 Synthetic population generation with public health characteristics for spatial agent-based modelsEmma Von Hoene, Amira Roess, Hamdi Kavak and Taylor Anderson PLOS Computational Biology, 2025, vol. 21, issue 3, 1-22 Abstract: Agent-based models (ABMs) simulate the behaviors, interactions, and disease transmission between individual “agents” within their environment, enabling the investigation of the underlying processes driving disease dynamics and how these processes may be influenced by policy interventions. Despite the critical role that characteristics such as health attitudes and vaccination status play in disease outcomes, the initialization of agent populations with these variables is often oversimplified, overlooking statistical relationships between attitudes and other characteristics or lacking spatial heterogeneity. Leveraging population synthesis methods to create populations with realistic health attitudes and protective behaviors for spatial ABMs has yet to be fully explored. Therefore, this study introduces a novel application for generating synthetic populations with protective behaviors and associated attitudes using public health surveys instead of traditional individual-level survey datasets from the census. We test our approach using two different public health surveys to create two synthetic populations representing individuals aged 18 and over in Virginia, U.S., and their COVID-19 vaccine attitudes and uptake as of December 2021. Results show that integrating public health surveys into synthetic population generation processes preserves the statistical relationships between vaccine uptake and attitudes in different demographic groups while capturing spatial heterogeneity at fine scales. This approach can support disease simulations that aim to explore how real populations might respond to interventions and how these responses may lead to demographic or geographic health disparities. Our study also demonstrates the potential for initializing agents with variables relevant to public health domains that extend beyond infectious diseases, ultimately advancing data-driven ABMs for geographically targeted decision-making.Author summary: In this study, we introduce a new method for generating synthetic populations of individuals or “agents” with characteristics that include health protective behaviors and attitudes, which are crucial for modeling disease spread. Traditional methods for parameterizing agents often overlook the complex relationships between demographic factors and health behaviors like vaccination. Additionally, detailed spatial data capturing these behaviors are limited, meaning agent behaviors are represented more uniformly across geographic space. By fitting public health surveys with spatially aggregated census data, we created agent populations that reasonably reflect real-world populations for disease spread simulations. We focused on Virginia, U.S. and generated a population with COVID-19 vaccine uptake and attitudes as of December 2021. Our results show that this approach captures the statistical relationships between demographic variables and vaccine uptake, along with the spatial variation in these behaviors. The approach is flexible so that it can be applied to various public health studies beyond just infectious diseases. Our work highlights the potential of public health surveys for enhancing synthetic population generation, offering a valuable approach for initializing models with more realistic populations to explore public health challenges. Date: 2025 Downloads: (external link) Related works: Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text Persistent link: https://EconPapers.repec.org/RePEc:plo:pcbi00:1012439 DOI: 10.1371/journal.pcbi.1012439 Access Statistics for this article More articles in PLOS Computational Biology from Public Library of Science |
Is your work missing from RePEc? Here is how to contribute.
Questions or problems? Check the EconPapers FAQ or send mail to .
EconPapers is hosted by the
School of Business at Örebro University.