Veterinary · Para-clinical · also MSVE VPE-321
How disease is measured, studied and controlled in populations, from incidence and prevalence to surveillance and outbreak investigation, plus the major zoonoses and One Health.
In the exams: Nepal NLEV Epidemiology & Public Health 6; SAVC epidemiology 15 and zoonoses 25; EGEL Plus epidemiology 15; ICAR AICE-JRF core group 06
Veterinary epidemiology is the science of disease in groups of animals: how much there is, where and when it occurs, why, and what will reduce it. It supplies the numbers behind every herd health plan, vaccination campaign, eradication programme, import permit and outbreak response. Zoonoses, diseases and infections naturally transmitted between animals and people, are where the veterinarian's population view protects human health most directly. Rabies, brucellosis, bovine tuberculosis, anthrax, leptospirosis and avian influenza are controlled most cheaply at their animal source, and this is the practical core of One Health.
This manual follows the order of a standard course. It starts with the vocabulary of disease occurrence and the measures of frequency and association, moves through study design, sampling, diagnostic test evaluation, surveillance, outbreak investigation and risk analysis, and ends with the major zoonoses, One Health, transboundary diseases and the World Organisation for Animal Health (WOAH). Formulas are written out in words so that each one can be worked by hand; the question bank gives worked numerical examples. Where practice, law or test thresholds differ between countries, the text says so. Doses are not given; follow national guidelines and product labels.
Examiners worldwide test the same core. Licensing and entrance papers (Nepal NLEV, the South African SAVC registration examination, Mexico's EGEL Plus, NAVLE, RCVS, AVBC, India's ICAR PG and PhD entrance and state veterinary officer examinations) ask definitions, calculations of prevalence, incidence, case fatality, sensitivity and predictive values, the shape of epidemic curves, the key features of zoonoses, and WOAH notification rules. Postgraduate and specialist examinations add study design, bias, sample size, risk analysis and surveillance design. The practitioner needs the same tools to read a paper, investigate a herd problem and report a notifiable disease correctly.
Epidemiology is the study of the distribution and determinants of health and disease in populations, and the use of that knowledge to control disease. The clinician looks at one animal and asks what is wrong with it. The epidemiologist looks at a herd, a flock, a district or a country and asks how much disease there is, who has it, where and when it occurs, why, and what will reduce it. Both are needed: a herd outbreak of abortion is diagnosed by the laboratory, but it is controlled by knowing which animals are at risk and how the agent moves between them. In veterinary work the unit of interest is often the herd, pen or flock, economic outcome matters as much as clinical outcome, and the denominator is often the hardest number to get. The classic model of disease causation is the epidemiological triad of agent, host and environment. Disease follows when a sufficient agent meets a susceptible host in an environment that allows transmission. The triad is useful for teaching but too simple for chronic, production and multifactorial diseases such as mastitis or calf pneumonia, where the web of causation, a network of many interacting risk factors, is the better picture. A determinant is any factor, intrinsic (breed, age, sex, genotype) or extrinsic (housing, feed, climate, management, contact), that changes the probability of disease.
Several standard terms describe the pattern of occurrence, and examiners ask them precisely. Sporadic disease occurs irregularly and haphazardly, in single cases or small clusters with no predictable pattern; tetanus in a herd is the usual example. Endemic disease, called enzootic when it concerns animals, is present in a population at a usual, fairly constant and predictable frequency; it may be common or rare, but it is expected. Epidemic disease, epizootic in animals, is an occurrence clearly in excess of what is normally expected for that population, place and time. The word says nothing about absolute numbers: one case of a disease that has been absent for decades can be an epidemic, while hundreds of cases of an endemic infection may not be. Pandemic disease, panzootic in animals, is an epidemic spread across a very wide area, usually several continents; the spread of highly pathogenic avian influenza H5N1 in birds since 2020 is often described as a panzootic. Exotic disease is one that does not normally occur in a country or region and has been introduced, or may be introduced, from outside; foot-and-mouth disease is exotic to Australia and North America. An outbreak is a cluster of cases linked in time and place, and in WOAH usage it is the occurrence of one or more cases in an epidemiological unit. Emerging disease is a new infection or infestation, or a known one appearing in a new area or species or with a significant change in behaviour; re-emerging disease is a known disease returning after a decline.
Agents move by direct transmission (contact, bite, coitus, vertical transfer across the placenta or egg) or indirect transmission (vehicles such as feed, water, milk and fomites; biological or mechanical vectors; airborne spread over distance). The incubation period runs from infection to clinical signs; the latent period runs from infection to the start of infectiousness; the infectious period is the time during which the host can transmit. When the latent period is shorter than the incubation period, animals shed before they look ill, which is why such diseases are hard to control by removing sick animals. A carrier sheds the agent without showing disease. The basic reproduction number, R0, is the average number of secondary cases produced by one infectious individual in a wholly susceptible population. If R0 is above 1 the infection can spread; below 1 it dies out. The threshold proportion of a population that must be immune to stop sustained spread in a homogeneously mixing population is 1 minus 1/R0. The iceberg concept reminds us that clinical cases are the visible tip: below the waterline lie subclinical infections, carriers and exposed but uninfected animals. Surveillance that counts only clinical cases always underestimates infection. A reservoir is the population or environment in which an agent persists and from which it is transmitted to a target population; a maintenance host can sustain the infection on its own, while a spillover or dead-end host cannot.
