# Evaluating Veterinary Surveillance System Attributes


## Key Takeaways

- **Systematic evaluation is paramount for reliable veterinary surveillance outputs**, which inform disease detection, trade, and policy; evaluations must be recurring processes, not one-off audits, to ensure ongoing system efficacy.
- **Core attributes of sensitivity, specificity, timeliness, and representativeness are critical for assessing surveillance system performance**, with sensitivity measuring the proportion of true events detected (e.g., a slaughterhouse inspection contributing 55.6% sensitivity for bovine tuberculosis detection) and timeliness assessing the delay between event occurrence and usable output.
- **The balance between sensitivity and specificity is a fundamental trade-off**, where increasing sensitivity (e.g., by broadening case definitions) often reduces specificity, generating more false alerts that consume resources and erode confidence.
- **Representativeness ensures data accurately reflect the target population**, necessitating consideration of sampling frames (e.g., avoiding reliance solely on large commercial herds) and reporting biases (e.g., passive systems over-representing clinician-recognized diseases).
- **Evaluation design must align with explicit surveillance objectives**, defining system boundaries, target hazards, data collection, and intended users to avoid measuring irrelevant metrics and ensure defensible conclusions.
- **Common failure modes include silent sensitivity loss** due to declining diagnostic capacity or submission rates, **timeliness degradation** from workflow congestion, and **representativeness failures** marked by geographic clustering of reports not aligned with disease distribution.

---

Veterinary surveillance systems generate data that inform disease detection, trade decisions, and population health policy. Their outputs are only as reliable as the systems that produce them, which is why structured evaluation matters. This article explains how to assess veterinary surveillance systems using standard attributes, with emphasis on sensitivity, timeliness, and representativeness. It serves veterinary researchers and epidemiologists who design, operate, or audit surveillance programs and need a defensible framework for judging system performance. The content is cross-species and applies to national programs, regional schemes, and production-sector monitoring alike. Specific surveillance methods such as syndromic or risk-based approaches are not covered here.

Evaluation is not a single event but a recurring process. A systematic review of animal and public health surveillance systems found that most published evaluations addressed only one or two attributes and that comprehensive assessments were uncommon, in part because surveillance objectives were often not stated clearly enough to guide attribute selection. That finding has a practical consequence: an evaluation that begins without explicit objectives risks measuring the wrong things and producing misleading conclusions. The same review identified 23 distinct attributes used across 101 evaluated systems, which illustrates both the breadth of possible assessment targets and the need for a structured approach to choosing among them.

## At a Glance

| Attribute | Definition | Primary Question It Answers |
|---|---|---|
| Sensitivity | Proportion of true events detected by the system | How much disease is being missed? |
| Specificity | Proportion of non-events correctly excluded | How many alerts are false? |
| Timeliness | Delay between event occurrence and system output | How quickly can action be taken? |
| Representativeness | Degree to which system data reflect the target population | Do the data describe the population of interest? |
| Simplicity | Ease of operation and data flow | Can the system be sustained? |
| Flexibility | Ability to adapt to changing needs | Can the system handle new threats? |
| Acceptability | Willingness of stakeholders to participate | Will reporting continue? |
| Stability | Reliability of operation over time | Does the system function when needed? |
| Data quality | Completeness and accuracy of records | Can the data be trusted? |

## The Purpose of Surveillance Evaluation

Surveillance systems exist to answer questions about disease occurrence, distribution, and change. Evaluation determines whether a system answers those questions well enough to justify its cost. The World Organization for Animal Health (WOAH) publishes international standards for animal health surveillance that frame evaluation as a component of ongoing system management, not a one-off audit. These standards emphasize that surveillance should be evaluated regularly to ensure it provides valuable information efficiently.

