Sentinel Event Surveillance to Estimate Total Sars-CoV-2 Infections, United States

Abstract

Human infections with a novel coronavirus (SARS-CoV-2) were first identified via syndromic surveillance in December of 2019 in Wuhan China. Since identification, infections (coronavirus disease-2019; COVID-19) caused by this novel pathogen have spread globally, with more than 180,000 confirmed cases as of March 16, 2020. Effective public health interventions, including social distancing, contact tracing, and isolation/quarantine rely on the rapid and accurate identification of confirmed cases. However, testing capacity (having sufficient tests and laboratory throughput) to support these non-pharmaceutical interventions remains a challenge for containment and mitigation of COVID-19 infections. We undertook a sentinel event strategy (where single health events signal emerging trends) to estimate the incidence of COVID-19 in the US. Data from a recent national conference, the Conservative Political Action Conference, (CPAC) near Washington, DC and from the outbreak in Wuhan, China were used to fit a simple exponential growth model to estimate the total number of incident SARS-CoV-2 infections in the United States on March 1, 2020, and to forecast subsequent infections potentially undetected by current testing strategies. Our analysis and forecasting estimates a total of 54,100 SARS-CoV-2 infections (80 % CI 5,600 to 125,300) have occurred in the United States to March 12, 2020. Our forecast predicts that a very substantial number of infections are undetected, and without extensive and far-reaching non-pharmaceutical interventions, the number of infections should be expected to grow at an exponential rate.

A comparative study of SIR Model, Linear Regression, Logistic Function and ARIMA Model for forecasting COVID-19 cases

“…Moreover, the percentage of the population infected in each city would be less than 0.01%, 0.05% and 0.02%, for South Korea, Italy, and Iran, respectively. In another research Lover and McAndrew [30] used the exponential growth model and epidemiological parameters from the epidemic in Wuhan, China to forecast cumulative infections in the United States. Their forecast results showed that a significant number of infections are undetected, and without considerable non-pharmaceutical interventions, the number of infections are expected to grow exponentially.…”

Stochastic modelling for predicting COVID-19 prevalence in East Africa Countries

“…Moreover, different statistical models have been proposed by different authors worldwide, to predict COVID-19 prevalence in different countries based on spread pattern of the disease. This includes models for China ( Abenvenuto et al., 2020 , Anastassopoulou et al., 2020 , Li et al., 2020 , Liu, Beeler, & Chakrabarty, 2020 , Liu, Magal, Seydi, & Webb, 2020 , Roosa et al., 2020 , Fanelli & Piazza, 2020 , Hu et al., 2020 , , Liu, Magal, Seydi, & Webb, 2020 , Wu et al., 2020 ), Italy ( Fanelli & Piazza, 2020 ; Grasselli, Pesenti, & Cecconi, 2020 ; Russo et al., 2020 ; Jia et al., 2020 ), France ( Fanelli & Piazza, 2020 ; Massonnaud et al., 2020 ), USA ( Liu, Beeler & Chakrabarty, 2020 ; Lover & McAndrew, 2020 ; Wise et al., 2020 ), South Korea ( Zhan et al., 2020 ; Kim, 2020 ), India ( Gupta & Pal, 2020 ) and Kenya ( Yoner et al., 2020 ). Generally, time series forecasting models like ARIMA is the standard technique which gives a decent predictions and forecasts on time series data in quick time.…”

An expert judgment model to predict early stages of the COVID-19 pandemic in the United States

“…As with previous outbreaks of other diseases [5][6][7], forecasts from computational models [8][9][10][11] assisted in planning and outbreak response near the beginning of the pandemic. However, given the initial limitations in testing capacity for SARS-CoV-2, these models were confronted with imperfect data with which to explain and predict viral transmission dynamics.…”