A count alone, such as 12 sick cows, cannot be compared between herds of 20 and 2,000. It must be related to the population at risk. A proportion has the numerator included in the denominator and ranges from 0 to 1 (or 0 to 100%). A ratio divides one quantity by another that does not contain it, such as the ratio of males to females. A rate, in the strict sense, expresses change per unit of time, with animal-time in the denominator. Everyday veterinary language uses the word rate loosely for proportions (morbidity rate, mortality rate, case fatality rate); examiners accept the loose term but expect you to know which quantities are really proportions. Prevalence is the proportion of a population that has the disease at a specified time. Point prevalence refers to a single moment, period prevalence to an interval (all animals that had the disease at any time during it). Prevalence = number of existing cases divided by the number of animals in the population at that time. It is measured in cross-sectional surveys and is the right measure for planning workload and treatment needs, but it mixes the speed at which new cases arise with how long they last. Incidence measures new cases. Cumulative incidence, also called incidence risk or attack risk, is the proportion of initially disease-free animals that develop disease during a specified period: new cases divided by the number of disease-free animals at the start. It requires a closed population followed for the whole period and must always carry its time period (a risk of 10% in one month is very different from 10% in five years). Incidence rate, or incidence density, is new cases divided by the total animal-time at risk, for example 3 cases per 100 cow-years. It handles animals entering and leaving the population, and animals stop contributing time once they become cases. In a stable population where incidence and duration are roughly constant, prevalence is approximately incidence rate multiplied by average duration of disease, provided prevalence is low (below about 10%). The exact steady-state relation is P/(1-P) = I x D. This relation explains common exam paradoxes. A new treatment that keeps animals alive but does not cure them increases prevalence even though incidence has not changed. A disease that kills quickly, or is cured quickly, has low prevalence even when incidence is high, which is why acute outbreaks are poorly described by one cross-sectional survey. Chronic infections such as paratuberculosis, bovine tuberculosis and brucellosis accumulate: the prevalence is high relative to annual incidence because infected animals remain in the herd for years.
Morbidity rate is the proportion of the population at risk that shows clinical disease over a period; in outbreak reports it is usually cumulative incidence of clinical cases. Mortality rate is the proportion of the population at risk that dies of the disease (or of all causes, for crude mortality) during a period: deaths divided by population at risk. Case fatality rate (CFR) is the proportion of diseased animals that die of the disease: deaths divided by cases. It measures severity, not how common the disease is. The three are linked: mortality equals morbidity multiplied by case fatality, when all deaths come from cases in the same period. A common error is to divide deaths by the whole herd and call it case fatality; that is mortality. Attack rate is a cumulative incidence used for a short, defined outbreak period, often after a common exposure: the number of animals that became ill divided by the number exposed. Food-specific attack rates, calculated separately for animals that did and did not receive each feed, point to the vehicle in feed-borne outbreaks. The secondary attack rate is the number of cases among contacts of primary cases, divided by the number of susceptible contacts; primary cases are removed from the denominator. Proportional mortality is the proportion of all deaths attributable to one cause; it tells you nothing about risk, because the denominator is deaths, not animals. Crude rates compared across herds of different age structure mislead; specific (age-, breed-) or standardised rates remove that effect.
Analytical studies compare disease frequency in exposed and unexposed groups. The data are arranged in a two-by-two table: a = exposed and diseased, b = exposed and not diseased, c = unexposed and diseased, d = unexposed and not diseased. The relative risk (RR, risk ratio) is the risk in the exposed, a/(a+b), divided by the risk in the unexposed, c/(c+d). An RR of 1 means no association; above 1, the exposure is associated with increased risk; below 1, it is protective. RR can be calculated from cohort studies and trials, because they measure incidence in both groups. When incidence rates with animal-time denominators are compared, the result is an incidence rate ratio. In a case-control study the investigator chooses how many cases and controls to enrol, so the proportion diseased is fixed by design and risk cannot be calculated. The odds ratio (OR) is used instead: the odds of exposure among cases (a/c) divided by the odds of exposure among controls (b/d), which simplifies to ad/bc. The exposure odds ratio from a case-control study equals the disease odds ratio in the population. When the disease is rare (roughly below 5 to 10% in both groups), the OR approximates the RR. When disease is common, the OR is further from 1 than the RR. In matched-pair case-control studies, only discordant pairs carry information: OR = pairs with the case exposed and control unexposed divided by pairs with the case unexposed and control exposed. Cross-sectional data give a prevalence ratio or prevalence odds ratio.
Relative measures describe strength of association and are used to judge causation. Absolute measures describe how much disease an exposure causes, and they drive control decisions. Attributable risk (risk difference) is risk in exposed minus risk in unexposed: the excess risk among exposed animals attributable to the exposure, assuming a causal relationship. The attributable fraction in the exposed (etiologic fraction) is (RR - 1)/RR, or AR divided by the risk in the exposed: the proportion of disease in exposed animals that would be prevented by removing the exposure. Population attributable risk is the risk in the whole population minus the risk in the unexposed. The population attributable fraction (PAF) is the proportion of all cases in the population due to the exposure: PAF = Pe(RR - 1)/[1 + Pe(RR - 1)], where Pe is the proportion of the population exposed. A strong but rare risk factor can have a small PAF, while a weak but very common one can have a large PAF; control programmes should weigh both. For a protective exposure such as vaccination, the preventive fraction is 1 - RR; vaccine efficacy in a trial is (risk in unvaccinated minus risk in vaccinated) divided by risk in unvaccinated. A 95% confidence interval for an RR or OR that includes 1 is compatible with no association at the 5% significance level. Statistical significance is not biological importance.