The evaluation process itself follows a recognizable structure. A systematic review of existing evaluation approaches identified four common steps: defining the system under evaluation, designing the evaluation process, implementing the evaluation, and drawing conclusions and recommendations. The first step is frequently the most neglected. A system cannot be evaluated meaningfully until its boundaries are clear: what population it covers, what hazards it targets, what data it collects, and who acts on its outputs.

## Core Attributes and Their Interrelationships

### Sensitivity

Sensitivity in surveillance evaluation refers to the ability of a system to detect the events it is designed to capture. This includes both the sensitivity of individual diagnostic components and the sensitivity of the system as a whole, which depends on reporting behavior, sampling design, and case definition. For rare or emerging diseases, sensitivity is particularly demanding. Scenario tree modeling has been used to quantify the sensitivity of passive and active surveillance components separately, allowing designers to combine components into a cost-effective whole. In one worked example for bovine tuberculosis in Switzerland, slaughterhouse inspection contributed a sensitivity of 55.6 percent while passive clinical surveillance and human surveillance each contributed less than one percent, a disparity that would be invisible without component-level analysis.

### Specificity

Specificity concerns the proportion of non-events that the system correctly excludes. Low specificity generates false alerts that consume diagnostic resources and erode stakeholder confidence. The tension between sensitivity and specificity is fundamental: increasing sensitivity by broadening case definitions or testing more animals will typically reduce specificity. Evaluators must judge whether the balance suits the system's purpose. A system designed for early detection of a zoonotic emergency may tolerate lower specificity than one used for routine certification of disease-free status.

### Timeliness

Timeliness measures the interval between a relevant event and the point at which the system produces a usable output. The relevant interval depends on the disease and the decision it informs. For influenza virus surveillance, the WHO Global Influenza Surveillance and Response System coordinates year-round collection of virological and epidemiological data to support vaccine strain selection, a process where delays of weeks can compromise the entire production cycle. For a slow-moving endemic disease, timeliness requirements are less demanding. Evaluators should define the critical time interval before measuring it, instead of reporting raw delays without context.

### Representativeness

Representativeness describes how accurately the system's data reflect the true distribution of the target condition in the target population. A system that samples only large commercial herds will not represent smallholder populations. A passive system that relies on voluntary reporting will over-represent diseases that clinicians recognize and under-represent those they do not. Representativeness is often the hardest attribute to verify because it requires knowledge of the underlying population that the surveillance system itself is meant to characterize.

## Designing the Evaluation

The evaluation design determines which attributes can be assessed credibly and what resources the process will consume. A structured design begins with a written evaluation brief that restates the surveillance objectives, identifies the intended users of the evaluation findings, and lists the attributes to be assessed. The brief should also specify the time window for data review, the geographic scope, and the species or production sectors covered. Without this framing, attribute selection risks reflecting convenience instead of surveillance purpose, a problem noted in systematic reviews of animal and public health surveillance evaluation [Drewe et al., systematic review of surveillance system attributes](https://pubmed.ncbi.nlm.nih.gov/22074638/).

The evaluation approach should match the question being asked. A rapid internal review suits routine performance monitoring, while an external evaluation with independent data collection suits formal program review or trade-related certification. Common evaluation steps include defining the system under evaluation, designing the evaluation process, implementing it, and drawing conclusions and recommendations [Calba et al., systematic review of surveillance evaluation approaches](https://pubmed.ncbi.nlm.nih.gov/25928645/). Each step should be documented in the evaluation file so that the basis for each attribute score is traceable.

