From Evidence to Action: Artificial Intelligence in Governmental Public Health Decision-Making

Presentation Type: Podium Abstract
Presentation Time: 02:00 PM - 02:12 PM

Abstract Keywords: Artificial Intelligence, Governance, Public Health, Informatics Implementation, Policy
Programmatic Theme: Public Health Informatics

AI is increasingly promoted to support governmental public health decision-making, yet its real-world evidence base remains unclear. We conducted a scoping review of AI applications used by public health professionals and elected officials. Only seven studies met inclusion criteria, all focused on COVID-19. Findings reveal limited documentation, minimal evaluation, and substantial gaps between AI innovation and accountable governmental use. We propose a conceptual framework on responsible AI development and implementation in public health emergencies.

Speaker(s):
Elizabeth Campbell, MS, MSPH, PhD
Johns Hopkins Bloomberg School of Public Health
Author(s):
Hannah Goodtree, MPH - Johns Hopkins Center for Outbreak Response Innovation; Sarah Gillani, MPH - Center for Outbreak Response Innovation at Johns Hopkins Center for Health Security; Oluremilekun Oyefolu, MD - Johns Hopkins Center for Outbreak Response Innovation; Alison Kelly, MPIA - Johns Hopkins Center for Outbreak Response Innovation; Caitlin Rivers, PhD - Center for Outbreak Response Innovation; Crystal Watson, DrPH - Johns Hopkins Center for Outbreak Response Innovation;

---

Toward Precision Population Health for Infectious Disease Control and Prevention in a Learning Health System: Lessons from the ARMOR Household Study

Presentation Type: Podium Abstract
Presentation Time: 02:12 PM - 02:24 PM

Abstract Keywords: Infectious Diseases and Epidemiology, Population Health, Data Mining, Public Health
Programmatic Theme: Public Health Informatics

Utilizing electronic health records and biobanking data from a large quarternary health system and data and specimens collected during the Antibody Response Monitoring for Occupational Resilience Household Transmission of COVID-19 (ARMOR Household) study, a prospective cohort study with daily household self-reporting and sample collection, we designed a scalable, modular infectious disease surveillance pipeline comprising: (1) data collection and automated query, (2) data harmonization and validation, (3) epidemiologic modeling, and (4) multilevel reporting.

Speaker(s):
Harry Reyes Nieva, PhD, MAS
Columbia University
Author(s):
Medini Annavajhala, PhD - Children's Hospital of Philadelphia; Simian Huang, MS - Columbia University; Jason Zucker, MD - Columbia University; Jennifer Chang, MPH - University of Michigan; Craig Heck, PhD - Columbia University; Magdalena Sobieszczyk, MD, MPH - Columbia University; Anne-Catrin Uhlemann, MD, MPH - Columbia University; Delivette Castor, PhD - Columbia University;

---

Estimating the Prevalence of Antimicrobial-Resistant Gonorrhea in United States Counties Using Machine Learning Models

Presentation Type: Podium Abstract

Click to View Presentation

Presentation Time: 02:24 PM - 02:36 PM

Abstract Keywords: Infectious Diseases and Epidemiology, Machine Learning, Public Health
Programmatic Theme: Public Health Informatics

We developed and evaluated machine learning models to estimate county-level prevalence of antimicrobial-resistant gonorrhea in U.S. counties without surveillance data. Using GISP data and publicly available demographic and epidemiologic indicators, models achieved moderate geographic generalizability, with best performance for ciprofloxacin resistance. These findings demonstrate the potential and limitations of leveraging population-level data to inform local AMR surveillance and targeted public health interventions.

Speaker(s):
Ze Wang, BS, BA
University of California, San Francisco (UCSF)
Author(s):
Wanjun Gu, BS, BA - University of California, San Francisco (UCSF); Jingwen Li, MDS, MPH - Centrymed Pharmaceutical Co., Ltd; Reza Yaesoubi, PhD - UCSF;

---

A Right to Be Counted: Improving Surveillance and Trust Through Policy

Presentation Type: Podium Abstract
Presentation Time: 02:36 PM - 02:48 PM

Abstract Keywords: Population Health, Governance, Privacy and Security, Policy, Data Sharing
Programmatic Theme: Public Health Informatics

Public health surveillance policies codify trade-offs between competing values, like individual privacy and community benefit. We present policy and ethical insights from mixed-methods, community-based participatory research (survey, focus groups) involving two vulnerable Hispanic communities. We found strong support for public health surveillance and data sharing (i.e., a 'moral obligation,') that challenges existing restrictive policies and may underestimate the 'right to be counted.' However, support depends on trust, and guardrails restricting non-health uses are needed.

Speaker(s):
Cason Schmit, JD.
Texas A&M University
Author(s):
Hye-Chung Kum, PhD, FAMIA - Population Informatics Lab, SPH, Texas A&M University; William Sage, MD, JD - Institute for Healthcare Access;

---

Geospatial Data Quality Infrastructure for AI-Enabled Learning Health Systems

Presentation Type: Paper - Regular

Click to View Presentation


Click to View Handout

Presentation Time: 02:48 PM - 03:00 PM

Abstract Keywords: Delivering Health Information and Knowledge to the Public, Healthcare Quality, Mobile Health, Geospatial (GIS) Data/Analysis, Global Health, Data Standards, Healthcare Quality, Informatics Implementation
Programmatic Theme: Public Health Informatics

Geospatial data are foundational to digital health navigation systems, yet the quality of facility registries supporting these systems is often assumed rather than systematically evaluated. In many low- and middle-income country (LMIC) health environments, facility registries may be incomplete, outdated, or inconsistent, potentially degrading the performance of AI-enabled navigation tools that rely on them. This study examines how geospatial registry quality affects the performance of an AI health navigation system within a Learning Health System (LHS) framework. Using a controlled simulation with synthetic facility data representing realistic registry conditions, we evaluate system performance under high-quality and degraded registry scenarios.

Results show that reduced registry quality leads to substantial declines in routing accuracy and population coverage. We further demonstrate how user feedback signals can be aggregated to identify registry errors and trigger corrective updates, producing measurable recovery in system performance over successive cycles. These findings suggest that geospatial registry quality should be treated as a measurable infrastructure component within LHS governance rather than as static background data.

Speaker(s):
Andy Skumanich, PhD
Innov8ai Inc.
Author(s):
Madani Gadjigo, MS - Innov8ai Inc.; Andy Skumanich, PhD - Innov8ai Inc.; Han Kyul Kim, MS - University of Southern California;

---

Constructing Prenatal Environmental Exposure Windows from EHR Address Data: A Geospatial Framework for a Neonatal Cohort Study

Presentation Type: Podium Abstract

Click to View Presentation

Presentation Time: 03:00 PM - 03:12 PM

Abstract Keywords: Data Transformation/ETL, Real-World Evidence Generation, Information Visualization, Data Standards, Data Mining, Interoperability and Health Information Exchange
Programmatic Theme: Clinical Research Informatics

Environmental exposures during pregnancy influence neonatal outcomes, yet linking electronic health record (EHR) data to geographic exposures presents substantial methodological challenges. We describe a systematic framework for constructing prenatal air-quality exposure windows using EHR address data for 4,602 neonates in a Level IV NICU. We detail challenges of multi-granularity address records, multi-ZIP linkage, and Area Deprivation Index assignment, with interactive geospatial visualizations confirming cohort coverage.

Speaker(s):
Janet Northcote, PhD Candidate
University of Florida
Author(s):
Janet Northcote, PhD Candidate - University of Florida; Lisiane Pruinelli, PhD, MS, RN, FAMIA, FAAN - University of Florida;

---