Studies are descriptive or analytical. Descriptive studies (case reports, case series, surveys of prevalence) describe who, where and when, and generate hypotheses. Analytical studies test hypotheses about exposure and outcome. In observational analytical studies the investigator observes exposures that animals already have. A cross-sectional study measures exposure and disease at the same time in a sample of the population. It is quick and cheap and estimates prevalence, but it cannot establish that exposure preceded disease, and it over-represents long-lasting cases. A case-control study starts with disease: animals or herds with the disease (cases) are compared with those without (controls) for past exposures. It suits rare diseases and diseases with long induction periods, and can study many exposures at once, but it is prone to selection and recall bias and yields only odds ratios. Controls must come from the same source population that produced the cases. A cohort study starts with exposure: exposed and unexposed disease-free animals are followed forward (prospective) or through existing records (retrospective, historical) and incidence is compared. It establishes temporal sequence, gives incidence and RR, and can study several outcomes of one exposure, but it is expensive, slow, and inefficient for rare diseases. In experimental studies the investigator allocates the exposure. A randomised controlled trial, in which animals, pens or herds are randomly allocated to treatment or control and, ideally, owners and assessors are blinded, is the strongest single design for judging a treatment or vaccine. Laboratory trials control conditions tightly; field trials test the intervention in real farms. When the herd or pen is the unit of allocation, the analysis must account for clustering.
Random error arises from sampling variation and shrinks as the sample grows; it is quantified by confidence intervals and P values. Systematic error, or bias, does not shrink with sample size. Selection bias arises when the animals studied differ systematically from the target population in a way related to exposure and disease: for example, a referral hospital's cases are more severe than those in general practice, and slaughterhouse surveys miss animals that died on farm. Information (measurement) bias arises from inaccurate measurement of exposure or disease; misclassification of disease by an imperfect test is one form, and recall bias, where owners of sick animals remember exposures better, is another. Non-differential misclassification (equal in both groups) usually biases towards no association; differential misclassification can bias either way. Confounding is the mixing of the effect of the exposure with that of a third factor that is associated with the exposure and independently causes the disease, and is not on the causal pathway. Age often confounds: older cows are both more likely to be housed in a certain way and more likely to have mastitis. Confounding is controlled at the design stage by randomisation, restriction or matching, and at the analysis stage by stratification (Mantel-Haenszel methods) or multivariable models. Interaction (effect modification) is different: the effect of the exposure genuinely differs between levels of a third factor, and it should be reported, not removed.
Association is not causation. Koch's postulates, formulated for single infectious agents, fail for carriers, multifactorial diseases and agents that cannot be cultured. Evans's postulates and the Bradford Hill considerations are used instead. The main Hill considerations are strength of association, consistency across studies, specificity, temporality (the cause must precede the effect, and this one is essential), biological gradient (dose-response), plausibility, coherence with known facts, experimental evidence, and analogy. No single consideration except temporality is required, and none proves causation on its own. In practice, veterinary epidemiologists combine this reasoning with the component-cause model: a sufficient cause is a set of component causes that together produce disease, and a necessary cause is one that must be present in every sufficient cause. Mycobacterium bovis is necessary for bovine tuberculosis but not sufficient; host susceptibility, dose and contact are also needed. Removing any one component of a sufficient cause prevents the disease produced by that set, which is why control can succeed without knowing every cause.
Testing every animal (a census) is rarely affordable, so surveys test a sample and infer to the population. The target population is the one about which conclusions are wanted; the study population is the one actually sampled; the sampling frame is the list from which units are drawn, such as a register of herds or a list of animals in a herd. Errors in the frame (missing backyard flocks, out-of-date herd lists) cause bias that no sample size can fix. Probability sampling gives every unit a known, non-zero chance of selection and allows valid statistical inference. Simple random sampling draws units by random numbers from the full frame. Systematic sampling picks every kth unit after a random start, where k is the population size divided by the sample size; it is practical in a race or at a slaughter line but can be biased if the list has a periodic pattern. Stratified sampling divides the population into strata (region, herd size, species) and samples randomly within each; it improves precision and guarantees representation of each stratum. Cluster sampling selects groups (villages, herds) and tests all or some animals in each; it is cheaper in the field because travel is reduced, but animals in a cluster resemble each other, which reduces precision. The design effect expresses this loss: the ratio of the variance under cluster sampling to that under simple random sampling of the same size. Multistage sampling selects herds first and then animals within herds. Non-probability sampling (convenience, purposive, judgement) is common in practice, for example sampling animals at a market, and is legitimate for detecting disease or for risk-based surveillance, but it cannot give unbiased prevalence estimates.
For simple random sampling from a large population, the sample size needed to estimate prevalence is n = Z squared x P(1 - P) / d squared, where Z is the standard normal value for the chosen confidence (1.96 for 95%), P is the expected prevalence and d is the desired absolute precision (half the width of the confidence interval). If nothing is known about P, use 0.5, which gives the maximum sample size. Precision is absolute: a precision of 5% around an expected 20% means an interval from about 15% to 25%. When the sample is a substantial fraction of the population (usually more than about 10%), the finite population correction reduces it: n adjusted = n / (1 + n/N). For cluster designs the simple random sample size is multiplied by the design effect. With imperfect tests, the prevalence measured is apparent prevalence; true prevalence can be estimated as (AP + Sp - 1)/(Se + Sp - 1), the Rogan-Gladen estimator.
A different question is whether disease is present at all. Surveys to detect disease, or to substantiate freedom, specify a design prevalence (the minimum prevalence expected if the disease were present, chosen on biological grounds) and a confidence level. For a large population and a perfect test, the number that must be tested so that at least one positive is found with probability equal to the confidence level is n = ln(1 - confidence) / ln(1 - design prevalence). For 95% confidence and a design prevalence of 5%, about 59 animals are needed; for 1%, about 299. For small populations a hypergeometric formula is used and the numbers are smaller. If the test has sensitivity below 100%, more animals are needed, and if specificity is imperfect a single positive does not prove presence, so confirmatory testing is built in. Herd-level sensitivity, the probability of classifying an infected herd as positive, is 1 - (1 - p x Se) raised to the power n when n animals are tested in a herd with within-herd prevalence p and the herd is called positive if any animal tests positive.
