Power spectral density and similarity analysis of COVID-19 mortality waves across countries

Area of Science:

• Epidemiology
• Data Science
• Public Health

Background:

• The COVID-19 pandemic prompted the creation of extensive mortality datasets by institutions like Johns Hopkins University CSSE.
• While global COVID-19 mortality trends were documented, detailed country-specific pattern comparisons were lacking.
• Understanding these variations is crucial for pandemic response and preparedness.

Purpose of the Study:

• To analyze the frequency domain of COVID-19 mortality waves across multiple countries.
• To quantify the similarities and differences in mortality patterns between nations using advanced analytical techniques.
• To identify recurring temporal patterns in pandemic-induced mortality.

Main Methods:

• Utilized the Fast Fourier Transform (FFT) to compute the Power Spectral Density (PSD) of COVID-19 mortality data.
• Analyzed data from 199 countries spanning January 2020 to March 2023.
• Applied cosine similarity analysis to compare PSD patterns across all studied countries.

Main Results:

• Identified two primary frequencies in global COVID-19 mortality: one cycle approximately every 10.4 months (1.15 waves/year) and another every 4.4 months (2.7 waves/year).
• Cosine similarity analysis indicated a high degree of similarity (global median index of 0.84) in the dominant mortality wave frequencies across countries.
• The distribution of similarity indices showed a negative skew (-0.54), suggesting a general trend towards similar mortality wave characteristics.

Conclusions:

• The study reveals consistent, cyclical patterns in COVID-19 mortality waves globally.
• These findings offer valuable insights for planetary health, aiding in the planning of interventions for future pandemics.
• Understanding these periodicities can inform the timing and deployment of public health strategies, such as confinement measures.