Exploring COVID-19 Daily Records of Diagnosed Cases and Fatalities Based on Simple Nonparametric Methods

Area of Science:

• Epidemiology
• Mathematical Modeling
• Public Health

Background:

• Effective containment of epidemics like COVID-19 relies on accurate epidemiological parameters, such as the effective reproduction number.
• Traditional parametric models (e.g., SIR) have limitations due to time-dependent parameters and delayed reporting of fatalities.
• Periodic patterns in incidence time series can arise from disease transmission cycles.

Purpose of the Study:

• To estimate crucial epidemiological parameters for the COVID-19 pandemic, focusing on Germany, using nonparametric methods.
• To compare these estimates with data from seven other countries.
• To analyze the time-dependency of parameters and their impact on epidemic dynamics.

Main Methods:

• Utilized nonparametric methods, including time-delay correlation analyses, on publicly available time-series data for prevalence and fatalities.
• Calculated instantaneous and asymptotic fatality-case ratios based on time lags between diagnoses and deaths.
• Estimated the basic reproduction number (R0) and instantaneous reproduction number using delay autocorrelation analysis.

Main Results:

• Identified a highly accurate 13-day duration from diagnosis to death in Germany and Switzerland.
• Calculated a temporal median instantaneous fatality-case ratio of 0.02 for Germany with a 13-day lag.
• Estimated the basic reproduction number (R0) for Germany to be between 2.4 and 3.4, with observed resonances potentially linked to weekly sampling patterns.

Conclusions:

• Nonparametric methods provide reliable estimates for time-dependent epidemiological parameters, overcoming limitations of traditional models.
• The study highlights the consistency between the time course of the instantaneous reproduction number and per capita growth.
• Comparative analysis across eight countries reveals convergence in fatality-case ratios later in the pandemic.