cs.MAOct 22, 2026
SaveToward an Agentic AI Approach to Public Health Data Collection
Organizations: Department of Systems and Information Engineering, University of Virginia, Charlottesville, USA · Engineering and Applied Sciences Department, RAND, Arlington, USA
Abstract
Disease surveillance increasingly integrates automated data collection, yet predefined sources and procedures require manual adaptation as reporting practices change. We explore agentic artificial intelligence (AI) as an approach to coordinating surveillance data collection across heterogeneous sources. Specifically, we built a multi-agent framework that integrates source discovery, source assessment, evidence structuring, event linking, and validation. Three retrospective applications illustrate its promise. The workflow recovered all 13 individual cases in a hantavirus outbreak, with links to 85 distinct publisher domains compared with 28 in the reference benchmark, reconstructed Canadian measles counts with a final total 0.33% below the national reference and higher cumulative counts on 79.8% of comparable days, and increased the share of days with recent Sierra Leone mpox reports from 28.2% to 65.6% by incorporating attributed media. These demonstrations also reveal limitations in record completeness and accuracy, while prospective gains in timeliness remain untested. We envision agentic AI as a complementary layer of public health data infrastructure, coordinating evidence-linked collection under surveillance goals and standards defined by public health professionals.
Figures & tables
Figure 1: Evolution of public health surveillance and the proposed agentic AI workflow. a , Conceptual evolution of public health data surveillance from manual, in which collections are performed by humans, to semi-automated systems built around fixed sources and rule-based processing, and toward agentic AI workflows combining end-to-end automation, adaptive execution, and expert oversight. b , Proposed agentic AI paradigm for public health data surveillance. A public health mission is translated into a shared task contract that coordinates specialized agents for source discovery, source assessment, evidence structuring, event linking, and validation. Gaps and conflicts redirect the workflow or trigger expert review, and a final quality gate releases structured outputs with traceable evidence.
Figure 2: Workflow-reconstructed results and reference comparisons across three case studies. a , Hantavirus cases (2026): workflow-reconstructed cumulative cases compared with World Health Organization (WHO) reference counts by release date. b , Distinct case-linked publisher domains per case for the workflow-reconstructed results in dark blue and the GitHub benchmark in light blue. c , Canadian measles (2025): cumulative reported cases by release date for the workflow-reconstructed results and the Public Health Agency of Canada (PHAC) reference. d , Number of days per month with workflow cumulative counts above PHAC counts, shown by light-green bars on the left axis, and their percentage among days with both series available, shown by the dark-green line on the right axis. NA indicates no comparable days. e , Sierra Leone mpox (2025): cumulative confirmed cases by release date for the workflow-reconstructed results and the WHO reference. f , Percentage of calendar days by time since the latest available report in 2025, without media in light red and with media in dark red.
Appendix figures & tables12 assets
Supplementary material from the paper’s appendix.
Appendix
| Stage | Key activities | Output |
|---|---|---|
| 1. Task specification | Define collection scope; assemble disease terminology; specify requested fields and evidence requirements. | Collection specification and record schema. |
| 2. Source discovery | Plan source categories and queries; search; refine queries around evidence gaps; consolidate repeated results. | Candidate sources with discovery information. |
| 3. Evidence acquisition | Assess relevance, publisher and permitted use; retrieve documents; retain extractable text and table context. | Assessed sources and traceable evidence fragments. |
| 4. Structured extraction | Extract explicitly reported observations; separate statistical meanings; validate and normalise records. | Structured observations with evidence and quality flags. |
| 5. Verification across sources | Link related reports; identify duplicates; compare aligned observations; retain support and disagreement. | Event groups, supporting evidence and unresolved discrepancies. |
| 6. Dataset assembly | Assess quality and evidence gaps; perform controlled recovery; retain any supplied review decisions and separate output categories. | Accepted dataset, pending issues and excluded records. |
Table S1: Major stages, key activities and outputs of the workflow.
Figure S1: The general structure of the data collection workflow. Arrows connect the six stages and show additional search and review paths. Dark blue denotes language model components, medium blue denotes rule based components and light blue denotes retrieval and processing tools. Human-review paths show an optional capability that was not used in the reported studies.
Figure S2: Langflow visualisation of the workflow. The canvas groups processing components into six stages. Each node represents a workflow component; its ports indicate the inputs and outputs passed between components. Solid connections show the specified processing sequence, while dashed connections indicate conditional choices. The connections also display the paths for additional evidence collection and optional human review. The human-review capability was not used in the reported studies.
| Case | Workflow all domains | Workflow direct evidence | Benchmark all domains |
|---|---|---|---|
| Case 1 | 54 | 19 | 3 |
| Case 2 | 41 | 14 | 4 |
| Case 3 | 35 | 12 | 4 |
| Case 4 | 32 | 14 | 2 |
| Case 5 | 12 | 3 | 4 |
| Case 6 | 11 | 2 | 5 |
Table S2: Source domains linked to each case in the hantavirus evaluation cohort.
