Headwinds in Public Health: How AI can Help Drive Progress

Summary

The discussion focuses on how public health systems could detect and respond to disease outbreaks more quickly by improving data integration, using carefully governed AI, and coordinating human, animal, and community health services.

Main ideas

• Public Health Practice’s work: Dr. Scott McNabb describes his company as a public health consulting and education organization. Its projects include workforce training, online learning, and helping organizations develop policies and programs with measurable outcomes. He cites work in Africa and the Middle East, including support for Africa CDC workforce development, training Saudi public health professionals, and consulting work in Sudan and Armenia.
• A fragmented surveillance system: McNabb says U.S. public health data is divided among state systems and specialized programs. This can create silos, duplicate work, and leave relevant information disconnected—for example, a disease case recorded in one registry may not appear in another. Some illnesses may also go unreported when people treat symptoms themselves rather than visiting a doctor.
• Using AI to improve reporting: McNabb proposes using ethical generative AI alongside electronic medical record systems. In his model, clinicians would enter information once, as they do now. AI could then identify possible reportable conditions, protect patient privacy, and turn relevant information into usable public health outputs—such as case lists or epidemic curves—so officials can connect cases and act sooner. The aim is to support public health work without adding another reporting burden for clinicians.
• “One Health” coordination: The speakers discuss connecting human health services with veterinary medicine and other relevant sources of information. Veterinarians or veterinary laboratories may detect some diseases before human clinicians do, and diseases can also involve insects and ticks. A coordinated approach could help officials identify emerging threats and investigate links among cases.
• Organizational and political barriers: McNabb argues that separate budgets and chains of command can discourage agencies from working together. He gives an example from Armenia in which information about a vector-borne disease had to pass through senior officials before reaching local physicians. He says integration requires leadership support, shared plans, and agreement about how resources should be used.
• Local ownership in international public health: McNabb reflects on the limits of top-down international assistance, describing past practices that could be paternalistic or disconnected from local priorities. He sees greater ownership by African countries and institutions as a more sustainable direction. He also says global standards and cooperation remain important, including potential future work with the World Health Organization.
• Technology and connectivity: The speakers discuss satellite internet, including Starlink, as a possible way to improve connectivity where conventional infrastructure is limited. McNabb describes it as a potentially reliable option, including in areas affected by damaged infrastructure.
• Disruption as an opportunity: McNabb sees political and technological upheaval as a reason to rethink existing public health systems rather than abandon the field. He encourages students and practitioners to remain engaged and focus on practical improvements.

Proposed approach and research plan

McNabb describes a potential way to test a more integrated public health model:

1. Keep the clinician’s existing workflow: Clinicians enter information into the electronic medical record as usual, without needing to submit it a second time.
2. Review data in parallel: AI examines the records for possible reportable conditions while protecting personally identifying information.
3. Create public health outputs: Where appropriate, the system converts information into formats officials can use, such as case lists or epidemic curves.
4. Connect related cases: Officials use the combined information to identify possible links, investigate outbreaks, and take preventive action.
5. Include multiple health sectors: Bring together public health agencies, clinicians, veterinarians, veterinary labs, pharmacies, and other relevant sources of information.
6. Pilot the model locally: McNabb proposes an applied research project in two or three counties, built from the grassroots and involving local communities.
7. Measure whether it works: Establish a monitoring and evaluation plan with agreed indicators, then assess whether the approach improves public health outcomes rather than assuming it will.

Speakers and sources featured

Speakers

• Patrick Begley — Host of 10 Minute Major.
• Dr. Scott McNabb — Guest; founder of Public Health Practice and a former CDC employee.

Organizations and systems discussed

• Public Health Practice
• Centers for Disease Control and Prevention (CDC)
• Africa CDC and the African Union
• World Health Organization (WHO)
• Emory University
• Clemson University
• University of Georgia
• Saudi public health institutions
• Google, Microsoft, and Starlink
• State and local public health agencies, medical providers, and veterinary services