Veterinary Epidemiology: Study Designs for Zoonotic Disease Research

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Key Takeaways

• Study design selection is dictated by the research question and disease natural history: Cross-sectional studies are optimal for estimating prevalence and identifying endemic patterns, while case-control designs are efficient for rare outcomes by retrospectively assessing exposures. Cohort studies are essential for quantifying incidence and understanding disease progression over time, particularly for occupational zoonoses.
• Appropriate measures of association are critical for valid inference: In cross-sectional studies with common outcomes (prevalence >10%), prevalence ratios (estimated via log-binomial or Poisson regression) are preferred over odds ratios, which tend to overestimate the association. Case-control studies inherently yield odds ratios, which approximate relative risk only when the outcome is rare.
• Zoonotic research faces unique challenges due to multi-species involvement and complex transmission pathways: Logistical difficulties in concurrent sampling of reservoir hosts, vectors, and humans necessitate careful consideration of sampling frames. Ecological niche modeling can complement field studies by mapping potential transmission risk areas, but its validity hinges on input data quality.
• Reporting standards like STROBE are crucial for transparency and critical appraisal: Adherence to guidelines ensures clear documentation of sampling strategies, response rates, and analytical choices, mitigating common failure modes such as selection bias and misclassification of exposure or outcome.
• Nested case-control and case-cohort designs offer efficient alternatives within established cohorts: These designs leverage banked biological samples to reduce laboratory costs while preserving the temporal logic of cohort studies, making them particularly useful for investigating biomarker-associated exposures or multiple outcomes from a single cohort.

Zoonotic diseases occupy a distinct position in veterinary epidemiology because the exposure pathway, the reservoir host, and the human outcome often require different sampling frames within a single investigation. This article reviews the principal observational study designs applied to zoonotic disease research in animal populations, with attention to their underlying logic, their common failure modes, and their appropriate interpretation. It serves veterinary researchers designing field studies, graduate students preparing protocols, and practitioners who must critically appraise published zoonotic literature.

The central question addressed is how a chosen study design shapes the validity of inferences about transmission, risk factors, and disease burden across species. Cross-sectional, case-control, and cohort designs each answer different questions, and each carries specific biases that become more consequential when the exposure involves an animal reservoir or an arthropod vector. The article also considers how ecological niche modeling and systematic review methods complement primary study designs in zoonotic research.

At a Glance

The Logic of Study Design Selection

The choice of study design follows from the natural history of the zoonosis and the research question. A newly emerged pathogen with unknown reservoir hosts demands a different approach than an endemic infection with well-characterized transmission. The investigator must first define the target population, the exposure of interest, and the outcome, then select a design that can measure their association with acceptable validity.

Cross-sectional studies measure exposure and outcome simultaneously in a defined population at one point in time. They are well suited to estimating seroprevalence of zoonotic pathogens across species and geographic regions. A systematic review of Rift Valley fever virus seroprevalence in Africa identified 174 studies across 31 countries from 1968 to 2016, with substantial variation by time, species, and country, and noted that few studies included both livestock and humans or livestock and wildlife in the same sampling frame. This limitation reflects a structural feature of cross-sectional zoonotic research: the logistical difficulty of sampling multiple host species concurrently.

The measure of association in cross-sectional studies deserves explicit attention. A review of veterinary cross-sectional studies with dichotomous outcomes found that odds ratios are frequently reported and often misinterpreted as risk ratios, particularly when disease prevalence is high. The authors recommend prevalence ratios as the more appropriate measure in this design, since odds ratios overestimate the prevalence ratio when the outcome is common. Logistic regression can be adapted to estimate prevalence ratios directly, and this choice should be made at the analysis planning stage instead of after data collection.

Case-Control Studies in Zoonotic Research

Case-control studies select participants based on outcome status and compare their exposure histories. This design is efficient for zoonoses with low incidence in the target population, where a cohort study would require impractically large sample sizes. The case definition must be explicit and consistently applied across species or diagnostic laboratories, and control selection should reflect the population that gave rise to the cases.