| Measure | Workflow | GitHub benchmark |
|---|---|---|
| Correct recovery | 59/87 (67.82%) | 74/87 (85.06%) |
| Field completeness | 70/87 (80.46%) | 84/87 (96.55%) |
| Correctness among populated fields | 59/70 (84.29%) | 74/84 (88.10%) |
Table S3: Overall field quality for the hantavirus case study datasets.
| Month | Official | Secondary | Total |
|---|---|---|---|
| Jan | 0 | 0 | 0 |
| Feb | 6 | 2 | 8 |
| Mar | 9 | 4 | 13 |
| Apr | 11 | 3 | 14 |
| May | 20 | 5 | 25 |
| Jun | 15 | 17 | 32 |
Table S4: Canadian measles: monthly source observations.
| Period | Comparable / calendar days | Higher / equal / lower days | Higher (%) | Mean (cases) | Relative (%) |
|---|---|---|---|---|---|
| Jan | 0/31 | — | — | — | — |
| Feb | 25/28 | 13/0/12 | 52.0 | +8.80 | +20.16 |
| Mar | 31/31 | 23/0/8 | 74.2 | +70.65 | +27.68 |
| Apr | 30/30 | 30/0/0 | 100.0 | +125.90 | +15.98 |
| May | 31/31 | 29/0/2 | 93.5 | +326.32 | +18.94 |
| Jun | 30/30 | 29/0/1 | 96.7 | +139.57 | +4.55 |
Table S5: Canadian measles: daily reported-count advantage by calendar period.
| Period | Comparable / calendar days | Higher / equal / lower days | Higher (%) | Mean (cases) | Relative (%) |
|---|---|---|---|---|---|
| First month | 28/28 | 16/0/12 | 57.1 | +12.46 | +24.77 |
| First 3 months | 89/89 | 69/0/20 | 77.5 | +80.29 | +19.82 |
| First 6 months | 181/181 | 157/0/24 | 86.7 | +143.63 | +8.43 |
| First 9 months | 273/273 | 225/0/48 | 82.4 | +105.82 | +3.83 |
| Year | 331/365 | 264/0/67 | 79.8 | +88.71 | +2.77 |
Table S6: Canadian measles: daily reported-count advantage by tracking window.
| Month | National official | Media | Regional | Total |
|---|---|---|---|---|
| Jan | 5 | 1 | 0 | 6 |
| Feb | 0 | 1 | 0 | 1 |
| Mar | 0 | 2 | 0 | 2 |
| Apr | 0 | 4 | 0 | 4 |
| May | 0 | 9 | 0 | 9 |
| Jun | 0 | 2 | 0 | 2 |
Table S7: Sierra Leone mpox: monthly source observations.
| Period | Shared days | All sources days (%) | Without media days (%) | Gain (pp) |
|---|---|---|---|---|
| Jan | 21 | 21 (100.0) | 21 (100.0) | 0.0 |
| Feb | 28 | 10 (35.7) | 0 (0.0) | 35.7 |
| Mar | 31 | 16 (51.6) | 0 (0.0) | 51.6 |
| Apr | 30 | 14 (46.7) | 0 (0.0) | 46.7 |
| May | 31 | 31 (100.0) | 0 (0.0) | 100.0 |
| Jun | 30 | 18 (60.0) | 0 (0.0) | 60.0 |
Table S8: Mpox report freshness on the shared calendar, by month and half-year.
| Period | Comparable / calendar days | Higher / equal / lower days | Higher (%) | Mean (cases) | Relative (%) |
|---|---|---|---|---|---|
| Jan | 0/31 | — | — | — | — |
| Feb | 28/28 | 12/0/16 | 42.9 | -1.89 | -9.55 |
| Mar | 31/31 | 7/0/24 | 22.6 | -2.74 | -3.62 |
| Apr | 30/30 | 15/0/15 | 50.0 | +131.43 | +70.78 |
| May | 31/31 | 31/0/0 | 100.0 | +1073.90 | +94.26 |
| Jun | 30/30 | 25/0/5 | 83.3 | +262.67 | +7.41 |
Table S9: Sierra Leone mpox: daily reported-count advantage by calendar period.
| Period | Comparable / calendar days | Higher / equal / lower days | Higher (%) | Mean (cases) | Relative (%) |
|---|---|---|---|---|---|
| First month | 10/31 | 10/0/0 | 100.0 | +1.80 | +11.25 |
| First 3 months | 69/90 | 19/0/50 | 27.5 | -3.16 | -5.38 |
| First 6 months | 160/181 | 97/0/63 | 60.6 | +283.19 | +23.37 |
| First 9 months | 252/273 | 152/0/100 | 60.3 | +199.75 | +7.66 |
| Year | 334/365 | 178/56/100 | 53.3 | +160.47 | +4.87 |
Table S10: Sierra Leone mpox: daily reported-count advantage by tracking window.
More in this topic
Data Analysis Agents
Data analysis agents use language models and tools to explore datasets, run analyses, and produce answers or reports grounded in evidence.
See all 106 papers