorded the range of the human cardiac and respiratory rates. Application of probability theory to medicine was Post-Renaissance thinking and 'The Enlightenment' cautiously and tendentiously accepted by British and The scientific revolution that began during the 1 6th French medical statisticians, who were largely con­ century posited that the physical universe was orderly cerned with the descriptive statistics of the major and could be explained mathematically (Dampier, 1948). public health issues (e.g., Figure 1.3), rather than with This argument was extended to the biological world, statistical inference. Nevertheless, during the 19th where it was considered that 'laws of mortality' must century, strong links were forged between epidemio­ exist. Graunt's mortality studies included attempts to logists, mathematicians and statisticians through the formulate such laws by constructing life tables (see common influence of Louis (Lilienfeld, 1978) 13, and, by Chapter 4), Edmund Halley (1656-1742) constructed the 20th century, rigorous methods of statistical infer­ life tables for Breslau (Benjamin, 1959), and Daniel ence were developing (Stigler, 1 986) and were being Bernoulli (1700-1 782) applied life-table methods to applied in medicine and agriculture. These methods smallpox data, thereby demonstrating that inoculation necessitate observation of events in populations, was efficacious in conferring lifelong immunity (Speiser, rather than in the individual, and are thus central 1982). A hundred years later, William Farr (Halliday, to the development of quantitative epidemiology (see 2000) produced a simple mathematical model of the Chapter 2). 1 865 rinderpest epidemic in the UK (see Chapter 19). The formulation of physical and biological events, Quantitative analysis of biological (including med­ however, now, as then, needs to be very carefully ical) phenomena evolved in the 1 8th century, when the assessed, and may convey an illusion of certainty Age of Enlightenment saw a growth in literature dealing and security that is not warranted (Gupta, 2001). with the relationship between probability and the Moreover, it is not always considered to be socially need for objectivity in science and society (the 'Prob­ beneficial14, and is not a substitute for rigorous, albeit abilistic Revolution'). The mathematical foundation of sometimes onerous, analysis of field data (The Eco­ probability was laid by Jakob Bernoulli in his Ars nomist, 2002). Additionally, there may be a tendency to Conjectandi, which was published posthumously in use whatever numerical data are available, regardless 1713. He developed a theory of 'inverse probability', of their relevance and quality (Gill, 1993)15. which stated that the frequency of an event would approach its probability of occurrence if the number of observations was large enough. This theory was 13 An interesting 'family tree', showing the links between 1 8th-20th century statisticians, public-health physicians and epidemiologists, is mathematically refined by Simeon-Denis Poisson, who depicted by Lilienfeld and Lilienfeld ( 1 980). proposed a 'law of large numbers', which stated that, if 14 See, for example, Gregory (2002) for a brief theological discussion. an event was observed a large number of times, one 15 Chambers (1997) amalgamates Gupta's and Gill's points, with could assume that the probability of its future occur­ sights set particularly on economists: 'Quantification alld statistics call mis­ lead, distract, be wasteful, simply not make sCllse or conflict with COlnlllOIi values rence would correspond to its observed frequency.... Yet professionals, especially economists and consultants tight for time, have The logical consequence of this conclusion is that, if a strong felt need for statistics. At worst they grub aroulld alld grab what lIum­ there are sufficient observations, sound predictions bers they can, feed them into their computers and prillt out not just numbers, but can be made. Thus, in relation to therapy, Pierre-Simon more and more elegant graphs, bar charts, pie charts allli three-dimensional wonders... Numbers can also reassure by appCllrilig to extmd control, precision Laplace (1814) suggested that a preferred method of and knowledge beyond their real limits... wrong numbers, one might add, are treatment 'will manifest itself more and more in the worst of all because al/numbers pose as true.' Porter (1 995) provides a philo­ measure that the number (of observations) is increased'. sophical discussion of quantification in ge