---
title: "LLMPopSim: Using LLM-based agentic simulation for local public health estimation"
id: "medrxiv-18-evaluating-agentic-simulation-for-local-public-health-estimation"
canonical_url: "https://medichelpline.com/clinical-feed/medrxiv-18-evaluating-agentic-simulation-for-local-public-health-estimation"
content_type: "clinical_feed_article"
specialty: "General"
source_name: "medRxiv (Clinical Preprints)"
source_url: "https://www.medrxiv.org/content/10.64898/2026.09.19.26363431v1?rss=1"
published_at: "2026-09-21T12:00:00.000Z"
evidence_level: "Verified Feed"
license: "CC-BY-NC-4.0 / Informational Use"
---
# LLMPopSim: Using LLM-based agentic simulation for local public health estimation
## Provenance & Clinical Metadata
- **Canonical URL:** https://medichelpline.com/clinical-feed/medrxiv-18-evaluating-agentic-simulation-for-local-public-health-estimation
- **Specialty:** [General](https://medichelpline.com/clinical-feed/general.md)
- **Primary Source:** medRxiv (Clinical Preprints)
- **Source URL:** [Original Journal Publication](https://www.medrxiv.org/content/10.64898/2026.09.19.26363431v1?rss=1)
- **Published At:** 2026-09-21T12:00:00.000Z
- **Evidence Rating:** Verified Feed
## Executive GIST (TL;DR)
- The authors introduce **LLMPopSim**, a generative population simulation framework that integrates U.S. Census and CDC PLACES data to construct synthetic individuals and simulate health behaviors with a large language model (**LLM**). - The framework combines demographic and socioeconomic inputs from the U.S. Census Bureau American Community Survey (ACS) and community-level health estimates from the CDC PLACES project to build geographically grounded synthetic populations. - The developers used historical data from Hawaiʻi to create the model and tested temporal and geographic transfer on held-out 2022 cohorts from Hawaiʻi and New York State. - Proof-of-concept behaviors simulated were **colorectal cancer screening** and **mammography** uptake. - Across four state–outcome evaluations, mean absolute error (MAE) ranged from **3.5 to 15.0 percentage points** and correlations between simulated and observed ZCTA-level prevalence ranged from **0.26 to 0.69**. - Performance varied by outcome and by measures of population fidelity: colorectal cancer screening preserved geographic ranking better but systematically overestimated prevalence and compressed geographic variation; mammography showed lower absolute error but weaker geographic correlation and inconsistent preservation of between-community variability. - Prediction error tended to be larger in communities with lower observed screening prevalence and varied across community characteristics without a consistent socioeconomic gradient. - The study identifies calibration, distributional fidelity, and subgroup performance as key challenges for generative population simulation and positions LLMPopSim as an empirical foundation for synthetic-population approaches to simulate heterogeneous public health responses. - All source data used were publicly available aggregate datasets (ACS 5-year estimates and CDC PLACES releases); no individual-level or restricted-access data were used. - The authors declare no competing interests and state appropriate ethical oversight; the work is a preprint and not peer reviewed, so it should not be used to guide clinical practice.
## Clinical Analysis & Structured Key Points
Evaluating agentic simulation for local public health estimation | medRxiv Skip to main content Evaluating agentic simulation for local public health estimation View ORCID Profile Michelle A Nakatsuka , Robert J Steele , Xu Han , View ORCID Profile Eric K Oermann doi: https://doi.org/10.64898/2026.09.19.26363431 Michelle A Nakatsuka 1 NYU Grossman School of Medicine; Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Michelle A Nakatsuka For correspondence: michelleamurata{at}gmail.com Robert J Steele 2 NYU Langone Health; Find this author on Google Scholar Find this author on PubMed Search for this author on this site Xu Han 2 NYU Langone Health; Find this author on Google Scholar Find this author on PubMed Search for this author on this site Eric K Oermann 3 NYU Langone Department of Neurosurgery Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Eric K Oermann Abstract Info/History Metrics Data/Code Preview PDF Abstract Large language model (LLM)-based generative agents can reproduce aspects of individual human behavior, but whether they can be scaled to geographically grounded populations that reproduce real-world health behaviors remains unclear. Here, we introduce LLMPopSim, a generative population simulation framework that integrates U.S. Census and Centers for Disease Control and Prevention data to construct synthetic individuals and uses an LLM to simulate individual health behaviors whose aggregate outcomes can be evaluated at the community level. We developed the framework using historical Hawaiʻi data and evaluated temporal and geographic generalizability using held-out 2022 cohorts from Hawaiʻi and New York State, with colorectal cancer screening and mammography as proof-of-concept behaviors. Across the four state-outcome evaluations, mean absolute error ranged from 3.5 to 15.0 percentage points and correlations between simulated and observed ZCTA-level prevalence ranged from 0.26 to 0.69. Performance differed across dimensions of population fidelity: colorectal cancer screening predictions preserved geographic ranking more strongly but systematically overestimated prevalence and compressed geographic variation, whereas mammography achieved lower absolute error but weaker geographic correlation and inconsistent preservation of between-community variability. Prediction error was greatest in communities with lower observed screening prevalence and varied across community characteristics without a uniform socioeconomic gradient. These findings demonstrate that individually represented LLM-based synthetic agents can aggregate into population-level patterns that retain measurable features of real-world health behavior across temporal and geographic transfer, while identifying calibration, distributional fidelity and subgroup performance as key challenges for generative population simulation. LLMPopSim provides an empirical foundation for developing synthetic populations that may ultimately enable simulation of heterogeneous population responses to public health interventions. Competing Interest Statement The authors have declared no competing interest. Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: The study used only openly available, aggregate human data that were publicly accessible before initiation of the study. Demographic and socioeconomic data were obtained from the U.S. Census Bureau American Community Survey (ACS) 5-year estimates (2018, 2020, and 2022), available through the U.S. Census Bureau: https://www.census.gov/programs-surveys/acs/data.html. Community-level health characteristics and preventive health behavior estimates were obtained from the Centers for Disease Control and Prevention (CDC) PLACES project (2020, 2022, and 2024 releases), available through the CDC PLACES Data Portal: https://www.cdc.gov/places/tools/data-portal.html. All source human data were publicly available at the aggregate geographic level. No restricted-access or individually identifiable human data were used. Individual-level populations analyzed in the simulation were synthetically generated from these aggregate data. I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes Data Availability All source data used in this study are publicly available. Demographic and socioeconomic data were obtained from the U.S. Census Bureau American Community Survey (ACS) 5-year estimates (2018, 2020, and 2022), available at https://www.census.gov/programs-surveys/acs/data.html. Community-level health characteristics and preventive health behavior estimates were obtained from the Centers for Disease Control and Prevention (CDC) PLACES project (2020, 2022, and 2024 releases), available at https://www.cdc.gov/places/tools/data-portal.html. Individual-level populations used for behavioral simulation were synthetically generated from these publicly available aggregate data as described in the manuscript. https://www.census.gov/programs-surveys/acs/data.html https://www.cdc.gov/places/tools/data-portal.html Copyright The copyright holder for this preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license . Back to top Previous Next Posted September 21, 2026. Download PDF Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Evaluating agentic simulation for local public health estimation Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Evaluating agentic simulation for local public health estimation Michelle A Nakatsuka , Robert J Steele , Xu Han , Eric K Oermann medRxiv 2026.09.19.26363431; doi: https://doi.org/10.64898/2026.09.19.26363431 Share This Article: Copy Citation Tools Evaluating agentic simulation for local public health estimation Michelle A Nakatsuka , Robert J Steele , Xu Han , Eric K Oermann medRxiv 2026.09.19.26363431; doi: https://doi.org/10.64898/2026.09.19.26363431 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Areas All Articles Addiction Medicine (619) Allergy and Immunology (903) Anesthesia (335) Cardiovascular Medicine (4834) Dentistry and Oral Medicine (479) Dermatology (419) Emergency Medicine (651) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1633) Epidemiology (15980) Forensic Medicine (33) Gastroenterology (1206) Genetic and Genomic Medicine (7067) Geriatric Medicine (734) Health Economics (1067) Health Informatics (5062) Health Policy (1438) Health Systems and Quality Improvement (1776) Hematology (587) HIV/AIDS (1347) Infectious Diseases (except HIV/AIDS) (16310) Intensive Care and Critical Care Medicine (1176) Medical Education (671) Medical Ethics (155) Nephrology (727) Neurology (7297) Nursing (369) Nutrition (1084) Obstetrics and Gynecology (1244) Occupational and Environmental Health (1006) Oncology (3618) Ophthalmology (1057) Orthopedics (398) Otolaryngology (458) Pain Medicine (474) Palliative Medicine (140) Pathology (709) Pediatrics (1805) Pharmacology and Therapeutics (739) Primary Care Research (769) Psychiatry and Clinical Psychology (5922) Public and Global Health (9772) Radiology and Imaging (2434) Rehabilitation Medicine and Physical Therapy (1456) Respiratory Medicine (1250) Rheumatology (643) Sexual and Reproductive Health (773) Sports Medicine (581) Surgery (783) Toxicology (107) Transplantation (305) Urology (293)
## Related Clinical Research