A diagnostic test is judged against a reference standard (gold standard) that classifies animals as truly diseased or not. The results form a two-by-two table: true positives (TP), false positives (FP), false negatives (FN) and true negatives (TN). Diagnostic sensitivity (Se) is the proportion of truly diseased animals that test positive: TP/(TP + FN). Diagnostic specificity (Sp) is the proportion of truly non-diseased animals that test negative: TN/(TN + FP). These are properties of the test in a defined population and do not depend on prevalence, though they can vary with stage of infection, breed, age and the spectrum of disease in the evaluation population. Diagnostic sensitivity and specificity must not be confused with analytical sensitivity (the smallest amount of analyte the assay detects) and analytical specificity (freedom from cross-reaction). A highly sensitive test has few false negatives, so a negative result rules disease out; a highly specific test has few false positives, so a positive result rules disease in. For a test that gives a continuous value, such as ELISA optical density or skin thickness, moving the cut-off trades one for the other. A receiver operating characteristic (ROC) curve plots sensitivity against 1 - specificity for every cut-off; the area under the curve summarises overall discrimination, 0.5 being useless and 1.0 perfect. Youden's index, Se + Sp - 1, is one way to choose a cut-off. When no perfect reference exists, latent class models estimate Se and Sp from two or more tests in two or more populations. WOAH's Manual of Diagnostic Tests sets a validation pathway and asks that tests be fit for a stated purpose, such as demonstrating freedom, confirming clinical cases or estimating prevalence.
The clinician and the regulator ask a different question: given this result, what is the probability that the animal is truly diseased? The positive predictive value (PPV) is TP/(TP + FP) and the negative predictive value (NPV) is TN/(TN + FN). Predictive values depend strongly on prevalence. PPV = (P x Se) / [P x Se + (1 - P)(1 - Sp)]. At low prevalence even a very specific test yields many false positives relative to true positives, so PPV falls. This is the central problem of the end stage of eradication programmes: when a disease becomes rare, most reactors to a screening test are false positives, and confirmatory tests or herd investigation are needed before slaughter. NPV is high at low prevalence and falls as prevalence rises. Likelihood ratios combine Se and Sp into one number that does not depend on prevalence. LR+ = Se/(1 - Sp); LR- = (1 - Se)/Sp. Pre-test odds multiplied by the likelihood ratio give post-test odds. The apparent prevalence measured by a test is AP = P x Se + (1 - P)(1 - Sp).
Two tests can be interpreted in parallel, where an animal is positive if either test is positive, or in series, where it is positive only if both are positive (in practice, only animals positive on the first test receive the second). Assuming the tests are conditionally independent, parallel interpretation gives Se = 1 - (1 - Se1)(1 - Se2) and Sp = Sp1 x Sp2: sensitivity rises and specificity falls. It suits situations where missing a case is costly, such as exotic disease incursion or clearing an infected herd. Series interpretation gives Se = Se1 x Se2 and Sp = 1 - (1 - Sp1)(1 - Sp2): specificity rises and sensitivity falls. It suits confirming positives before costly action, as with a screening Rose Bengal test followed by a complement fixation test for brucellosis. Tests based on the same biological principle, such as two antibody tests, are usually conditionally dependent, and the real gain is smaller than the formula predicts. When there is no gold standard, agreement between two tests or two observers is measured by Cohen's kappa: kappa = (observed agreement - expected agreement) / (1 - expected agreement), where expected agreement is what would occur by chance given the marginal totals. Kappa of 0 means chance agreement only and 1 means perfect agreement; values between about 0.6 and 0.8 are commonly described as substantial, although the descriptive bands are conventions, not rules. Kappa measures agreement, not accuracy: two tests can agree well and both be wrong.
Surveillance is the systematic, ongoing collection, collation and analysis of information related to animal health and the timely dissemination of that information so that action can be taken; that is essentially the definition in the WOAH Terrestrial Animal Health Code. The final clause matters: data that do not reach a decision-maker in time are not surveillance. Monitoring is the ongoing collection of data on health and disease without a predefined action threshold, for example tracking the trend of mastitis in a region. Surveillance has four common purposes: early detection of exotic or emerging disease; demonstrating freedom from disease, usually for trade; measuring the level and distribution of endemic disease to plan control; and finding cases to support eradication. The design follows the purpose. Early detection needs high sensitivity and speed; freedom needs a documented probability of detection at a design prevalence; prevalence estimation needs representative sampling. Systems are evaluated against standard attributes: sensitivity, positive predictive value, timeliness, representativeness, simplicity, flexibility, acceptability, data quality and cost. These pull against each other: widening a case definition raises sensitivity but lowers PPV. Reporting must be linked to fair compensation; where culling without compensation is the only response, farmers stop reporting and sensitivity falls.
Passive surveillance relies on farmers, private veterinarians and laboratories to report cases they see. It is cheap and covers the whole country continuously, and it is the backbone of early detection, but it depends on awareness, trust and compensation, and it misses mild or unfamiliar disease. Active surveillance is initiated by the veterinary authority: planned surveys, testing at slaughter, herd visits. It is more complete for the target disease but costly and periodic. Targeted or risk-based surveillance concentrates effort on the subpopulations most likely to be infected or to spread infection, such as animals near a border, markets, animals with clinical signs, or imported stock; it gives more detection per unit cost but needs explicit weighting when used to claim freedom. Sentinel surveillance monitors selected units (sentinel herds, unvaccinated sentinel animals, sentinel chickens for arboviruses) that are expected to show infection early. Syndromic surveillance uses non-specific indicators such as drops in milk yield, abortions, feed intake or laboratory submissions to detect unusual patterns before a diagnosis is made. Participatory surveillance uses structured interviews, mapping and ranking with livestock keepers to find disease in pastoral or remote systems; it was used in the final stages of rinderpest eradication. Abattoir surveillance, laboratory-based surveillance and wildlife surveillance complete the picture.
