OH-MEMA: An Integrated One Health Mixed-Effects Modeling Approach for Syndromic Surveillance.

Area of Science:

• One Health
• Epidemiology
• Data Science

Background:

• Integrating diverse One Health time series data for surveillance is challenging due to fragmented tools.
• Existing workflows separate data preparation, modeling, and interpretation, hindering transparency and usability.

Purpose of the Study:

• Introduce OH-MEMA (One Health Mixed-Effects Modeling and Analytics), an interactive visual analytics framework.
• Integrate heterogeneous One Health data streams (clinical, environmental, wastewater) for enhanced syndromic surveillance and pandemic preparedness.

Main Methods:

• Web-based interface for uploading and analyzing multi-source datasets.
• Mixed-effects modeling with fixed and random effects, including spatial, temporal, and demographic variables.
• Visualizations of time series, evaluation metrics (MAE, RMSE, correlation), and analytic provenance via a model tree.

Main Results:

• Quantitative validation demonstrated robust predictive performance of mixed-effects models across counties and outcomes.
• Qualitative evaluation by experts showed improved interpretability, manageable workload, and effective workflow integration.
• System validated using NASA Task Load Index and open-ended interviews with epidemiologists and surveillance analysts.

Conclusions:

• OH-MEMA offers an interpretable, human-in-the-loop platform for exploratory forecasting and model analysis in syndromic surveillance.
• The framework successfully integrates data, modeling, and interpretation for user-centered decision-making in One Health.
• Supports analytical reasoning and enhances pandemic preparedness through unified data streams.