- [Neuro-Symbolic AI for Automated Pathology Quality Measurement: Performance and Methods](https://medichelpline.com/clinical-feed/medrxiv-9-neuro-symbolic-ai-for-automated-pathology-quality-measurement.md)
- [BMI and Recurrence-Free Survival in Early-Onset Colorectal Cancer: A Cohort Analysis](https://medichelpline.com/clinical-feed/medrxiv-3-weighing-the-odds-body-mass-index-and-recurrence-free-survival-in-early-onset.md)
- [Structured aerobic exercise linked to lower cancer recurrence and improved survival](https://medichelpline.com/clinical-feed/medical-news-today-0-aerobic-exercise-could-help-improve-outlook-after-cancer-treatment.md)
- [Obesity-associated inflammation and cancer: mechanisms linking adipose immune changes to tumor risk](https://medichelpline.com/clinical-feed/frontiers-in-immunology-6-obesity-associated-inflammation-and-cancer-molecular-mechanisms-evidence-gaps.md)
- [Healthy lifestyle and pan-cancer risk: UK Biobank prospective analysis](https://medichelpline.com/clinical-feed/british-journal-of-cancer-2-healthy-lifestyle-and-cancer-risk-a-pan-cancer-analysis-in-the-uk-biobank.md)

## Navigation
- [← Back to General Feed](https://medichelpline.com/clinical-feed/general.md)
- [← All Clinical Specialties](https://medichelpline.com/clinical-feed.md)
## Medical & Regulatory Disclaimer

> [!CAUTION]
> MedicHelpline content is structured for research, educational, and professional discovery purposes. It does not constitute individual medical advice, clinical diagnosis, or treatment recommendations.
> Always verify dosing, contraindications, and regulatory alerts against official product labeling and primary regulatory sources before clinical decision-making.