An outbreak investigation follows a logical sequence, although in practice steps overlap and control measures often start at once. First, prepare: assemble the team, sampling kit, protective equipment and contacts with the laboratory and authorities. Second, establish that an outbreak exists by comparing current case numbers with expected numbers from records, and rule out artefacts such as a change in reporting or a new diagnostic test. Third, verify the diagnosis through clinical examination, post-mortem examination and laboratory confirmation. Fourth, write a case definition with criteria of time, place, animal and clinical or laboratory findings; start broad (suspect cases) and narrow later. Fifth, find and count cases systematically and record them in a line list, one row per animal or herd, with dates of onset, location, species, age, and exposures. Sixth, describe the outbreak by time, place and animal: the epidemic curve, a map, and attack rates by group. Seventh, develop hypotheses about source, mode of transmission and exposures, and test them with an analytical comparison, often a case-control or retrospective cohort study within the affected population. Eighth, implement control and prevention measures, which may include isolation, movement restriction, treatment, vaccination, culling, disinfection, vector control or removal of a contaminated feed. Ninth, communicate: report to the owner, the veterinary authority and, for listed diseases, through national channels to WOAH; and follow up to confirm that measures worked.
The epidemic curve is a histogram of new cases by date of onset, with a time interval of roughly a quarter to a third of the incubation period. Its shape suggests the type of outbreak. A point-source epidemic, in which animals were exposed to a common source over a short time, rises steeply and falls, with all cases within about one incubation period; the likely time of exposure can be estimated by counting back one average incubation period from the peak, or one minimum incubation period from the first case. A continuous or intermittent common-source epidemic rises and then plateaus or fluctuates while the source persists, such as a contaminated water supply. A propagated (progressive) epidemic, spread from animal to animal, shows successive peaks separated by about one incubation period (or one serial interval) and grows over several generations; foot-and-mouth disease and influenza spread this way. Mixed patterns occur when a common source seeds secondary spread. Spatial analysis complements the curve: a map of cases and farms may show clustering around a feed mill, along a river or downwind of an infected premises.
Once cases are counted, attack rates in animals exposed and not exposed to each suspected source are compared, and the relative risk for each exposure is calculated. The exposure with a high attack rate among exposed, a low attack rate among unexposed, and which accounts for most cases, is the likely vehicle. For instance, if lambs fed one batch of concentrate have a much higher attack rate than lambs not fed it, and most cases ate it, that batch is implicated, and the feed is withdrawn and sampled. In disease-control emergencies the veterinary authority acts under national legislation to declare infected and surveillance zones, stop movement, trace forward and backward (animals, products, vehicles and people that moved into or out of the premises during the risk period), and apply stamping out or vaccination according to the contingency plan. Tracing windows are set from the incubation period.
Countries that import animals and animal products must balance trade against the risk of introducing disease. Under the World Trade Organization Agreement on the Application of Sanitary and Phytosanitary Measures (the SPS Agreement), import measures must be based on international standards, or on a scientific risk assessment where a country chooses a higher level of protection. For animal health, the SPS Agreement names WOAH as the relevant standard-setting body, and the Terrestrial and Aquatic Codes contain the standards. The WOAH framework has four components. Hazard identification lists the pathogenic agents that could be associated with the commodity and could cause harm; if none is identified, the analysis stops. Risk assessment evaluates the likelihood and consequences of entry, establishment and spread. Risk management identifies, selects and implements measures (testing, quarantine, treatment, zoning, certification) to reduce risk to the importing country's appropriate level of protection. Risk communication is the interactive exchange of information and opinion throughout the process with all interested parties, and it begins at the start, not at the end.
Within risk assessment, WOAH describes four steps. Entry assessment (formerly release assessment) describes the biological pathways by which the commodity could carry the hazard into the importing country and estimates the likelihood. Exposure assessment describes the pathways by which susceptible animals or humans in the importing country could be exposed, and the likelihood. Consequence assessment describes the direct consequences (animal infection, production losses, public health) and indirect consequences (control costs, compensation, trade losses, environmental effects). Risk estimation integrates the three into an overall measure. Qualitative risk assessment expresses outcomes in words (negligible, low, medium, high) and is the most common in practice; it must be transparent about its reasoning. Quantitative assessment expresses outcomes numerically, often using stochastic models with probability distributions and Monte Carlo simulation, and must report uncertainty and variability. Food safety follows the Codex Alimentarius model with slightly different vocabulary: hazard identification, hazard characterisation (dose-response), exposure assessment and risk characterisation. On farms the same logic ranks entry pathways (purchased animals, vehicles, visitors, feed, wildlife) for biosecurity plans. Zoning and compartmentalisation, recognised in the WOAH Codes, allow trade from a disease-free zone or a biosecure compartment of a country that is not free as a whole.
