Mapping Incidence and Prevalence Peak Data for SIR Modeling Applications.

Area of Science:

• Epidemiology
• Mathematical Biology
• Computational Science

Background:

• Infectious disease modeling is crucial for epidemic response.
• Compartmental models like SIR are used for forecasting disease dynamics.
• Early data can lead to unstable model fits and unrealistic forecasts.

Purpose of the Study:

• To develop a novel method for incorporating peak hospitalization data into SIR model fitting.
• To enhance the accuracy and stability of infectious disease forecasting.
• To assess the impact of using incidence versus prevalence data in modeling.

Main Methods:

• Formulated a system of two equations to computationally solve for SIR model parameters based on peak incidence.
• Updated the Dirichlet-Beta State Space modeling framework to utilize hospital incidence data.
• Assessed the method's accuracy and computational speed via simulation.

Main Results:

• The new method allows for the calculation of SIR parameter estimates, including transmission and recovery rates.
• Misspecifying prevalence data as incidence data leads to a noticeable loss in accuracy.
• The updated Dirichlet-Beta State Space model demonstrates practical potential for forecasting using hospital incidence data.

Conclusions:

• Incorporating peak hospitalization data significantly stabilizes compartmental model fits.
• Accurate data type specification (incidence vs. prevalence) is critical for reliable infectious disease modeling.
• The developed method offers a practical and accurate approach to infectious disease forecasting.