Selection bias is the dominant threat in case-control zoonotic studies. Controls drawn from veterinary teaching hospitals may not represent the exposure distribution of the source population, particularly for zoonoses with geographic or occupational clustering. Recall bias also operates when exposure ascertainment depends on owner or farmer memory, which is especially problematic for intermittent exposures such as tick bites or wildlife contact.

The odds ratio from a case-control study approximates the relative risk only when the disease is rare. For zoonoses with high prevalence in animal reservoirs, this approximation fails, and the investigator should consider whether a cross-sectional or cohort design would better serve the research question.

Cohort Studies and Incidence Estimation

Cohort studies follow a defined population over time, measuring exposure at baseline and observing incident outcomes. They provide the most direct estimate of disease risk associated with a zoonotic exposure and allow examination of multiple outcomes from a single exposure. The costs are substantial: long follow-up periods, attrition, and the need for large populations when the outcome is rare.

For zoonotic diseases, cohort designs are most feasible in production animal settings where populations are enumerated and follow-up is structured by management cycles. They are less practical for wildlife reservoirs, where individual identification and repeated sampling are difficult. Prospective cohorts also face the challenge that zoonotic exposures may change during follow-up, particularly for vector-borne pathogens where seasonal and climatic factors alter exposure intensity.

Ecological Niche Modeling as a Complementary Approach

Ecological niche modeling has become a standard tool for mapping zoonotic disease transmission risk by relating occurrence data to environmental variables. The approach can identify geographic areas where conditions support pathogen maintenance or vector populations, informing surveillance priorities and intervention planning. However, the validity of these models depends on the quality and type of input data, the justification of the study extent, and the choice of algorithm. Models built on biased occurrence data or poorly defined study areas can produce incomplete or incorrect inferences about disease distribution.

These models complement instead of replace traditional epidemiologic designs. They generate hypotheses about transmission risk that require field-based confirmation through cross-sectional or cohort studies, and they can help define sampling frames for those studies by identifying high-risk regions.

Systematic Review and Evidence Synthesis

Systematic review methods provide a structured approach to synthesizing zoonotic disease evidence across multiple studies. The PRISMA guidelines support transparent reporting of search strategies, inclusion criteria, and risk-of-bias assessment. A systematic review of alpha-gal syndrome following tick exposure identified 103 unique studies published from 2009 to 2020, of which 76.7 percent were case reports or case series. This distribution illustrates a common pattern in zoonotic research: early evidence accumulates through case reports, which document novel associations but cannot establish incidence or quantify risk. Systematic reviews make this evidence base explicit and identify where stronger study designs are needed.

The STROBE statement provides parallel guidance for reporting observational studies, and its use in veterinary zoonotic research supports critical appraisal of design and analysis choices.

Selecting the Primary Design: A Decision Framework

The choice between cross-sectional, case-control, and cohort designs for a zoonotic research question depends on three factors: the frequency of the outcome, the latency between exposure and outcome, and the resources available for follow-up. For a rare disease such as clinical leptospirosis in a low-prevalence dog population, a case-control design is usually the only feasible option because a cohort would require an impractically large sample to accrue enough cases. For a common outcome such as seropositivity to Rift Valley fever virus in endemic livestock, cross-sectional sampling provides a rapid prevalence estimate across species and regions, as demonstrated in a systematic review of RVFV seroprevalence in Africa from 1968 to 2016 systematic review of Rift Valley fever virus seroprevalence in livestock, wildlife and humans.

When the exposure is rare but the outcome is common, cohort enrollment from exposed populations is efficient. When both exposure and outcome are common, cross-sectional designs answer descriptive questions quickly but cannot establish temporal sequence. When the research question concerns incidence, recurrence, or time to event, only a cohort design with defined follow-up will suffice.

The table below summarizes the selection criteria for each design.

Cross-Sectional Surveys in Zoonotic Seroprevalence Work

Cross-sectional surveys dominate zoonotic seroprevalence research because they are logistically simple and can sample multiple species simultaneously. The RVFV review identified 174 seroprevalence studies across 31 African countries, yet only 8 of 126 articles included both livestock and human sampling systematic review of Rift Valley fever virus seroprevalence in livestock, wildlife and humans. That gap matters because single-species surveys cannot reveal transmission dynamics at the livestock-human interface.