## Attribute Scoring Framework

A scoring framework converts qualitative judgments into comparable scores. The framework below uses a five-point scale for each attribute, with explicit anchors at scores 1, 3, and 5. Intermediate scores of 2 and 4 are permitted when the system meets some but not all criteria for the adjacent anchor.

| Score | Sensitivity | Timeliness | Representativeness |
|-------|-------------|------------|---------------------|
| 1 | Detects only clinically obvious cases in high-risk populations, no active component | Reporting interval exceeds the action interval for the target disease by more than 2-fold | Sampling restricted to one geographic region or one production type with no documented coverage assessment |
| 2 | Partial active surveillance present, detection limited to certain production classes | Reporting interval exceeds the action interval by up to 2-fold | Coverage assessed but gaps identified in more than one population stratum |
| 3 | Active and passive components both contribute, detection probability estimated for target disease | Reporting interval meets the action interval for the target disease | Coverage documented across major geographic and production strata, gaps quantified |
| 4 | Detection probability quantified with scenario tree or equivalent method, sensitivity above 90% for target prevalence | Reporting interval allows intervention within one transmission cycle | Coverage documented with periodic reassessment, sampling frame updated |
| 5 | Sensitivity quantified and optimized across components, uncertainty bounds reported | Reporting interval permits intervention within a fraction of the transmission cycle | Coverage verified against an external census or registry, bias assessed |

The framework assumes the evaluator has defined the target disease, the action interval, and the target prevalence before scoring. These parameters should appear in the evaluation brief. For rare or emerging diseases, sensitivity estimation may require scenario tree modeling because field prevalence data are unavailable [Hadorn and Stärk, evaluation of surveillance for rare and emerging infectious diseases](https://pubmed.ncbi.nlm.nih.gov/18651991/).

## Data Sources and Verification

Each attribute score requires evidence. Sensitivity scoring draws on detection probability estimates, diagnostic test performance data, and reporting rate assessments. Timeliness scoring requires dated records from sample collection, laboratory submission, result reporting, and case notification. Representativeness scoring requires the sampling frame, coverage maps, and population denominator data.

Verification steps include:

- Compare laboratory accession logs against case reports to estimate the proportion of diagnosed cases that enter the surveillance database.
- Review a sample of submitted specimens to confirm that diagnostic tests match the case definition.
- Check reporting dates against laboratory report dates to measure the interval between diagnosis and notification.
- Compare the distribution of sampled animals against census data for age, sex, geographic location, and production type.

Where records are incomplete, the evaluator should state the limitation and adjust the attribute score accordingly. A system with excellent laboratory capacity but incomplete field reporting cannot receive a high sensitivity score.

## Documenting Findings

The evaluation report should present each attribute score, the evidence supporting it, and the limitations of that evidence. A standardized template improves comparability across evaluations and over time.

### Sample Evaluation Report Template

**Evaluation title:**
**Evaluator and affiliation:**
**Evaluation dates:**
**Surveillance system name and owner:**
**Target disease(s):**
**Surveillance objectives (as stated):**

| Attribute | Score (1-5) | Evidence reviewed | Limitations | Recommended action |
|-----------|-------------|-------------------|-------------|---------------------|
| Sensitivity | | | | |
| Specificity | | | | |
| Timeliness | | | | |
| Representativeness | | | | |
| Data quality | | | | |
| Stability | | | | |

**Overall assessment:**
**Priority recommendations (ranked):**
**Resources required for recommended actions:**
**Suggested timeline for re-evaluation:**

The template serves cross-species and cross-production applications. For wildlife surveillance, the population denominator may be uncertain and the representativeness score should reflect that uncertainty. For food animal production systems, timeliness thresholds should align with slaughter and marketing cycles. For companion animal surveillance, reporting completeness may depend on practitioner engagement instead of regulatory mandate.

## Species and Production System Adjustments

The correct evaluation approach varies with the surveillance context. In intensive poultry production, flock-level reporting and rapid laboratory confirmation dominate, and timeliness is measured in hours to days. In extensive beef cattle systems, individual animal identification may be incomplete, and representativeness scoring must account for the sampling frame limitations. In wildlife, detection often relies on hunter harvest reports or carcass submissions, and sensitivity is typically lower than in domestic populations.

The evaluation brief should specify which production system or species the evaluation covers. A single evaluation spanning multiple species may require separate attribute scores for each species stratum, because a system that is highly sensitive for cattle may be nearly insensitive for small ruminants. The same principle applies across regions within a country, where laboratory access and reporting infrastructure differ.

