A history-dependent approach for accurate initial condition estimation in epidemic models

Area of Science:

• Mathematical modeling of dynamical systems
• Epidemiology and public health
• Computational biology and bioinformatics

Background:

• Mathematical models are crucial for understanding complex systems like disease spread.
• Accurate initial conditions are vital for reliable model predictions, but often unknown.
• Current methods for estimating initial conditions in infectious disease models can be biased.

Purpose of the Study:

• To develop and validate a history-dependent method for estimating initial conditions in infectious disease models.
• To address the limitations of history-independent assumptions in initial condition estimation.
• To improve the accuracy and reliability of epidemic modeling.

Main Methods:

• Developed a history-dependent initial condition estimation method based on a master equation.
• Modeled the time-varying likelihood of becoming infectious during a latent period.
• Compared the new method against history-independent approaches using simulated and real-world data.

Main Results:

• The history-dependent method significantly reduced estimation bias compared to history-independent methods.
• The method demonstrated robustness across scenarios with measurement errors and epidemic shifts (e.g., vaccination).
• A 55% reduction in estimation error was observed using COVID-19 data from Seoul, Republic of Korea.

Conclusions:

• The history-dependent method provides a more accurate estimation of initial conditions for infectious disease models.
• Improved initial condition estimation enhances epidemic model precision, aiding public health policy.
• A user-friendly package, Hist-D, is available to implement this advanced estimation technique.