References & Citations

Evidence-based public health intelligence requires source discipline. Public facts on this site should trace to peer-reviewed research, official guidance, or clearly labeled APHI analysis.

Categories

AI in Disease Surveillance & Forecasting

1
Peer-Reviewed
Using artificial intelligence to improve public health: a narrative review
PMC10637620. National Center for Biotechnology Information, U.S. National Library of Medicine.
Key Finding: AI applications in public health include spatial modeling, risk prediction, misinformation control, syndromic surveillance, disease forecasting, and pandemic modeling.
View Publication →
2
Government Report
Artificial Intelligence and Machine Learning | Technologies | CDC
Centers for Disease Control and Prevention, Data Modernization Initiative.
Key Finding: Machine learning can help identify patterns in health data that may indicate public health threats, supporting situational awareness when paired with human review.
View Report →
3
Research Article
Artificial intelligence in early warning systems for infectious disease surveillance
Frontiers in Public Health, 2025.
Key Finding: AI enables rapid and scalable support for case-based and event-based surveillance, with approximately 8,000 articles processed daily.
View Article →
4
Journal Article
Harnessing the power of artificial intelligence for disease-surveillance purposes
BMC Proceedings, 2025.
Key Finding: Deep learning and natural language processing integrate diverse data sources (epidemiological, web, climate, wastewater) for enhanced surveillance capabilities.
View Article →

CDC AI/ML Implementation

5
Government Report
CDC's Vision for AI in Public Health
Centers for Disease Control and Prevention, 2026.
Key Finding: CDC frames AI as a tool to accelerate detection and response, reduce administrative burden, support operational excellence, and strengthen partnerships across public health agencies, academia, industry, and government.
View Report →
6
Implementation Case
MedCoder: AI-Powered Death Certificate Coding
CDC National Vital Statistics System.
Key Finding: MedCoder integrates natural language processing, machine learning, rules-based programming, and regression testing to code causes of death for vital statistics.
View Implementation →
7
Implementation Case
TowerScout: Automated Cooling Tower Detection
CDC collaboration with UC Berkeley.
Key Finding: Web application automatically detects cooling towers from satellite imagery, accelerating CDC's ability to respond to Legionnaires' disease outbreaks.
View Case Study →
8
Program Report
FluSight: AI/ML for Influenza Forecasting
CDC Influenza Division.
Key Finding: CDC describes FluSight as a public health forecasting effort that supports influenza situational awareness and planning.
View Program →

WHO AI Governance & Ethics

9
International Standard
Ethics & Governance of Artificial Intelligence for Health
World Health Organization, 2021. ISBN: 978-92-4-002920-0.
Key Finding: First harmonized set of international ethics guidance on using health AI in accordance with ethical norms and human rights standards.
View Publication →
10
Initiative Report
Global Initiative on AI for Health (GI-AI4H): Strategic Priorities
NPJ Digital Medicine, 2025.
Key Finding: WHO initiative harmonizes governance standards for AI and advances ethical, regulatory, implementation, and operational dimensions in low- and middle-income countries.
View Article →

Health Equity & Algorithmic Fairness

11
CDC Guidance
Health Equity and Ethical Considerations in Using Artificial Intelligence in Public Health and Medicine
CDC, Preventing Chronic Disease, 2024.
Key Finding: AI poses risks that could exacerbate existing disparities. Misuse can increase disparities for socially and economically disadvantaged populations, including Indigenous Peoples, racialized communities, and those in rural/remote regions.
View Guidance →
12
Research Review
Integrating health equity in artificial intelligence for public health in Canada
Frontiers in Public Health, 2025.
Key Finding: Key equity challenges include gaps in AI epistemology, biases across the AI lifecycle, digital divide, unrepresentative training datasets, and privacy concerns disproportionately impacting priority populations.
View Review →
13
Policy Analysis
Artificial intelligence in public health: promises, challenges, and an agenda for policy makers
The Lancet Public Health, 2025.
Key Finding: Core challenges include equity, accountability, data privacy, robust digital infrastructures, and workforce skills. Issues include dual valence and automation bias.
View Article →

Privacy-Preserving Methods

14
Research Article
Federated Learning for Health Data Collaboration
Privacy-preserving machine learning literature.
Key Finding: Federated learning can support model development across institutions without centralizing sensitive data, but governance, data quality, and validation remain central constraints.
View Study →
15
Privacy Framework
Differential Privacy in Public Health Data Releases
Applied privacy techniques in population health analytics.
Key Finding: Differential privacy provides mathematical guarantees that individual-level data cannot be reverse-engineered from aggregate statistics.

Model Performance & Validation

16
Validation Study
Wastewater Surveillance and Forecasting
Public health surveillance literature.
Key Finding: Wastewater data can provide population-level infectious disease signals. Machine learning may help interpret trends, but local sampling, laboratory methods, and clinical context affect usefulness.
17
Validation Standard
Outbreak Signal Review Evaluation
APHI evaluation requirement.
Key Finding: Any future APHI detection claim must be based on a documented comparison against an existing surveillance workflow, including data sources, denominator, alert burden, missed signals, false positives, and review date.

Policy & Governance Frameworks

21
Federal Framework
NIST AI Risk Management Framework
National Institute of Standards and Technology, 2023.
Key Finding: Comprehensive framework for managing AI risks with core functions (Govern, Map, Measure, Manage) emphasizing trustworthiness, safety, security, and resilience throughout the AI lifecycle.
View Framework →
22
Federal Guidance
OMB Memoranda on AI Governance
Office of Management and Budget, 2025.
Key Finding: M-25-21 establishes standards for accelerating federal AI use through innovation, governance, and public trust. M-25-22 provides guidance on efficient AI acquisition in government, balancing innovation with responsible deployment.
View OMB Guidance →
23
Legal Standard
Official AI and Machine Learning Definitions
15 U.S.C. 9401(3), United States Code.
Key Finding: Federal statutory definitions of Artificial Intelligence ("a machine-based system that can... make predictions, recommendations, or decisions influencing real or virtual environments") and Machine Learning ("characterized by providing systems the ability to automatically learn and improve on the basis of data or experience").

Implementation & Adoption

24
Implementation Standard
Public Health AI Adoption Requirements
APHI implementation analysis.
Key Finding: Adoption depends on data governance, workflow fit, workforce readiness, procurement constraints, auditability, and trust. APHI will not publish adoption percentages without a public source.
25
Use Case Boundary
Overdose Prevention Prioritization
APHI concept-stage use case.
Key Finding: Population-level prioritization may support prevention planning, but individual diagnosis, enforcement, or denial of services are excluded uses.
26
Use Case Boundary
Vaccination Outreach Planning
APHI concept-stage use case.
Key Finding: Outreach planning tools must account for access, trust, local context, and equity. Performance claims require public evidence or partner-approved evaluation.

About These References

This page documents the scientific and institutional evidence base supporting claims made throughout the APHI website. References are categorized for easy navigation and include:

  • Peer-Reviewed Publications: Articles from scientific journals (e.g., Lancet, NEJM, Nature, Frontiers)
  • Government Reports: Official documentation from CDC, WHO, and other health agencies
  • Implementation Cases: Real-world deployments and program evaluations
  • Validation Studies: Model performance assessments and impact analyses

Numeric claims are used only when tied to a public source or a documented partner evaluation. Composite or internal metrics are not presented as APHI performance claims.

Last Updated: May 2026 | Review Cadence: Quarterly