Panozzo, Kimberly; Rahman, MD Ishfaq Ur; Ames, April L; Saltzman, Barbara S; Alam, Bhuiyan M
description
Abstract Monitoring the spread of the emerging Covid-19 pandemic has taken the world by storm. From the initial reports of a novel coronavirus spreading in Wuhan, China, to the Fall of 2020, people around the world have turned to GIS and online maps to track its spread. To control the pandemic, scientists have begun to analyze and model the spread of cases and evaluate significant factors that may influence its diffusion. Unfortunately, many early Covid-19 cases went undetected due to lack of testing and disparities in access. Additionally, testing rates and efficacy are not consistent across geographies and time. For these reasons, the 'official' confirmed Covid-19 cases represent a biased and considerable undercount. This study aims to estimate the weekly number of Covid-19 cases at the county level by (1) estimating the number of expected deaths for 2020 using the National Center for Health Statistics Mortality data from 2014 to 2018, (2) calculating the difference between expected deaths and observed deaths at each count by week and (3) use most current case fatality rate (CFR) to estimate the number of contracted cases of Covid-19 (lagged 18 days to represent approximate date of transmission). Tentative results show over 800,000 deaths above normal as of September 1<sup>st</sup><span> </span>with significant spatial differences across both time and space. Results also show considerable differences between excess deaths calculate at the state level to excess deaths using county level data. In sum, cases of Covid-19 are significantly undercounted across locations.