A recurring analytical error in cross-sectional zoonotic studies is the reporting of odds ratios when prevalence ratios are the appropriate measure. When disease prevalence exceeds roughly 10 percent, the odds ratio overestimates the prevalence ratio, and the magnitude of overestimation grows with both prevalence and effect size. A review of veterinary cross-sectional studies found that many articles misinterpreted odds ratios as risk ratios, producing statements such as "the risk is X times greater" when the study design did not support that language odds ratio or prevalence ratio in veterinary cross-sectional studies. For zoonotic seroprevalence surveys where seropositivity is often common, direct estimation of prevalence ratios using log-binomial or Poisson regression with robust variance is preferred.

Sampling strategy determines external validity. Probability sampling with stratification by species, production system, and ecological zone is necessary when results will inform regional control decisions. Convenience sampling from diagnostic laboratory submissions overrepresents clinically affected animals and biases prevalence upward. The RVFV review noted that study design quality varied widely and that STROBE-guided evaluation identified frequent deficiencies in reporting of sampling frames and response rates systematic review of Rift Valley fever virus seroprevalence in livestock, wildlife and humans.

Case-Control Designs for Zoonotic Outbreak Investigation

Case-control studies are the workhorse of zoonotic outbreak investigation because they can be completed within days of case identification. The design is particularly valuable when the exposure is difficult to measure prospectively, such as tick exposure preceding alpha-gal syndrome. A systematic review of alpha-gal syndrome found that 76.7 percent of the 103 included studies were case reports or case series, with only a minority using analytic designs that could quantify exposure-outcome associations systematic review of tick exposures and alpha-gal syndrome. This distribution reflects the difficulty of mounting case-control studies for a syndrome with delayed onset and multiple possible tick vectors, but it also limits the strength of causal inference available from the literature.

Selection of controls is the principal threat to validity. Controls must be drawn from the same population that produced the cases, and they must have the same opportunity for exposure. In a veterinary zoonotic investigation, controls might be animals from the same herd, premises, or neighbourhood, depending on the hypothesised transmission route. For a foodborne zoonosis, controls should come from the same production system and supply chain as cases. For a vector-borne zoonosis, controls should be matched on geographic location and season because vector density varies across both.

Matching on too many variables can overadjust and obscure the exposure of interest. Matching on species and production system is usually appropriate, matching on age is appropriate when age is strongly associated with both exposure and outcome. Matching on the hypothesised exposure itself is never appropriate.

Cohort Studies for Zoonotic Transmission Dynamics

Cohort studies provide the strongest observational evidence for incidence and causal relationships in zoonotic disease research. They are essential when the research question concerns the rate of new infections, the duration of infectiousness, or the effect of an intervention over time. Prospective cohorts are particularly valuable for zoonoses with occupational exposure, such as brucellosis in abattoir workers or Q fever in dairy goat farmers, because exposure status can be measured before outcome onset.

The principal design decisions in a veterinary zoonotic cohort are the definition of the at-risk population, the frequency of follow-up sampling, and the handling of losses to follow-up. For livestock cohorts, individual animal identification and movement records are usually available through herd management systems, which facilitates follow-up. For wildlife cohorts, individual tracking is rarely feasible, and repeated cross-sectional sampling of a defined population may be the only practical approach.

Retrospective cohorts using existing diagnostic or production records are less costly but depend on the completeness and accuracy of historical data. A retrospective cohort is appropriate when reliable exposure records exist and when the outcome can be ascertained from diagnostic databases or slaughter records. The main risk is misclassification of exposure or outcome due to inconsistent record keeping.

Nested Case-Control and Case-Cohort Designs

Nested case-control studies arise within an established cohort when biological samples have been banked at enrollment. Cases are identified during follow-up, and controls are selected from cohort members who remained at risk at the time each case was diagnosed. This design preserves the temporal logic of the cohort while limiting laboratory costs to a fraction of the full cohort. It is particularly useful for zoonotic research when the exposure of interest is a biomarker, such as antibody titre, pathogen load, or genetic marker, that is expensive to measure on every cohort member.

