Comparing three basic models for seasonal influenza

Area of Science:

• Epidemiology
• Mathematical Modeling
• Infectious Disease Dynamics

Background:

• Seasonal influenza poses a significant public health challenge, with distinct transmission patterns potentially existing between influenza A and B strains.
• Understanding these differences is crucial for accurate forecasting and effective public health interventions.

Purpose of the Study:

• To compare three mathematical models of seasonal influenza transmission using the Spatiotemporal Epidemiological Modeler (STEM).
• To assess the impact of varying transmission parameters (magnitude, background rate, season length) on model accuracy for influenza A and B.
• To evaluate the predictive capabilities of different influenza transmission models.

Main Methods:

• Utilized the open-source Spatiotemporal Epidemiological Modeler (STEM) software.
• Developed and compared three distinct models of seasonal influenza transmission, varying assumptions about strain-specific parameters.
• Optimized models using 10 years of surveillance data from Israel and employed cross-validation for accuracy assessment.

Main Results:

• Models incorporating variations in transmission amplitude (maximum transmissibility) between influenza A and B demonstrated increased predictive accuracy compared to a baseline model.
• Allowing for further variations in the shape of the seasonal forcing function yielded minimal improvement in predictive ability.
• The study highlights the importance of transmission magnitude differences in modeling influenza dynamics.

Conclusions:

• Accounting for differences in transmission amplitude is key to enhancing the predictive accuracy of seasonal influenza models.
• While transmission magnitude is important, variations in the seasonal forcing function's shape have a limited impact on predictive performance.
• These findings can inform the development of more robust influenza forecasting systems.