Scenario construction and evolutionary analysis of nonconventional public health emergencies based on Bayesian networks.

Objectives: The objective was to aggregate the various scenarios that occur during nonconventional public health emergencies (NCPHEs) and analyze the evolutionary patterns of NCPHEs to better avoid risks and reduce social impacts. The aim was to enhance strategies for handling NCPHEs.

Study Design: News reports were crawled to obtain the scenario elements of NCPHEs and categorized into the spreading stage or derivation stage. Finally, the key scenario nodes and scenario evolution process were analyzed in combination with a corresponding emergency response assessment of each scenario by experts.

Methods: Dempster-Shafer (DS) theory and Bayesian networks (BNs) were applied for data reasoning, and a spread-derived coupled scenario-response theoretical model of NCPHEs for major public health emergencies was constructed. The scenario evolution path of COVID-19 was derived by combining seven types of major scenario states and corresponding emergency response measures extracted from 952 spreading scenarios.

Results: The 26 NCPHE spread scenarios and 41 NCPHE derivation scenarios were summarized. Optimized and pessimistic NCPHE scenario pathways were generated by combining the seven major spreading scenarios to help decision makers predict the development of NCPHEs and take timely and effective emergency response measures for key scenario nodes.

Conclusion: This study provides a new approach for understanding and managing NCPHEs, emphasizing the need to consider the specificity and complexity of such emergencies when developing decision-making strategies. Our contextual derivation model and emergency decision-making system provide practical tools with which to enhance NCPHE response capabilities and promote public health and safety.