## Common Failure Modes

Evaluations fail in predictable ways. The most common failure is scoring attributes without linking them to surveillance objectives, which produces misleading results [Drewe et al., systematic review of surveillance system attributes](https://pubmed.ncbi.nlm.nih.gov/22074638/). A second failure is relying on a single data source, such as laboratory records, when field reporting data would change the score. A third is treating attribute scores as static, when surveillance systems change with personnel turnover, funding shifts, and disease emergence.

The evaluator should also distinguish between the surveillance system and the diagnostic tests it uses. A highly sensitive diagnostic test does not make a surveillance system highly sensitive if specimens are rarely collected. Conversely, a moderately sensitive test combined with high submission rates can produce acceptable system sensitivity. The evaluation assesses the system, not its components in isolation.

Where the evidence base is limited, the report should say so directly. Quantifying sensitivity for a disease with no recent incursions requires modeling assumptions that should be stated. Timeliness data may be missing for historical periods. These gaps do not invalidate the evaluation, but they should be visible to the reader.

## Recognized Complications and Failure Modes

Surveillance systems fail in characteriztic ways, and early detection of these failures depends on monitoring operational indicators instead of waiting for formal re-evaluation. The most common complication is silent sensitivity loss, where the system continues to generate reports but detects a decreasing fraction of true cases. This occurs when diagnostic capacity declines, when submission rates fall, or when case definitions drift from the original standard. Early detection requires tracking the ratio of suspect to confirmed cases, the interval between sample collection and laboratory result, and the number of participating reporting units over time. A falling suspect-to-confirmed ratio may indicate that reporting has become overly specific, while a rising ratio with stable confirmations suggests that the system is being overwhelmed by false alarms.

Timeliness degradation is frequently insidious. Reporting intervals that were acceptable at system design may lengthen gradually as staffing changes or as laboratory workflows become congested. The discriminating check is to maintain a running median of the interval from disease onset to report receipt, stratified by reporting source, and to investigate any month where the median exceeds the previous twelve-month average by more than one reporting cycle. Representativeness failures typically emerge as geographic or demographic clustering of reports that mirrors reporting effort instead of disease distribution. Comparing report density against the underlying population at risk, using census or production data, will expose this pattern.

The table below summarizes common failure modes and the checks that discriminate between them.

| Observation | Likely cause | Discriminating check |
|---|---|---|
| Falling suspect-to-confirmed ratio | Case definition drift or increased false-positive reporting | Audit recent cases against the original case definition |
| Stable report volume with declining confirmed cases | Diagnostic sensitivity loss or laboratory degradation | Review laboratory quality assurance records and test performance |
| Increasing median onset-to-report interval | Timeliness erosion in one reporting stream | Stratify intervals by reporting source and compare against baseline |
| Geographic clustering of reports | Representativeness failure or true outbreak | Compare report density to population-at-risk distribution |
| Rising proportion of incomplete submissions | Data quality decline or reporter fatigue | Track completeness scores by submitting practice or region |
| No reports for an extended period | Passive surveillance collapse or true absence | Verify reporting channels remain functional and known to reporters |

## Common Errors in Evaluation Practice

Less experienced evaluators frequently conflate system activity with system performance. A surveillance system that generates many reports is not necessarily sensitive, timely, or representative, and evaluation must assess each attribute independently against stated objectives. The corrective action is to return to the surveillance objectives and to score each attribute against those objectives instead of against report volume.

A second recurring error is the selection of attributes before the evaluation questions are fixed. The systematic review by Drewe and colleagues found that surveillance objectives were often not stated in published evaluations, which made the reasons for choosing particular attributes unclear and created potential for misleading results [Drewe et al., systematic review of surveillance system attributes](https://pubmed.ncbi.nlm.nih.gov/22074638/). The corrective action is to document the evaluation questions first, then select attributes that map directly to those questions, and only then choose measurement methods.

