State-of-the-art learning COVID-19 vaccine effectiveness using LSTM

Area of Science:

• Epidemiology
• Computational Biology
• Public Health

Background:

• COVID-19 vaccines are critical tools for mitigating disease severity.
• Understanding vaccine effectiveness, especially with evolving variants and booster recommendations, remains essential.

Purpose of the Study:

• To evaluate the impact of COVID-19 vaccine administrations, including booster doses, on hospitalization risks.
• To assess the effectiveness of adherence to vaccine policies in reducing infection and hospitalization rates.

Main Methods:

• Utilized a Long Short-Term Memory (LSTM) model to create a dynamic environment simulating COVID-19 vaccine effects.
• Generated hypothetical vaccination profiles to predict counterfactual outcomes and estimate vaccine effectiveness.
• Analyzed real-world data from May 2021 to April 2023.

Main Results:

• Booster doses were found to significantly reduce the risk of COVID-19 hospitalization.
• Bivalent boosters demonstrated comparable or slightly superior effectiveness to monovalent boosters.
• Adherence to vaccine policies was highlighted as a key factor in reducing hospitalizations.

Conclusions:

• COVID-19 vaccine boosters are effective in preventing hospitalizations.
• Reinforces the importance of public health strategies emphasizing vaccine policy compliance.
• Provides data-driven insights into vaccine dynamics to support informed public health decision-making.