The case-cohort design selects a random subcohort at baseline and compares all cases with this subcohort, regardless of when cases occur. It is more efficient than nested case-control when multiple outcomes are of interest because the same subcohort serves as the comparison group for every outcome. For zoonotic research examining several pathogens from the same cohort, case-cohort designs reduce laboratory costs substantially.

Both designs require careful definition of the risk set and attention to the sampling fraction. The odds ratio from a nested case-control study approximates the risk ratio when the outcome is rare, and it approximates the incidence rate ratio when controls are sampled from the risk set using incidence density sampling.

Documenting Design Decisions for Reproducibility

Reporting standards for observational veterinary research have improved, and journals increasingly expect adherence to structured reporting guidelines. The STROBE statement provides a checklist for reporting observational studies, and its use in the RVFV systematic review revealed that many studies failed to report basic elements such as sampling strategy, response rates, and handling of missing data systematic review of Rift Valley fever virus seroprevalence in livestock, wildlife and humans. For cross-sectional studies specifically, authors should report the prevalence ratio or odds ratio with a clear statement of which measure was used and why odds ratio or prevalence ratio in veterinary cross-sectional studies.

The study protocol should specify the target population, the sampling frame, the sample size calculation with its assumptions, the case definition and its validation, the exposure measurement and its quality control, and the planned statistical analysis. For zoonotic research, the protocol should also state how multi-species data will be handled, whether the analysis will be stratified by species, and how the One Health context will be addressed WHO One Health framework and CDC One Health and zoonotic disease resources.

Species differences change the correct design choice in specific ways. In livestock, individual-level cohort studies are feasible because animals are identified and tracked. In wildlife, population-level sampling with repeated cross-sectional surveys is often the only option. In companion animals, case-control studies are common because clinical caseloads provide cases but denominators for incidence are rarely available. In human populations linked to animal exposures, cohort studies of occupational groups are feasible but may be limited by mobility and loss to follow-up.

The choice of analytical method must match the design. Cross-sectional studies with common outcomes should use prevalence ratios. Case-control studies use odds ratios by design. Cohort studies report relative risks or incidence rate ratios. Using the wrong measure for the design produces biased estimates and misleading interpretations, particularly when the outcome is common odds ratio or prevalence ratio in veterinary cross-sectional studies.

Recognized Failure Modes in Zoonotic Study Designs

Several recurring design failures undermine the validity of zoonotic research. Selection bias dominates cross-sectional seroprevalence work when sampling frames favour accessible herds or clinically affected animals. The systematic review of Rift Valley fever virus seroprevalence in Africa identified marked heterogeneity in sampling strategies across 174 studies, with convenience sampling and non-random herd selection limiting generalizability of prevalence estimates Systematic literature review of Rift Valley fever virus seroprevalence. Detection of this failure requires comparing the study population's age, sex, and management profile against the target population, and scrutinising whether participation rates are reported.

Misclassification of exposure or outcome is a second major failure mode. In zoonotic disease work, diagnostic test imperfections compound this problem. Serological cross-reactivity between related pathogens, waning antibody titres, and the timing of sampling relative to exposure all produce misclassification that biases effect estimates toward the null. Early detection depends on reviewing the diagnostic test's sensitivity and specificity in the species under study, and on pilot-testing field sampling protocols.

A third failure mode is the inappropriate choice of effect measure. Cross-sectional studies with binary outcomes frequently report odds ratios when prevalence ratios are the more interpretable and appropriate measure, particularly when disease prevalence exceeds 10%. A review of veterinary cross-sectional studies found that odds ratios were commonly misinterpreted as risk ratios, leading to overestimation of associations Odds Ratio or Prevalence Ratio? An Overview of Reported. The discriminating check is simple: if the outcome is common, request prevalence ratios or log-binomial regression.

Common Errors and Corrective Actions

Less experienced investigators often conflate statistical association with causal inference in cross-sectional designs. Because exposure and outcome are measured simultaneously, temporal precedence cannot be established. The corrective action is to frame cross-sectional results as hypothesis-generating and to design follow-up cohort or case-control studies for causal confirmation.

A second frequent error involves the study base. In case-control studies, controls must be sampled from the same population that gave rise to the cases. Selecting controls from a referral hospital population when cases come from community surveillance introduces Berkson's bias. The corrective action is to define the source population explicitly before case ascertainment begins and to document the sampling frame for both groups.