A third error is treating a single attribute score as a summary of system value. Evaluations that address only one or two attributes are common, but comprehensive assessment requires attention to the relationships between attributes, since improving sensitivity may reduce timeliness or increase cost [Calba et al., systematic review of surveillance evaluation approaches](https://pubmed.ncbi.nlm.nih.gov/25928645/). The corrective action is to present attribute scores as a profile instead of a single index and to discuss trade-offs explicitly.

Students and early-career evaluators also tend to rely on passive surveillance data without verifying its completeness. Under-reporting is a well-recognized limitation of passive systems, and cost-intensive active surveillance is often required to raise sensitivity to acceptable levels [Hadorn and Stärk, evaluation of surveillance for rare and emerging diseases](https://pubmed.ncbi.nlm.nih.gov/18651991/). The corrective action is to quantify the sensitivity contribution of each surveillance system component separately, using methods such as scenario tree modeling, instead of treating the system as a single undifferentiated stream.

## Limitations of the Evidence and Areas of Expert Disagreement

The evidence base for veterinary surveillance evaluation remains fragmented. The systematic review by Drewe and colleagues identified 23 different attributes assessed across 101 surveillance systems, but most evaluations addressed only one or two attributes and comprehensive evaluations were uncommon [Drewe et al., systematic review of surveillance system attributes](https://pubmed.ncbi.nlm.nih.gov/22074638/). This fragmentation means that comparative judgments about which attributes matter most, and how they should be weighted, rest on limited empirical support.

Expert opinion differs on several points. There is no consensus on whether sensitivity or timeliness should take priority when resources are constrained, and the answer depends heavily on the disease under surveillance and the consequences of delayed detection. For rare and emerging diseases, the priority is early detection at very low prevalence, which places a premium on sensitivity even when the cost per case detected is high [Hadorn and Stärk, evaluation of surveillance for rare and emerging diseases](https://pubmed.ncbi.nlm.nih.gov/18651991/). For endemic diseases with established control programs, timeliness and representativeness may matter more for monitoring trends and evaluating interventions.

A further area of disagreement concerns the standardization of evaluation methods. Some authorities advocate for a generic evaluation framework applicable across species and production systems, while others argue that evaluation must be tailored to the specific surveillance objectives and context. The international standards published by the World Organization for Animal Health provide a common reference point for member countries, but they do not resolve the question of how to weight competing attributes in a specific evaluation [WOAH animal health surveillance standards](https://www.woah.org/en/what-we-do/animal-health-and-welfare/disease-data-collection/).

## Escalation and Referral Criteria

Referral to specialist consultation is warranted when the evaluation reveals a problem that exceeds the evaluator's expertise or when the consequences of getting the answer wrong are severe. Quantitative sensitivity analysis using scenario tree modeling or Bayesian methods is a common trigger for specialist involvement, since these techniques require statistical expertise that generalizt evaluators may not possess. Laboratory involvement is indicated when diagnostic performance is suspected to be a limiting factor, since test sensitivity and specificity must be verified under local conditions instead of assumed from published values.

Regulatory reporting obligations vary by jurisdiction and by disease. Where a surveillance system detects a notifiable disease, the reporting pathway is defined by the relevant animal health authority, and the evaluation should verify that this pathway functions correctly as part of the timeliness assessment. The Terrestrial Animal Health Code of the World Organization for Animal Health sets international standards for notification and reporting, and evaluations of systems that support international trade should reference these standards explicitly [WOAH Terrestrial Animal Health Code](https://www.woah.org/en/what-we-do/standards/codes-and-manuals/terrestrial-code-online-access/).

Escalation is also appropriate when the evaluation identifies a risk to public health, since many emerging infectious diseases of animals are zoonotic. In such cases, coordination with public health surveillance authorities is necessary, and the evaluation should assess the interface between animal and human surveillance components instead of treating them in isolation.

