Integrating Spatiotemporal Features in LSTM for Spatially Informed COVID-19 Hospitalization Forecasting

2025-06-06

AI Summary

To address the insufficient robustness of state-level hospitalization forecasting during COVID-19 variant outbreaks, this paper proposes an LSTM-based forecasting framework integrating social proximity modeling and multi-step ensemble learning. We innovatively construct a spatial interaction feature—State Proximity via Facebook’s Social Connectedness Index (SPH)—to capture inter-state population mobility dynamics. A parallel LSTM architecture is designed to jointly model long- and short-term temporal dependencies, while a multi-horizon consistency-weighted ensemble strategy enhances prediction stability. During the Omicron surge, our model reduces mean absolute error by 27, 42, 54, and 69 hospitalizations per state over 7-, 14-, 21-, and 28-day horizons, respectively, compared to the CDC’s ensemble model. Ablation studies confirm SPH’s substantial contribution to accuracy improvement. This work establishes a novel paradigm for spatially aware time-series forecasting in public health emergency response.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Intelligent Robots: State Estimation
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

The COVID-19 pandemic's severe impact highlighted the need for accurate, timely hospitalization forecasting to support effective healthcare planning. However, most forecasting models struggled, especially during variant surges, when they were needed most. This study introduces a novel Long Short-Term Memory (LSTM) framework for forecasting daily state-level incident hospitalizations in the United States. We present a spatiotemporal feature, Social Proximity to Hospitalizations (SPH), derived from Facebook's Social Connectedness Index to improve forecasts. SPH serves as a proxy for interstate population interaction, capturing transmission dynamics across space and time. Our parallel LSTM architecture captures both short- and long-term temporal dependencies, and our multi-horizon ensembling strategy balances consistency and forecasting error. Evaluation against COVID-19 Forecast Hub ensemble models during the Delta and Omicron surges reveals superiority of our model. On average, our model surpasses the ensemble by 27, 42, 54, and 69 hospitalizations per state on the $7^{th}$, $14^{th}$, $21^{st}$, and $28^{th}$ forecast days, respectively, during the Omicron surge. Data-ablation experiments confirm SPH's predictive power, highlighting its effectiveness in enhancing forecasting models. This research not only advances hospitalization forecasting but also underscores the significance of spatiotemporal features, such as SPH, in refining predictive performance in modeling the complex dynamics of infectious disease spread.

Problem

• Develops LSTM model for COVID-19 hospitalization forecasting
• Incorporates spatiotemporal features to improve prediction accuracy
• Addresses limitations of existing models during variant surges

Innovation

• Spatiotemporal LSTM for COVID-19 hospitalization forecasting
• Social Proximity to Hospitalizations (SPH) feature
• Multi-horizon ensembling strategy balances accuracy