New confinement index and new perspective for comparing countries - COVID-19

Area of Science:

• Epidemiology
• Data Science
• Public Health

Background:

• Comparing national COVID-19 lockdown measures and mortality presents significant challenges.
• Existing comparison strategies lack consensus, hindering effective analysis.

Purpose of the Study:

• To propose a novel comparison method for countries using three-dimensional trajectories.
• To overcome difficulties in comparing lockdown strategies and COVID-19 deaths.

Main Methods:

• Developed a confinement index using Google Community Mobility Reports and Principal Component Analysis.
• Applied longitudinal clustering to group European countries by COVID-19 fatalities and confinement levels.
• Utilized artificial longitudinal data to validate clustering methods and indices.

Main Results:

• The Calinski-Harabasz index proved superior for identifying the correct number of clusters.
• K-means clustering with Euclidean distance was the most effective method.
• Analysis revealed two distinct clusters for both mobility and fatality datasets.

Conclusions:

• Northern European countries exhibited higher mobility during initial lockdowns.
• COVID-19 fatalities showed a decline approximately 40 days after the first recorded death.