Data Governance · 3 min read
Belgium's COVID-19 modelling held up, but its data plumbing did not
A 2020–2024 scoping review in BMC Infectious Diseases credits Belgium with strong genomic surveillance and forecasting, but flags weak dataset linkage, delayed variant estimates and untested contact tracing.
Key takeaways
- Scoping review covers Belgian COVID-19 modelling and analytical studies from 2020 to 2024, published in BMC Infectious Diseases (DOI: 10.1186/s12879-026-14748-8).
- Authors identified genomic surveillance, social contact analysis and epidemic forecasting as the main analytical strengths.
- Documented gaps: limited sub-national analyses, weak individual-level dataset integration, delayed variant-specific transmission parameter estimates.
- Superspreading was insufficiently characterised and digital contact tracing lacked formal evaluation.
- Authors call for improved data governance, faster access to linked datasets and integrated epidemic intelligence systems.

A scoping review of Belgian COVID-19 quantitative research from 2020 to 2024 has delivered a split verdict: strong analytical methods, weak data foundations. The study, led by Dupas and colleagues at the University of Hasselt and KU Leuven, appeared in BMC Infectious Diseases (DOI: 10.1186/s12879-026-14748-8) and was reported by Scienmag on November 20, 2026.
The finding matters beyond epidemiology. Every organization that ran analytics under crisis pressure should recognize the pattern described here. Good models emerged anyway. The gaps sat underneath them, in data governance and access.
What worked?
The review identified three domains where Belgian quantitative research performed well:
- Genomic surveillance — sequencing-based tracking of viral variants, which anchored transmission monitoring across the population.
- Social contact analysis — survey-based measurement of how people mix, the core input for transmission models.
- Epidemic forecasting — projection of epidemic trends to support public health decisions.
These three, per the authors, provided the essential scientific frameworks for tracking transmission patterns and projecting trends during the prolonged crisis. That is a real capability. Belgium produced decision-grade evidence under time pressure.
Where were the gaps?
The review documented five documented weaknesses. The pattern is familiar to anyone who has audited a rushed analytics operation.
- Limited sub-national analyses — national aggregates masked regional variation in a country with distinct regional health authorities.
- Weak integration of individual-level datasets — the data existed but rarely linked.
- Delayed estimation of variant-specific transmission parameters — the answers arrived after the decisions.
- Insufficient characterisation of superspreading — the disproportionate role of clustered transmission went under-modelled.
- Limited formal evaluation of digital contact tracing — a tool was deployed without rigorous performance assessment.
Note what is not said. The review does not quantify any of these gaps. No figures on how many studies were reviewed, no delay measured in weeks, no estimate of superspreading's share of transmission. A scoping review maps a field; it does not measure it. Treat the conclusions as directional.
What should data teams take from this?
The authors' recommendations read like a post-incident review for any data organization. Advancing evidence-based decision-making during crises requires, in their words, improved data governance, faster access to linked datasets, and the development of integrated epidemic intelligence systems capable of delivering timely analytical insights.
Translated to practice, that means three commitments made before the next emergency, not during it:
- Pre-negotiated data-sharing agreements, so linked datasets do not wait on legal review mid-crisis.
- Sub-national analytical capacity, because aggregates hide the heterogeneity that drives decisions.
- Evaluation built into every deployed tool from day one, not retrofitted after the budget is spent.
The skeptical read
A 2026-dated review covering 2020–2024 arrives after the policy window has closed. The Belgian epidemic is over; the recommendations target future emergencies. That is the correct framing, but it depends on institutional memory surviving until the next event. The review's authors — Dupas, Bagaforo, Hubin, Gilbert, Dellicour, and Hens — have documented both the capability and the shortfall. Whether anyone funds the fix is a separate question the paper cannot answer.
The evidence base was strong where the data was strong. Where the data was fragmented, the analysis was late or absent. That is not an epidemiology lesson. That is an infrastructure lesson.
doi.org (Original)
More from Tom Whitfield
Show full bio
Staff writer covering regulation, policy and their impact on Analytics Digest is a professional daily at Analytics Digest.
17 articles