A zoonosis is a disease or infection naturally transmitted between vertebrate animals and humans. Rabies is caused by rabies virus, a Lyssavirus in the family Rhabdoviridae; other lyssaviruses, mostly in bats, cause the same disease. Nearly all human deaths follow dog bites in Asia and Africa, and WHO estimates tens of thousands of human deaths each year. Wildlife cycles maintain the virus in foxes, raccoons, skunks, jackals, mongooses and bats in different regions, and vampire bats transmit it to cattle in Latin America. The virus enters through a bite or through saliva on broken skin or mucosa, travels along peripheral nerves to the central nervous system, and then spreads out to salivary glands. The incubation period is highly variable, usually weeks to a few months, and occasionally longer. Clinical rabies in animals appears in furious or dumb (paralytic) forms, and once signs appear it is almost invariably fatal. Diagnosis is post-mortem on brain tissue; the direct fluorescent antibody test is the long-standing reference test, with molecular tests and rapid immunodiagnostic tests also used. Negri bodies on histology are supportive but absent in a proportion of cases. A dog, cat or ferret that bites a person and is healthy is commonly observed for 10 days; if it remains healthy, it was not shedding virus at the time of the bite. Human post-exposure prophylaxis (PEP) combines immediate thorough washing of the wound with soap and water for about 15 minutes, vaccination, and rabies immunoglobulin for severe (WHO category III) exposures. Schedules differ by country and are set by national guidelines. Mass vaccination of dogs, reaching around 70% coverage in the dog population, is the method that eliminates dog-mediated rabies; the global Zero by 30 plan aims to end human deaths from dog-mediated rabies by 2030.
Brucellosis is caused by Gram-negative intracellular coccobacilli of the genus Brucella. B. abortus mainly infects cattle and buffalo, B. melitensis sheep and goats (and the most pathogenic for humans), B. suis pigs, and B. canis dogs; B. ovis causes epididymitis in rams and is not considered a significant zoonosis. In ruminants the placenta and fetal tissues are rich in erythritol, which favours growth, so infection causes abortion in late gestation, retained placenta, orchitis and epididymitis in males, and hygroma in cattle. Aborted fetuses, placentae and uterine discharges contain enormous numbers of organisms, and infected animals shed in milk. Humans are infected by consuming unpasteurised milk and dairy products, by contact with birth products (veterinarians, farmers, abattoir workers) and by laboratory exposure. Human brucellosis causes undulant fever, sweats, joint pain, spondylitis and orchitis, and can become chronic. Serological screening uses the Rose Bengal test and buffered antigen tests; the complement fixation test and ELISAs are used for confirmation; the milk ring test and milk ELISA screen herds through bulk milk; the fluorescence polarisation assay is another option. Culture is definitive. Live vaccines include strain 19 and RB51 for cattle and Rev.1 for small ruminants; they are given to young female stock, and strain 19 and Rev.1 can induce antibodies that interfere with serology and can infect humans. Control rests on vaccination where prevalence is high, test and removal where it is low, movement control, hygiene at calving, and pasteurisation. Treatment of animals is not recommended.
Bovine tuberculosis is caused mainly by Mycobacterium bovis, a member of the M. tuberculosis complex, in cattle, buffalo, deer and many other species. Transmission between cattle is mainly by aerosol; calves and humans can be infected by drinking raw milk. Before milk pasteurisation M. bovis caused a substantial share of human tuberculosis, especially extrapulmonary disease in children. M. bovis is intrinsically resistant to pyrazinamide, which matters when human cases are treated. Diagnosis in live cattle uses the tuberculin skin test, a delayed-type hypersensitivity reaction read at 72 hours. The single intradermal test injects bovine purified protein derivative (PPD) in the caudal fold or the neck. The single intradermal comparative cervical test injects bovine and avian PPD at separate sites on the neck and compares the increases in skin thickness, which improves specificity where environmental mycobacteria cause cross-reactions. Interpretation thresholds are set by national rules. The interferon-gamma assay is a blood test used in parallel to increase sensitivity. Anergic animals, often with advanced disease, may not react. Control is by test and slaughter, movement testing, meat inspection with traceback of lesions found at slaughter, and pasteurisation. Wildlife reservoirs complicate eradication: the European badger in the UK and Ireland, the brushtail possum in New Zealand, white-tailed deer in parts of the USA and African buffalo in southern Africa.
Anthrax is caused by Bacillus anthracis, a large Gram-positive, spore-forming rod. Virulence requires a capsule of poly-D-glutamic acid and a three-part toxin made of protective antigen, lethal factor and oedema factor. Spores survive in soil for decades, and outbreaks follow floods, droughts and soil disturbance in endemic areas. Herbivores are infected by ingesting spores and often die suddenly; carcasses show dark, unclotted blood from the natural openings, rapid bloating, absence of rigor mortis, and a greatly enlarged spleen if opened. A suspected anthrax carcass must not be opened, because vegetative bacilli sporulate on exposure to air. Diagnosis uses a blood smear stained with polychrome methylene blue, which shows square-ended bacilli with a pink capsule (the M'Fadyean reaction), together with culture and PCR in approved laboratories. Carcasses are disposed of by incineration or deep burial with disinfection, according to national rules. The live attenuated Sterne spore vaccine (a capsule-negative, toxin-producing strain) is used in endemic areas. Humans acquire cutaneous anthrax, by far the commonest form, through handling infected animals, hides and wool; inhalational anthrax (woolsorters' disease) and gastrointestinal anthrax from eating meat of infected animals are rarer and more often fatal.
