Lessons learned from modeling COVID-19 and steps to take at the start of the next pandemic focusing on ensemble models

Abstract The COVID-19 pandemic spurred many computational modeling efforts. Many mistakes were made, and many lessons were learned. This study attempts to list the key lessons learned from a modeling perspective, highlighting both the successes and shortcomings observed during the pandemic. Additionally, this work attempts to compile a set of critical steps and best practices that the authors believe would prove helpful and should be implemented before the start of the next pandemic to avoid inaccuracies in modeling pandemic scenarios with special attention to ensemble models. This will help to improve preparedness and ensure that computational models can more effectively guide decision-making in future pandemics. The paper counts 17 main recommendations for actions. The recommendations are primarily experience-driven. Those can be briefly summarized as improve data gathering, making resources accessible, and to use ensemble models to explain the pandemic.