Forecasts of Covid-19 evolution by nearest epidemic trajectories detection

Area of Science:

• Epidemiology
• Public Health
• Data Science

Background:

• The COVID-19 pandemic necessitates accurate short-term forecasting for effective public health interventions.
• The emergence of new variants, such as the B.1.1.7 (UK) variant, has altered epidemic dynamics.
• Decentralized epidemic management requires localized forecasting capabilities.

Purpose of the Study:

• To develop a robust method for short-term forecasting of COVID-19 epidemics in small administrative units (districts).
• To demonstrate the applicability of the method using data from three Polish cities during the third epidemic wave.
• To provide a tool for informed, local decision-making regarding anti-COVID-19 restrictions.

Main Methods:

• Utilizing historical epidemic evolution data to identify similar patterns.
• Applying pattern recognition algorithms for short-term epidemic forecasting.
• Analyzing epidemic data from March-April 2021 in Poland, focusing on the B.1.1.7 variant's impact.

Main Results:

• Successful one- and two-week COVID-19 forecasts were generated for selected Polish districts.
• Observed differences in epidemic progression during the third wave compared to previous waves, attributed to the B.1.1.7 variant.
• The forecasting algorithm proved effective in identifying localized epidemic trends.

Conclusions:

• The proposed method offers a reliable approach for short-term COVID-19 forecasting at the district level.
• Localized forecasting enables adaptive management of public health interventions, including the timely implementation or release of restrictions.
• This approach supports efficient, data-driven epidemic control strategies in small administrative units.