Leptospirosis is caused by pathogenic spirochaetes of the genus Leptospira, classified into many serovars. Each serovar tends to have maintenance hosts, in which infection is chronic, renal and often mild, and incidental hosts, in which disease is severe. Rats maintain serovar Icterohaemorrhagiae, cattle Hardjo, pigs Pomona and Bratislava, and dogs Canicola, though these links are not exclusive. Leptospires are shed in urine and survive in warm, moist soil and fresh water, so human cases rise after heavy rains and floods. Occupational groups at risk include rice and sugarcane farmers, sewer workers, abattoir workers, dairy farmers and veterinarians. In animals the disease causes abortion, stillbirth, milk drop, haemolytic disease in calves and acute kidney and liver disease in dogs. In humans it ranges from a febrile illness to Weil's disease, with jaundice, kidney failure and haemorrhage, and to pulmonary haemorrhage. The microscopic agglutination test (MAT) using live leptospires of relevant serovars is the reference serological test; PCR detects the organism in blood early and in urine later. Control combines rodent control, vaccination of cattle, pigs and dogs with serovars relevant to the area, drainage, protective clothing and treatment of carriers. Glanders is caused by Burkholderia mallei, an obligate animal pathogen of horses, donkeys and mules that does not survive long in the environment. It causes nodules and ulcers in the nasal mucosa, pulmonary lesions, and a cutaneous form called farcy with nodules along lymphatics. Donkeys tend to have acute disease and horses chronic infection. The complement fixation test is widely used for trade and surveillance, and the mallein test is a hypersensitivity test. Glanders is notifiable to WOAH and is controlled by testing and humane destruction of positive animals rather than treatment. Human infection is rare but serious, and laboratory workers are at risk; B. mallei is a recognised potential biological weapon.
Kyasanur Forest disease (KFD) is a tick-borne viral haemorrhagic fever caused by a flavivirus. It was recognised in 1957 in the Kyasanur forest of Shimoga (Shivamogga) district, Karnataka, India, after deaths of monkeys. The main vector is the tick Haemaphysalis spinigera. Small mammals maintain the virus, while langurs and bonnet macaques are highly susceptible, die, and amplify infection for ticks; monkey deaths are the classic early warning. Humans are infected by tick bites while entering forests to collect firewood, graze cattle or work, mainly in the dry season. Humans are dead-end hosts. The disease has spread to neighbouring states of the Western Ghats. Control uses a formalin-inactivated vaccine for people in affected villages, tick repellents, protective clothing, investigation of monkey deaths and acaricide treatment around death sites. A closely related virus, Alkhurma haemorrhagic fever virus, occurs in Saudi Arabia. Japanese encephalitis (JE) is a mosquito-borne flavivirus infection of Asia and the western Pacific. The main vectors are Culex mosquitoes, especially Culex tritaeniorhynchus, which breed in irrigated rice fields. Ardeid water birds (herons and egrets) are maintenance hosts, and pigs are the key amplifying hosts, developing high viraemia without severe illness. Infection of pregnant sows causes abortion, mummified and stillborn piglets, and boars may become infertile. Humans and horses are dead-end hosts that can develop encephalitis. Most human infections are asymptomatic, but encephalitis is severe, with high fatality and lasting sequelae in survivors, mostly in children. Control relies on human vaccination, separating pig rearing from human settlements, vector control and, in some countries, vaccination of pigs and horses.
Avian influenza is caused by influenza A viruses (family Orthomyxoviridae), subtyped by the haemagglutinin (H1 to H16 in birds) and neuraminidase (N1 to N9) surface proteins. Wild aquatic birds, particularly ducks, geese and gulls, are the natural reservoir of low pathogenicity avian influenza (LPAI). Highly pathogenic avian influenza (HPAI) is defined by laboratory criteria in the WOAH Manual: high virulence in intravenous inoculation of chickens (an intravenous pathogenicity index above 1.2 in six-week-old chickens, or at least 75% mortality in four- to eight-week-old chickens), or, for H5 and H7 viruses, a haemagglutinin cleavage site sequence with multiple basic amino acids like those of known HPAI viruses. HPAI has so far arisen from H5 and H7 subtypes, which is why LPAI viruses of these subtypes in poultry are also a regulatory concern. HPAI in chickens causes sudden high mortality, drop in egg production, cyanosis and oedema of the comb and wattles, and haemorrhages. The goose/Guangdong H5 lineage, first detected in 1996, has evolved into many clades; clade 2.3.4.4b has spread across Europe, Asia, Africa and the Americas in wild birds and poultry since about 2020, reaching wild and marine mammals and, in the USA from 2024, dairy cattle. Human infections with H5N1 are sporadic, follow close contact with infected birds or animals, and have had a high reported case fatality, although mild cases may be under-counted; sustained human-to-human transmission has not been established. H7N9 caused human cases in China from 2013 with little disease in poultry. Control of HPAI includes biosecurity, surveillance of wild birds and poultry, rapid stamping out with zoning, movement control, and vaccination in some countries under surveillance plans able to distinguish infected from vaccinated birds. People handling infected birds need protective equipment.
One Health is an integrated, unifying approach that aims to sustainably balance and optimise the health of people, animals and ecosystems, recognising that they are closely linked and interdependent. That wording comes from the definition adopted in 2021 by the One Health High-Level Expert Panel advising FAO, UNEP, WHO and WOAH, which now work together as the Quadripartite and published a One Health Joint Plan of Action for 2022 to 2026. Most emerging infectious diseases of humans are zoonotic, and many originate in wildlife. The standard One Health priorities are zoonoses (rabies, brucellosis, avian influenza, anthrax), food safety, antimicrobial resistance (AMR), vector-borne disease under climate and land-use change, and environmental health. Examples of One Health in practice include joint animal-human investigation of anthrax or Rift Valley fever outbreaks, sharing of laboratory results between veterinary and public health services, dog vaccination as the most cost-effective way to prevent human rabies, integrated surveillance of antimicrobial use and resistance in humans, animals and food, and national One Health platforms or committees. The veterinarian's contribution is specific: controlling infection at its animal source is usually cheaper and more effective than treating people, and veterinary services protect the food chain.
