Comparing AI Enabled and Business as Usual Evidence Surveillance for Public Health Intelligence
Abstract
Background: Timely, high‑quality evidence surveillance is central to public health protection, preparedness and decision‑making. Business‑as‑usual (BAU) surveillance is well established but resource‑intensive and increasingly challenged by the volume and fragmentation of public health information. AI‑enabled surveillance may support digital public health and horizon scanning, but evidence of its routine operational value is limited.
Methods: Within a Research and Guideline Development Unit, AI‑enabled weekly evidence surveillance operated alongside BAU surveillance between 26/12/2025 and 3/04/2026. AI surveillance used automated horizon scanning of international, European and national institutional sources, with aggregation and deduplication. BAU surveillance relied on expert‑led manual searching and synthesis. Outputs were compared by volume, thematic coverage, geographic scope and relevance to health protection.
Results: BAU surveillance identified over 400 items, compared with approximately 250 AI‑generated outputs. Both approaches consistently identified core surveillance reports and guidance from major institutions. BAU outputs offered greater granularity, including detailed guidance updates and implementation materials. AI outputs were more consolidated, prioritising distinct surveillance products, substantive updates and cross‑cutting intelligence to support situational awareness.
Conclusions: AI‑enabled surveillance supported consolidation and horizon scanning but was not mature enough to replace routine health protection surveillance. Expert‑led BAU processes remain essential for contextual interpretation and operational relevance. Further evaluation is required to support safe and effective integration of AI into public health surveillance systems.
Key messages: • Public Health Evidence Surveillance directly reflects the core activity being evaluated and anchors the work in health protection practice.
• Artificial Intelligence in Public Health captures the novel methodological component and aligns with digital public health literature.
Related
- Building smarter surveillance: a framework for artificial intelligence-enabled monitoring of overuse injuries in youth sports.
- Artificial intelligence–enabled nursing surveillance for pediatric cancer immunotherapy: integrating biomarkers, patient-reported outcomes, and long-term toxicity management
- Modernising public health surveillance in an EHDS-enabled Europe: building a coordinated European public health intelligence ecosystem
- Artificial intelligence-enabled healthcare access and user experience: comparative institutional analysis and survey evidence from China
- Family and informal caregivers’ acceptance, concerns and expectations regarding artificial intelligence-enabled dementia care: a protocol for qualitative evidence synthesis
BibTeX
@article{quintyne2026comparing,
title = {{Comparing AI Enabled and Business as Usual Evidence Surveillance for Public Health Intelligence}},
author = {KI Quintyne and C Gilbourne and S Kelly and M Williams and R Parlour},
journal = {The European Journal of Public Health},
year = {2026},
month = oct,
doi = {10.1093/eurpub/ckag169.1197},
url = {https://doi.org/10.1093/eurpub/ckag169.1197},
note = {CC BY-NC 4.0},
}
RIS
TY - JOUR
TI - Comparing AI Enabled and Business as Usual Evidence Surveillance for Public Health Intelligence
AU - KI Quintyne
AU - C Gilbourne
AU - S Kelly
AU - M Williams
AU - R Parlour
PY - 2026
DA - 2026/10/01
JO - The European Journal of Public Health
DO - 10.1093/eurpub/ckag169.1197
UR - https://doi.org/10.1093/eurpub/ckag169.1197
AB - Background: Timely, high‑quality evidence surveillance is central to public health protection, preparedness and decision‑making. Business‑as‑usual (BAU) surveillance is well established but resource‑intensive and increasingly challenged by the volume and fragmentation of public health information. AI‑enabled surveillance may support digital public health and horizon scanning, but evidence of its routine operational value is limited. Methods: Within a Research and Guideline Development Unit, AI‑enabled weekly evidence surveillance operated alongside BAU surveillance between 26/12/2025 and 3/04/2026. AI surveillance used automated horizon scanning of international, European and national institutional sources, with aggregation and deduplication. BAU surveillance relied on expert‑led manual searching and synthesis. Outputs were compared by volume, thematic coverage, geographic scope and relevance to health protection. Results: BAU surveillance identified over 400 items, compared with approximately 250 AI‑generated outputs. Both approaches consistently identified core surveillance reports and guidance from major institutions. BAU outputs offered greater granularity, including detailed guidance updates and implementation materials. AI outputs were more consolidated, prioritising distinct surveillance products, substantive updates and cross‑cutting intelligence to support situational awareness. Conclusions: AI‑enabled surveillance supported consolidation and horizon scanning but was not mature enough to replace routine health protection surveillance. Expert‑led BAU processes remain essential for contextual interpretation and operational relevance. Further evaluation is required to support safe and effective integration of AI into public health surveillance systems. Key messages: • Public Health Evidence Surveillance directly reflects the core activity being evaluated and anchors the work in health protection practice. • Artificial Intelligence in Public Health captures the novel methodological component and aligns with digital public health literature.
SN - 1101-1262
SN - 1464-360X
ER -