## Frequently Asked Questions

### How Do I Evaluate a Surveillance System When Budget or Staff Time Is Severely Limited?

Prioritize attributes that map directly to the system's stated objectives. If the system exists to detect incursions early, sensitivity and timeliness outrank completeness of case follow-up. Use existing data instead of collecting new data. Slaughterhouse inspection records, laboratory submission logs, and necropsy reports often support a retrospective assessment of sensitivity without prospective effort. Scenario tree modeling can quantify component sensitivity using historical data and expert opinion, which is less resource-intensive than field trials. Document explicitly which attributes were not assessed and why. The [systematic review of surveillance system attributes](https://pubmed.ncbi.nlm.nih.gov/22074638/) notes that most published evaluations address only one or two attributes, so a focused evaluation is consistent with current practice. State the limitation in the final report.

### What Should I Do When the Ideal Diagnostic Test or Sampling Frame Is Unavailable?

Use the best available test and adjust your interpretation accordingly. If confirmatory testing is unavailable, rely on case definitions that combine clinical signs with epidemiological linkage. Report sensitivity and specificity estimates as ranges instead of point values when test performance is uncertain. Bayesian methods can estimate test characteriztics without a gold standard, but they require prior information and careful model specification. For sampling, convenience samples may be the only feasible option. Analyze them for selection bias and state the direction of expected bias. The [WOAH animal health surveillance standards](https://www.woah.org/en/what-we-do/animal-health-and-welfare/disease-data-collection/) require that surveillance outputs be fit for purpose, and a biased sample may still support a negative finding if the bias favours detection.

### How Does the Evaluation Approach Differ Between Companion Animal and Production Animal Systems?

Companion animal surveillance relies heavily on passive reporting from private practitioners, so representativeness depends on practitioner participation and submission behavior. Evaluate the proportion of practices submitting and the geographic coverage of submissions. Production animal systems often combine mandatory reporting, slaughterhouse inspection, and active testing, allowing component-specific sensitivity estimates. The [scenario tree approach for rare disease surveillance](https://pubmed.ncbi.nlm.nih.gov/18651991/) was developed for bovine tuberculosis in Switzerland and illustrates how slaughterhouse inspection, clinical surveillance, and human surveillance can be modelled as separate components. Cost structures differ markedly. In production systems, trade consequences drive the economic analysis. In companion animals, the primary costs are diagnostic testing and practitioner time. Tailor the evaluation questions to the system's purpose and the species' epidemiology.

### What Records Should I Keep to Make Future Evaluations Easier?

Maintain a surveillance system description that states objectives, case definitions, population at risk, data sources, and reporting pathways. Record the date each component was implemented and any subsequent changes to case definitions or laboratory methods. Keep denominator data, including population estimates and the number of premises or animals under observation. Archive raw laboratory results, also summary reports. Document the number of submissions received, the number tested, and the number positive by species and geographic region. Record the time between key events, such as sample collection, laboratory receipt, and result reporting. The [CDC principles of epidemiology](https://www.cdc.gov/csels/dsepd/ss1978/index.html) emphasize that surveillance data are only as useful as the documentation supporting their interpretation. Store these records in a format that remains accessible after staff changes.

### How Do I Explain Evaluation Findings to a Practice Owner or Government Supervisor?

Frame the report around decisions the supervisor must make. State whether the system meets its objectives, where it falls short, and what specific changes would improve performance. Quantify the trade-off between sensitivity and cost where possible. For example, increasing slaughterhouse inspection sensitivity from 55 to 80 percent may require specific investments in training or testing frequency. Use the [systematic review of evaluation approaches](https://pubmed.ncbi.nlm.nih.gov/25928645/) to justify the evaluation structure you used. Present limitations honestly but briefly. Supervisors need to know whether a negative finding is reliable and whether a positive finding is likely to be a true detection. Offer a prioritized action list with estimated costs and expected benefits for each item.

