Area-specific autoencoder spatiotemporal graph neural networks for opioid overdose death prediction
JAMIA Open · 06.03.2026 · 9 (2)
DOI: 10.1093/jamiaopen/ooag063 ↗Kurz gefasst
Fachartikel in der Zeitschrift JAMIA Open, online erschienen am 29.04.2026. Verfasst von Xianhui Chen und 10 weiteren Autor:innen. Beteiligte Einrichtung: Computer Science and Engineering, The Ohio State University, Columbus, OH, United States. Die Arbeit nennt 22 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NIDA NIH HHS, National Institute on Drug Abuse, National Science Foundation.
Abstract (Originalsprache)
Background
Ohio has been severely impacted by the opioid crisis, with opioid overdose (OD) death rates exceeding national averages. Accurate OD death prediction supports proactive prevention and treatment allocation. Existing methods often focus on ZIP Code Tabulation Area (ZCTA)–level prediction for small-area resource allocation; however, performance at this resolution is poor due to substantial fluctuations in OD death counts, which introduce noise. This raises a critical methodological question: what is the optimal population threshold for OD death prediction that balances predictive accuracy with geographic resolution?
Methods
We perform a theoretical analysis of variance and error bounds to establish the minimal population required for robust prediction. Building on this analysis, we propose an Area-specific Autoencoder Spatiotemporal Graph Neural Network (AAE-STGNN) framework for OD death count prediction using urine drug test (UDT) data as dynamic features and Social Determinants of Health (SDoH) as static features. The framework consists of two key components: (1) an Area-specific Autoencoder (AAE), which learns latent spatial representations while incorporating the minimal population threshold, and (2) a Spatiotemporal Graph Neural Network (STGNN), which models geographic adjacency between areas and dynamic features across time.
Results
Empirical evaluations demonstrate that AAE-STGNN outperforms state-of-the-art (SOTA) approaches, achieving improved accuracy and robustness. We also provide the OD death count trend estimation to support public health decision-making.
Conclusions
These findings underscore the importance of selecting an optimal spatial granularity and leveraging spatiotemporal modeling techniques for data-driven public health surveillance and targeted intervention in the opioid crisis.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Computer Science and Engineering, The Ohio State University, Columbus, OH, United States.
- Department of Biomedical Informatics, The Ohio State University, Columbus, OH, United States.
- Center for Biostatistics, The Ohio State University, Columbus, OH, United States.
- Millennium Health, San Diego, CA, United States.
ORCID-Profile: Xianhui Chen
Förderung
- NIDA NIH HHS
- National Institute on Drug Abuse
- National Science Foundation
Zitieren
APA
Chen, X., Yin, C., Myers, J. V., Slover, B., Thomas, N., Marks, C., Kim, J., Fernández, S., Whitley, P., Fareed, N., & Zhang, P. (2026). Area-specific autoencoder spatiotemporal graph neural networks for opioid overdose death prediction. JAMIA Open, 9(2). https://doi.org/10.1093/jamiaopen/ooag063
Vancouver
Chen X, Yin C, Myers JV, Slover B, Thomas N, Marks C, et al. Area-specific autoencoder spatiotemporal graph neural networks for opioid overdose death prediction. JAMIA Open. 2026;9(2). doi:10.1093/jamiaopen/ooag063
BibTeX
@article{Chen2026,
author = {Chen, Xianhui and Yin, Changchang and Myers, John V and Slover, Brandon and Thomas, Neena and Marks, Charles and Kim, Joanne and Fernández, Soledad and Whitley, Penn and Fareed, Naleef and Zhang, Ping},
title = {Area-specific autoencoder spatiotemporal graph neural networks for opioid overdose death prediction},
journal = {JAMIA Open},
year = {2026},
volume = {9},
number = {2},
doi = {10.1093/jamiaopen/ooag063}
}
EUVIMED