A third error concerns sample size calculations that ignore clustering. Zoonotic pathogens often cluster at herd, flock, or household level. Analyzes that treat individual animals as independent when they are clustered within herds produce artificially narrow confidence intervals and inflated type I error rates. The corrective action is to account for clustering in both sample size calculations and statistical analysis, using multilevel models or generalized estimating equations.

Limitations of Current Evidence

The evidence base for many zoonotic diseases remains fragmented. The Rift Valley fever systematic review found that only 8 of 126 articles included seroprevalence data from both livestock and humans or livestock and wildlife, limiting inference about cross-species transmission dynamics Systematic literature review of Rift Valley fever virus seroprevalence. Similarly, the evidence linking tick exposures to alpha-gal syndrome consists predominantly of case reports and case series, which cannot establish incidence or quantify risk factors Tick exposures and alpha-gal syndrome: A systematic review of.

Expert opinion diverges on several methodological questions. Whether ecological niche modeling outputs should inform regulatory action remains contested. Some argue that model predictions, when built on biased occurrence data, can mislead instead of inform disease mapping Advances and Limitations of Disease Biogeography Using Ecological Niche. Others maintain that niche models, despite their limitations, provide the only feasible spatial risk estimates for data-poor zoonoses. The resolution lies in transparent reporting of input data quality and model assumptions.

Referral, Consultation, and Regulatory Reporting

Veterinarians encountering suspected zoonotic disease should escalate in three circumstances. First, when diagnostic capacity exceeds local laboratory capability, such as when molecular confirmation or specialised serology is required, referral to a reference laboratory is indicated. Second, when a disease has statutory reporting requirements, regulatory authorities must be notified without waiting for laboratory confirmation. The WOAH terrestrial animal health standards define notification obligations for listed diseases, and national authorities determine local reporting pathways WOAH Terrestrial Animal Health Code. Third, when a cluster of human illness is linked to animal exposure, public health authorities should be engaged through One Health coordination frameworks that link human and animal health surveillance WHO One Health Initiative.

Troubleshooting Table

Frequently Asked Questions

How Do I Choose a Study Design When Funding and Time Are Severely Limited?

When resources are constrained, a cross-sectional survey offers the most efficient route to generate prevalence estimates and hypothesis-generating associations. It requires a single point of contact with the study population and no follow-up period. If the outcome is rare or the exposure is uncommon, a case-control design becomes more efficient because it deliberately oversamples cases. For outbreak investigations, a retrospective case-control study can be completed within days if good records exist. Cohort studies are rarely feasible under tight budgets because they demand longitudinal follow-up and repeated sampling. Regardless of design, report the measure of association correctly. Cross-sectional studies should report prevalence ratios instead of odds ratios when the outcome is common, since odds ratios overestimate the association in that setting prevalence ratio reporting guidance in veterinary cross-sectional studies.

What Should I Do When Diagnostic Laboratory Capacity Is Unavailable in the Field?

Field conditions often preclude confirmatory testing. In that situation, define a clinical case definition that is explicit, repeatable, and validated against a reference standard where possible. Collect biological samples for later laboratory analysis, and store them according to the requirements of the intended assay. For seroprevalence surveys, dried blood spots on filter paper are a practical alternative to serum when cold chain access is unreliable. If you cannot confirm exposure status, classify it as unknown instead of negative, and plan a sensitivity analysis that tests whether misclassification would change your conclusions. The systematic review of Rift Valley fever virus seroprevalence studies noted that study design quality varied widely and that sampling strategies influenced reported prevalence, so document your constraints and their potential effect on validity systematic review of Rift Valley fever seroprevalence study designs.

How Does Study Design Differ When the Zoonosis Involves a Wildlife Reservoir?

Wildlife studies introduce sampling bias because detection probability varies by species, habitat, and behavior. Convenience sampling of hunted or trapped animals overrepresents accessible individuals and may miss the true spatial distribution of infection. Ecological niche modeling can complement field sampling by predicting suitable transmission areas from environmental variables, but the quality of those predictions depends on the quality and completeness of occurrence data advances and limitations of ecological niche modeling for disease mapping. For wildlife-linked zoonoses, consider a two-stage design: first use niche modeling to stratify sampling areas, then conduct targeted cross-sectional sampling within those strata. Account for clustering by herd, social group, or geographic site in both sample size calculations and analysis.

