COVID¹⁹ First Wave Mortality Dynamics

The first 4 months of COVID-19 pandemic in the US are characterized by mortality, caused directly by SARS-2 infection as well as by other casues, that change as social isolation policies are adopted.

Abstract

The epidemiology dynamics of COVID-19 have changed markedly with the adoption of large scale isolation. Improvement of therapeutic approaches, more granular confinment practices and social distancing policies, are changing both the mortality and infection progression profiles of new cases. The complexity of the process is highligthed by the effect on additional mortality by causes other than COVID, such as Heart Disease and Altzheimer. In the United States, the new dynamics identify a discontinuity at the end of June we will use here to mark the end of a first wave. The interactive representation of the non-linear dynamics of this 4 month period are the main topic of this report. Specifically, the time-delayed embedding and compression of the multivariate timeseries is performed by a neural network visualization designed to explore relationships between causes of mortality. It is also hoped that the comparison of 2020 data with the previous 5 years for this initial period provides a baseline and a methodology to explore the shifting dynamics of COVID-19 progression in subsequent months and years.

Data Preparation

The bulk of the data used in this study comes from The Centers for Diseases Control and Prevention, CDC. Since the analysis compares 2020 with the previous 5 years, 2015-2019, the data preparation required the aggregation of two sources, 2014-2018 and 2019-2020. The data preparation process is not trivial, as variable names change over time and privacy concerns flag out counts between 1 and 6 for any given mortality category. In order to facilitate this process, a data browser tool was developed and is made publicly available at episphere.github.io/mortalitytracker. A third data source of data for COVID mortality is used by this tool to assess the consistency of the counts . That same source, at Johns Hopkins University, provides the total confirmed case count. For the sake of permanent reproducibility, the

... ...