An interaction Neyman-Scott point process model for coronavirus disease-19

Area of Science:

• Epidemiology
• Spatial Statistics
• Computational Biology

Background:

• Coronavirus disease 2019 (COVID-19) rapid transmission causes significant global mortality and societal challenges.
• Effective control of infectious disease outbreaks requires understanding spatial patterns of patient visits and identifying local cluster centers.
• Analyzing contact tracing data is complex due to strong clustering and intricate interactions between cluster centers.

Purpose of the Study:

• To develop a novel statistical model for analyzing spatial patterns in COVID-19 patient visit data.
• To accurately describe the complex interaction behaviors of disease cluster centers.
• To provide visualizations that can inform public health interventions for infectious diseases.

Main Methods:

• Development of a novel interaction Neyman-Scott process to model patient visit events as offspring from parent cluster centers.
• Implementation of an auxiliary variable algorithm within Markov chain Monte Carlo for model inference, addressing intractable likelihood functions.
• Fitting the model to simulated and real COVID-19 contact tracing data from Seoul under various outbreak scenarios.

Main Results:

• The proposed model effectively describes the spatial patterns of patient visits in COVID-19 contact tracing data.
• The model captures the complex attraction and repulsion dynamics between disease cluster centers.
• Demonstrated accuracy across different simulated and real-world outbreak scenarios.

Conclusions:

• The developed interaction Neyman-Scott process provides a robust method for analyzing infectious disease spatial epidemiology.
• Accurate modeling of patient visit patterns can significantly enhance the effectiveness of public health interventions, including social distancing measures.
• The findings offer valuable insights for controlling future infectious disease outbreaks through data-driven spatial analysis.