What Records Must I Keep to Make My Zoonotic Study Defensible?

Maintain a study protocol that specifies the target population, sampling frame, inclusion and exclusion criteria, and the primary outcome definition. Record every enrollment decision, including refusals and losses, with reasons. Keep a chain of custody log for all biological samples, noting collection time, storage temperature, and transport. Document laboratory methods and lot numbers for reagents. Preserve raw data in a format that allows re-analysis, and record any deviations from the protocol with dates and justification. For studies that inform regulatory action, the World Organization for Animal Health terrestrial animal health standards specify surveillance and reporting expectations that may apply to your jurisdiction WOAH terrestrial animal health standards. Consult those standards early, because retrospective reconstruction of records is rarely possible.

How Do I Explain Study Limitations to a Producer or Public Health Official Without Undermining Confidence?

Frame limitations as specific, bounded uncertainties instead of general weaknesses. State what the study can and cannot estimate, and give the direction and magnitude of likely bias. For example, if a seroprevalence survey used convenience sampling, say that the estimate may overrepresent high-risk groups and therefore cannot be generalized to the regional herd population. Distinguish statistical uncertainty, expressed as confidence intervals, from design-related uncertainty, which no statistical method can correct. Explain that the study was designed to answer a particular question and that other questions require different designs. The One Health framework explicitly connects animal and human health surveillance, so officials will expect you to state how your findings relate to human exposure pathways WHO One Health initiative.

When Is It Appropriate to Use a Case Series Instead of a Formal Comparative Design?

A case series is appropriate when a condition is newly recognized, extremely rare, or when the goal is to describe clinical presentation and generate hypotheses. The systematic review of alpha-gal syndrome found that most published evidence consisted of case reports and series, which established the clinical spectrum but could not quantify risk systematic review of tick exposures and alpha-gal syndrome. Use a case series to document novel manifestations, unusual species involvement, or suspected new exposure routes. Do not use a case series to estimate incidence, measure association, or evaluate intervention effectiveness. If you publish a case series, state explicitly that it cannot support causal inference and recommend the comparative design needed to test the hypotheses generated.

Related Clinical & Scientific Guides

• Wildlife Disease Surveillance: Designing and Implementing a One Health Program
• Biosecurity Risk Assessment for Livestock Operations: A Practical Framework
• Rabies Post-Exposure Prophylaxis in Veterinary Personnel

References and Further Reading

• Systematic literature review of Rift Valley fever virus seroprevalence in livestock, wildlife and humans in Africa from 1968 to 2016.. 2018.
• Advances and Limitations of Disease Biogeography Using Ecological Niche Modeling.. 2016.
• Odds Ratio or Prevalence Ratio? An Overview of Reported Statistical Methods and Appropriateness of Interpretations in Cross-sectional Studies with Dichotomous Outcomes in Veterinary Medicine.. 2017.
• A review of dicrocoeliosis of ruminants including recent advances in the diagnosis and treatment.. 2002.
• Tick exposures and alpha-gal syndrome: A systematic review of the evidence.. 2021.
• Pauci- and Multibacillary Leprosy: Two Distinct, Genetically Neglected Diseases.. 2016.
• WHO One Health Initiative. WHO.
• CDC One Health and Zoonotic Disease Resources. CDC.
• MSD Veterinary Manual, Professional Edition. MSD Veterinary Manual.

Related Articles

• Zoonotic Disease Risk Assessment in Veterinary Practice
• Zoonotic Disease Risk Communication in Veterinary Practice
• Zoonotic Disease Surveillance in Animal Shelters: Methods and Challenges
• Veterinary Public Health and Wildlife Trade: Risk Assessment
• Wildlife Rehabilitation and Zoonotic Disease Risk: Protocols for Veterinary Professionals

This article is educational professional reference material for veterinary audiences. It is not a substitute for veterinary diagnosis, individual clinical judgment, current product labeling, or applicable regulatory requirements.