Unlocking the Value of Veterinary Epidemiological Data: A New Standard for Reuse and Innovation

In the era of One Health, data is our most valuable asset—but only if we can use it.

For decades, the field of veterinary epidemiology has generated vast amounts of data — from national surveillance programs and academic studies to industry production records. Yet, a significant portion of this data sits in silos, effectively “single-use” because it lacks the context required for others to interpret and trust it.

Our new set of community-driven guidelines, titled “Enhancing reusability of veterinary epidemiological data by creating contextual metadata”, proposes a practical solution. By shifting focus from simple archiving to the creation of rich metadata, these guidelines aim to transform disparate datasets into a reusable global asset.

While we are separately advancing work on comprehensive data governance systems to manage access and sovereignty, these new metadata standards represent the foundational technical step: ensuring that when data is shared, it is actually useful.

The Problem with “Generic” Metadata

Current data management practices often rely on generic standards (like Dublin Core or DataCite). While excellent for cataloging a digital object (providing a title, author, and date), these standards fail to capture the complexity of data used in veterinary epidemiology.

Veterinary data is multi-scale and messy. It spans genetics, population dynamics, economics, and biology. A dataset might describe “mortality,” but without knowing the data lineage — was it observed by a vet, reported by a farmer, or inferred from sensor data? — a regulator or researcher cannot confidently reuse it.

The new guidelines argue that rich metadata must go beyond basic description to include detailed information on data quality, lineage, and context.

The Solutions

A Domain-Specific Framework

Developed by a consortium of researchers and institutions, the guidelines propose a three-tiered metadata structure tailored specifically for animal health:

1. General Information: The basics (access rules, licenses, and persistent identifiers) to ensure data is findable.

2. High-Level Description: Contextual intelligence that allows a user to assess relevance instantly. This includes spatial and temporal resolution (e.g., “aggregated by region” vs. “GPS coordinates”) and data lineage (the journey of the data from collection to processing).

3. Specific Content Description: The deep technical details required for analysis, such as data dictionaries (defining exactly what a variable means) and, crucially, data quality assessments.

Human-first metadata

Broader data sharing efforts in other disciplines rightly promote the importance of machine-readable metadata. However, in veterinary epidemiology, lack of familiarity with, and access to, technical expertise in the field means that requring machine-readable metadata acts as a blocker to creating any metadata at all. EpiMundi is promoting the concept of human-first metadata, that can be easily written, read and understood by humans. This can then be developed into machine readable formats, but for us, the humans come first.

Why This Matters: Benefits for Key Stakeholders

1. For Regulators: Transparency and Trade

In an interconnected world, sanitary barriers and trade negotiations rely on trust in surveillance data. By adopting these standards, competent authorities can transparently document the quality and coverage of their national datasets.

• Enhanced ability to demonstrate freedom from disease or low-risk status to trading partners using rigorously documented evidence.
• Improved integration of fragmented national datasets (e.g., distinct silos for different species) into a coherent One Health surveillance picture.

2. For Academic Researchers: Reproducibility and Impact

The “FAIR” principles (Findable, Accessible, Interoperable, Reusable) are becoming a mandate for funding. These guidelines provide the practical templates to meet those requirements without reinventing the wheel.

• oving beyond the “methods” section of a PDF. By attaching rich metadata to datasets, researchers ensure their work can be cited and reused correctly in meta-analyses, increasing long-term impact.
• Reducing the “data waste” where valuable field data is lost once a project ends.

3. For Industry: Efficiency and Innovation

Private sector data (production records, sensor data, genomics) is often messy and heterogeneous. Adopting internal metadata standards allows companies to integrate historical data with new streams, fueling AI and machine learning initiatives.

• Internal efficiency. When data lineage is documented, teams stop wasting time deciphering what “Column X” meant in a spreadsheet from three years ago.
• Preparing for the future. As the paper notes, future iterations will support machine-readable formats (like JSON-LD) for automated dashboards and decision support tools.

A Focus on Data Quality

Uniquely, these guidelines tackle the “elephant in the room”: data quality. The authors note that the absence of guidance on reporting flaws often pushes data owners to hide them.

The framework encourages transparency by providing templates to report on:

• Completeness: Is the dataset a census or a sample? What is missing?
• Correctness: Are there known invalid values or data entry errors?
• Reliability: Are there inconsistencies between internal fields or external benchmarks?

Looking Ahead: Governance and Systems

This paper provides the technical instructions for describing data. However, describing data is only half the battle. EpiMundi has broad experience in the parallel challenge of data governance.

While metadata ensures data is understandable, governance systems ensure it is secure. We have developed governance frameworks that define ownership, access control, usage rights, decision-makiing and ethical usage — ensuring that when rich metadata makes data reusable, it is shared only according to the strict terms of the data owner.

Conclusion

Data reuse is not just an academic exercise; it is an efficiency imperative for the veterinary sector. By adopting these new community standards for metadata, we can transform isolated spreadsheets into a connected ecosystem of intelligence, driving faster responses to disease outbreaks and more efficient food production systems.