### When Should I Seek External Expertise for a Surveillance Evaluation?

Seek external input when the evaluation will inform trade negotiations, when legal or regulatory consequences follow from the findings, or when internal staff have a direct stake in the outcome. Independence matters when the system's performance reflects on the people conducting the evaluation. External expertise is also warranted when the statistical methods exceed local capacity, such as scenario tree modeling, Bayesian analysis, or complex sampling designs. The [WOAH terrestrial animal health code](https://www.woah.org/en/what-we-do/standards/codes-and-manuals/terrestrial-code-online-access/) sets international expectations for surveillance quality, and an evaluation used to support claims of freedom from disease should meet those standards. For routine internal reviews, a structured self-assessment using published frameworks is usually sufficient. Reserve external consultants for high-stakes evaluations where impartiality and advanced methods are essential.

## Related Clinical & Scientific Guides

* [Network Analysis for Infectious Disease Spread in Animal Populations](/knowledge/veterinary-medicine/veterinary-epidemiology/network-analysis-infectious-disease-spread-animal-populations)
* [Randomized Controlled Trials in Veterinary Field Settings](/knowledge/veterinary-medicine/veterinary-epidemiology/randomized-controlled-trials-veterinary-field-settings)
* [Regression Analysis in Veterinary Epidemiology: Logistic and Poisson Models](/knowledge/veterinary-medicine/veterinary-epidemiology/regression-analysis-veterinary-epidemiology-logistic-poisson-models)


## References and Further Reading

- [Evaluation of animal and public health surveillance systems: a systematic review.](https://pubmed.ncbi.nlm.nih.gov/22074638/). 2012.
- [Surveillance systems evaluation: a systematic review of the existing approaches.](https://pubmed.ncbi.nlm.nih.gov/25928645/). 2015.
- [Neurobehavioral function and low-level exposure to brominated flame retardants in adolescents: a cross-sectional study.](https://pubmed.ncbi.nlm.nih.gov/23151181/). 2012.
- [Overview of current toxicological knowledge of engineered nanoparticles.](https://pubmed.ncbi.nlm.nih.gov/21606847/). 2011.
- [Improving influenza vaccine virus selection: report of a WHO informal consultation held at WHO headquarters, Geneva, Switzerland, 14-16 June 2010.](https://pubmed.ncbi.nlm.nih.gov/21819547/). 2012.
- [Evaluation and optimization of surveillance systems for rare and emerging infectious diseases.](https://pubmed.ncbi.nlm.nih.gov/18651991/). 2008.
- [WOAH Animal Health Surveillance Standards](https://www.woah.org/en/what-we-do/animal-health-and-welfare/disease-data-collection/). WOAH.
- [CDC Principles of Epidemiology in Public Health Practice](https://www.cdc.gov/csels/dsepd/ss1978/index.html). CDC.
- [MSD Veterinary Manual, Professional Edition](https://www.msdvetmanual.com/). MSD Veterinary Manual.

## Related Articles

- [Bayesian Methods for Diagnostic Test Evaluation in Animals](/knowledge/veterinary-medicine/veterinary-epidemiology/bayesian-methods-diagnostic-test-evaluation-animals)
- [Evaluating Diagnostic Tests in the Absence of a Gold Standard](/knowledge/veterinary-medicine/veterinary-epidemiology/evaluating-diagnostic-tests-absence-gold-standard)
- [Syndromic Surveillance in Veterinary Practice](/knowledge/veterinary-medicine/veterinary-epidemiology/syndromic-surveillance-veterinary-practice)
- [Risk-Based Surveillance in Animal Health](/knowledge/veterinary-medicine/veterinary-epidemiology/risk-based-surveillance-animal-health)
- [Time Series Analysis for Veterinary Disease Surveillance](/knowledge/veterinary-medicine/veterinary-epidemiology/time-series-analysis-veterinary-disease-surveillance)

> 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.