Government officials and policymakers have tried to use the numbers to understand the impact of COVID-19. Figures such as the number of hospitalizations or deaths reflect some of this burden. Each data point tells only part of the story. But no one explains the true prevalence of the new coronavirus by revealing the number of people actually infected at a given time – An important figure that will help scientists understand whether herd immunity can be achieved even with vaccines.
Now, two scientists from the University of Washington have developed a statistical framework that includes key COVID-19 data -; such as the number of cases and deaths due to COVID-19 -; To model the true prevalence of this disease in the United States and individual states. Their approach, published the week of July 26 in the journal Proceedings of the National Academy of Sciences, shows that 60% of COVID-19 cases in the US were undetected as of March 7, 2021, the deadline for which the dataset was created. they have employment.
This framework can help authorities determine the true disease burden in their region -; both diagnosed and undiagnosed -; and accordingly direct sources, the researchers said.
“There are all kinds of different data sources we can use to understand the COVID-19 pandemic – the number of hospitalizations in a state or the number of positive tests. But each data source has its own flaws. Senior author Adrian Raftery, professor of sociology and statistics at UW, explains what is really going on. “What we want to do is develop a framework that corrects flaws in multiple data sources and leverages its strengths to give us an idea of the prevalence of COVID-19 in a region, state, or country as a whole.”
Data sources can be biased in different ways. For example, one commonly quoted COVID-19 statistic is the proportion of positive test results in a region or state. However, Raftery said that this figure alone cannot provide a clear picture of the prevalence of COVID-19, as access to and willingness to be tested varies by location.
Other statistical methods often attempt to correct for bias in a data source to model the true prevalence of disease in a region. For their approach, Raftery and lead author Nicholas Irons, a UW doctoral student in statistics, combined three factors: the number of confirmed COVID-19 cases, the number of deaths due to COVID-19, and the number of COVID-19 tests administered, each as reported by the COVID Monitoring Project. like day. Additionally, they included results from random COVID-19 tests of Indiana and Ohio residents as an “anchor” for their method.
The researchers used their framework to model the prevalence of COVID-19 through March 7, 2021 in the US and each of the states. At that time, according to their framework, an estimated 19.7% of US residents, or about 65 million people, were infected. Raftery and Irons show that this is unlikely for the US to achieve herd immunity without the ongoing vaccination campaign. Additionally, there was an undercount factor of 2.3 in the US, the researchers found, meaning that only 1 in 2.3 COVID-19 cases is confirmed by testing. In other words, about 60% of cases were not counted at all.
According to Irons, this COVID-19 undercount rate also varies widely by state and may have multiple causes.
“It may depend on the severity of the pandemic and the amount of testing in this situation,” Irons said. “If you have a situation where there is severe pandemic but limited testing, the undercount may be very high and you are missing the vast majority of infections that occur. Or you may have a situation where testing is widespread and there is no pandemic there, the undercount rate would be lower.”
Raftery also said the undercount factor fluctuates by state or region as the pandemic progresses, due to differences in access to medical care between regions, changes in the availability of tests, and other factors.
With the actual prevalence of COVID-19, Raftery and Irons calculated other useful figures for states such as the infection mortality rate, which is the percentage of infected people who succumb to COVID-19, as well as the cumulative incidence. Percentage of a state’s population with COVID-19.
Ideally, regular random testing of individuals would show the level of infection in a state, region, or even nationally, Raftery said. But in the COVID-19 pandemic, only Indiana and Ohio conducted random viral testing of residents, datasets critical in helping researchers develop their framework. In the absence of widespread random testing, this new method could help authorities assess the true burden of disease in this pandemic and the next.
“We think this tool can make a difference by giving those responsible a more accurate picture of how many people are infected and how many of them are missed by current testing and treatment efforts,” Raftery said.