|
?:abstract
|
-
COVID-19 is extraordinary both as once-in-a-lifetime pandemic and having abundant real-time case data, thus providing an extraordinary opportunity for timely independent analysis and novel perspectives. We investigate the weekly periodicity in the daily reported new cases and new deaths with the implied relationships to the societal and institutional responses using autocorrelation and Fourier transformation. The results show significant linear correlations between the weekly periodicity and the total cases and deaths, ranging from 50% to 84% for sizable groups of countries with population normalized deaths spanning nearly three orders of magnitude, from a few to approaching a thousand per million. In particular, the Strength Indicator of the periodicity in the new cases, defined by the autocorrelation with a 7-day lag, is positively correlated strongly to the total deaths per million in respective countries. The Persistence Indicator of the periodicity, defined as the average of three autocorrelations with 7-, 14- and 21-day lags, is an overall better indicator of the progression of the pandemic. For longer time series, Fourier transformation gives similar results. This analysis begins to fill the gap in modeling and simulation of epidemics with the inclusion of high frequency modulations, in this case most likely from human behaviors and institutional practices, and reveals that they can be highly correlated to the magnitude and duration of the pandemic. The results show that there is significant need to understand the causes and effects of the periodicity and its relationship to the progression and outcome of the pandemic, and how we could adapt our strategies and implementations to reduce the extent of the impact of COVID-19.
|
![[This page as RDF]](https://research.tib.eu/covid-19/static/rdf-icon.gif){kind=icon}