Transboundary animal diseases (TADs) are, in the FAO definition, epidemic diseases that are highly contagious or transmissible, have the potential for very rapid spread irrespective of national borders, and cause serious socio-economic and possibly public health consequences. Foot-and-mouth disease, peste des petits ruminants, classical and African swine fever, highly pathogenic avian influenza, lumpy skin disease, contagious bovine pleuropneumonia, Rift Valley fever and sheep and goat pox are typical examples. Their spread is driven by the movement of animals, animal products, people and vehicles, informal cross-border trade, wildlife and, for vector-borne diseases, wind and climate. They threaten trade because infected countries lose access to markets. The Global Framework for the Progressive Control of Transboundary Animal Diseases (GF-TADs), a joint FAO-WOAH initiative, coordinates regional strategies. Rinderpest was declared globally eradicated in 2011, the first animal disease to be eradicated, and PPR is the target of a global eradication strategy aiming for 2030. The Progressive Control Pathway for FMD guides countries in stages from understanding the disease to demonstrating freedom.
The World Organisation for Animal Health (WOAH) was founded in 1924 as the Office International des Epizooties (OIE); it has been called the World Organisation for Animal Health since 2003 and has used the acronym WOAH since 2022; its headquarters are in Paris. WOAH publishes the Terrestrial Animal Health Code and Aquatic Animal Health Code (standards for trade and disease control) and the corresponding Manuals of Diagnostic Tests and Vaccines. Since 2005 there has been a single list of notifiable diseases, replacing the former Lists A and B. The criteria for listing, set out in the Terrestrial Code, are, in summary: international spread of the agent has been proven; at least one country has demonstrated freedom or impending freedom; reliable means of detection and diagnosis and a precise case definition exist; and either natural transmission to humans with severe consequences has been proven, or the disease causes significant morbidity or mortality in domestic animals, or it has been shown to cause significant morbidity or mortality in wildlife. Members must send an immediate notification through the World Animal Health Information System (WAHIS) within 24 hours for events such as the first occurrence of a listed disease in the country or zone, a recurrence after the disease was declared eradicated, the first occurrence of a new strain, a sudden unexpected change in distribution or increase in incidence or severity, or an emerging disease with significant morbidity, mortality or zoonotic potential. They then submit follow-up reports, six-monthly reports on the presence or absence of listed diseases, and annual reports. WOAH officially recognises the disease status of members or zones for a small number of diseases: foot-and-mouth disease, contagious bovine pleuropneumonia, African horse sickness, peste des petits ruminants, classical swine fever, and bovine spongiform encephalopathy risk status. It also endorses official control programmes for some of these. The Performance of Veterinary Services (PVS) Pathway evaluates national veterinary services against Code standards. National law then decides which diseases must be reported by owners and practitioners, and the list differs by country.
Learn the definitions exactly, because examiners test the difference between prevalence and incidence, mortality and case fatality, and epizootic and enzootic. Practise the two-by-two table until you can compute sensitivity, specificity, PPV, NPV, relative risk and odds ratio in under a minute and say which study design allows which measure. For each major zoonosis, memorise agent, main animal host, route to humans, key diagnostic test and the control method. When you see a clinical case in the teaching hospital, ask the population questions: how many others are at risk, how did it get here, and does it need to be reported?
Use epidemiology on farm visits: count cases against the number at risk before calling something an outbreak, draw an epidemic curve from the farm records, and compare attack rates between groups to find the likely source. Choose tests knowing that a positive at low prevalence is often false, and confirm before advising culling. Manage zoonotic risk for clients and staff: rabies exposure, brucellosis at calving, leptospirosis in dairies, psittacosis and avian influenza in birds. Know your country's notifiable disease list and the reporting route, and report suspicions promptly; delay costs more than a false alarm.
At this stage you design and defend programmes: sampling plans for regional prevalence surveys, risk-based surveillance for a border zone, test-and-slaughter schemes whose predictive values change as prevalence falls, and compensation rules that keep farmers reporting. You judge import risk assessments, set contingency plans for FMD, HPAI or ASF, run simulation exercises, and lead outbreak response teams with public health colleagues. Your written reports feed WOAH notifications and trade negotiations, so assumptions about design prevalence, test performance and zoning must be explicit and defensible.
Research questions here include estimating test accuracy without a gold standard using latent class models, building transmission and economic models to compare control strategies, spatial and network analysis of animal movements, evaluating surveillance sensitivity through scenario trees, and quantifying spillover at the wildlife-livestock-human interface. You also review and advise: critique study designs for bias and confounding, set standards for evidence of freedom, and bring epidemiological rigour to One Health programmes that span ministries. The best work turns into policy, so it must be transparent about uncertainty and honest about what the data cannot show.
This manual covers the core definitions, calculations and zoonoses in the syllabus but does not yet cover: statistical tests and regression modelling in depth (chi-square, logistic and Poisson regression, survival analysis); mathematical transmission models beyond R0; economic evaluation of disease control (cost-benefit, partial budgets); spatial statistics; detailed WOAH Code chapters for individual diseases; country-specific notifiable disease lists and legal procedures; human post-exposure prophylaxis schedules and drug doses, which must be taken from current national guidelines; and other important zoonoses such as Q fever, salmonellosis, hydatidosis, cysticercosis, toxoplasmosis, Rift Valley fever, Nipah and Crimean-Congo haemorrhagic fever, which are covered in veterinary public health and microbiology. The WOAH list of diseases, the list of diseases with official status and the rules on avian influenza reporting are revised periodically; check the current Terrestrial Code.
Original practice questions written to the syllabus and pattern of the exams named in each question's examTags. They are not past papers and do not reproduce any board's questions.
These notes are free to read here, with no account.
What the Laisha app adds:
Free for ten days, then a subscription. We would rather write that here than have you find it later.