GIS, Public Health, and Agent-Based Modeling in the Era of COVID-19

Directions Magazine is pleased to welcome Eva Reid as a regular columnist. In her debut contribution, she examines how health data modeling—particularly through GIS and agent-based systems—will influence disease tracking, vaccine deployment, and geospatial decision-making in the months ahead.

Geospatial professionals have long emphasized that mapping is only the beginning. As Dr. Este Geraghty, Chief Medical Officer and Health Solutions Director at ESRI, recently noted in discussion about GIS and public health innovation, spatial technology is far more than visualization—it is an analytical engine for understanding complex systems. That perspective is especially relevant as new modeling tools such as Epistemix’s agent-based platform, FRED (Framework for Reconstructing Epidemiological Dynamics), gain attention.

The Longstanding Link Between Place and Health

The relationship between geography and health outcomes is not a recent discovery. For more than a century and a half—and arguably since classical antiquity—scholars have recognized that location, environment, and population characteristics influence disease patterns. This spatial-health connection has fueled major advancements in medical research and public health strategy. Geographic Information Systems provide a structured way to examine how environmental context, population density, infrastructure, and mobility intersect to shape health outcomes.

Over the past two decades, GIS has become a standard instrument in public health research. It supports mapping of healthcare resources and disease incidence while enabling monitoring, assessment, and forecasting. From outbreak surveillance to estimating infection risk and anticipating emergency response needs, spatial analytics enhances situational awareness. As technology evolves, so too does the precision of public health surveillance, including improved tracking of virus transmission pathways and more efficient allocation of protective equipment and pharmaceuticals.

GIS in Epidemiology and Disease Surveillance

Geospatial analysis has played a central role in understanding recent infectious disease outbreaks such as H1N1 and Ebola, as well as ongoing monitoring of Zika, West Nile virus, and measles. Epidemiologists rely on spatial data to address pressing questions: how rapidly is a virus spreading, where is it moving, how strained are hospital systems, and where are medical resources located?

Mathematical simulations and statistical forecasting remain fundamental tools in epidemiology. These approaches estimate outbreak trajectories and inform intervention strategies. When combined with GIS, modeling gains geographic depth. Spatially enabled “what-if” scenarios can illustrate how disease may propagate across neighborhoods, regions, or states and help prioritize vaccine distribution strategies based on place-specific risk.

Yet integrating GIS modeling into pandemic response presents institutional challenges. According to Dr. Geraghty, many epidemiologists do not work directly with GIS tools on a daily basis. Within local health departments, GIS capacity is often described as moderate to limited—not due to disinterest, but because geospatial expertise is frequently placed within IT divisions. GIS, however, is not merely information technology; it represents informatics and analytical science embedded in organizational workflows.

Data Constraints in Health Modeling

Another obstacle lies in the availability and structure of population data. Health models frequently depend on datasets that may contain gaps, quality limitations, or aggregation requirements designed to protect individual privacy. Census information, for example, originates at the individual level but is released in aggregated geographic units. These data inform infrastructure investment decisions, social program targeting, and identification of vulnerable populations. While indispensable, aggregated datasets restrict visibility into individual behavior and can obscure variability within communities.

Additional limitations arise when variables are unavailable at consistent geographic scales or when datasets become outdated between enumeration cycles. Environmental risk information can also be difficult to obtain or standardize. Moreover, demographic and socioeconomic indicators often fail to capture behavioral dimensions, such as whether individuals will seek medical services or comply with public health guidance.

Agent-Based Modeling and the Emergence of FRED

To address these complexities, agent-based modeling (ABM) has emerged as a powerful complementary approach. ABM simulates the behavior of individuals—referred to as “agents”—and their interactions within defined environments. By modeling both macro- and micro-level dynamics, this methodology captures how individual decisions influence broader system outcomes. Beyond public health, agent-based techniques have been applied in transportation planning and infrastructure investment analysis.

Epistemix, a company founded on research led by Dr. Donald Burke and John Grefenstette at the University of Pittsburgh, developed FRED as an agent-based epidemiological modeling system. Although the framework has existed for years, its collaboration with ESRI—initiated slightly over a year ago—combines behavioral modeling with geographic intelligence. The partnership seeks to integrate spatial context and human behavior into disease spread simulations.

Responding to COVID-19 Through Spatial and Behavioral Insight

While earlier applications of FRED examined diseases such as measles, the platform proved particularly relevant during the COVID-19 pandemic. Researchers recognized that the model could help examine herd immunity thresholds, vaccine allocation strategies, and the influence of human movement patterns on transmission dynamics. Identifying gaps in available data becomes part of the modeling process itself, guiding more informed public health decisions.

Dr. Geraghty has emphasized the importance of actionable insight during periods of uncertainty. Even imperfect models grounded in scientific methodology provide valuable guidance when policymakers must act quickly. In a rapidly evolving crisis, clarity—even if partial—supports responsible planning.

The Centers for Disease Control and Prevention (CDC) has estimated vaccine dose requirements for each state based on multiple factors. However, as vaccination campaigns progress, mapping clinic locations and population counts alone is insufficient. Decision-makers must consider environmental conditions, mobility networks, and social behavior. Facilities equipped with ultra-cold storage may exist near target populations, but accessibility and willingness to seek vaccination vary. Behavioral patterns, transportation barriers, and levels of trust influence uptake rates.

Public health planning must also confront vaccine hesitancy, particularly in communities historically underserved or disproportionately affected by systemic inequities. Incorporating these dynamics into modeling frameworks is complex. Agent-based systems like FRED offer potential pathways for representing behavioral variability and community interaction patterns within geographic space.

One significant contribution of this approach is its ability to conceptualize herd immunity geographically. Communities differ in contact patterns, density, and mobility. Consequently, vaccination thresholds necessary to interrupt transmission may vary spatially. If models reveal that certain areas can achieve protective effects at lower vaccination rates due to interaction patterns, resource allocation strategies could be refined. Such insights may ease logistical pressures and address concerns about distribution site capacity or cold-chain infrastructure.

Looking Ahead

As health departments navigate ongoing pandemic challenges and future disease threats, partnerships that merge GIS, epidemiology, and behavioral modeling may redefine public health practice. Integrating tools like FRED within geospatial platforms enhances the capacity to simulate complex realities rather than relying solely on aggregated statistics.

The coming months will test how effectively agencies translate spatial intelligence into policy. However, collaborations between organizations such as ESRI and Epistemix illustrate a broader transformation underway: public health is increasingly informed by geography, data science, and computational modeling working in concert. In confronting global health crises, this convergence of GIS and agent-based analytics may prove indispensable.