Government officials and policymakers try to use numbers to understand the impact of COVID-19. Figures such as hospitalizations or deaths reflect part of this burden. Each data point only tells part of the story. But no one describes the true universality of the new coronavirus by revealing the actual number of people infected at a given time-this is an important number to help scientists understand whether herd immunity can be achieved, even if vaccinated.

Now, two scientists at the University of Washington (UW) have developed a statistical framework that combines key COVID-19 data (such as the number of cases and deaths caused by COVID-19) to simulate the disease in the United States and individual State’s true prevalence. Their method was published in the Proceedings of the National Academy of Sciences in the week of July 26. It is estimated that as of March 7, 2021 (the last date of the data set), as many as 60% of the COVID-19 cases are not found and they are available for employment.

Researchers say this framework can help officials determine the true burden of disease in their area—confirmed and undiagnosed—and channel resources accordingly.

"We can use a variety of different data sources to understand the COVID-19 pandemic-the number of people hospitalized in a state, or the number of people who test positive. But each data source has its own flaws, which can lead to What actually happened gives a biased picture," said senior author Adrian Lovetree, a professor of sociology and statistics at the University of Washington. "What we want to do is to develop a framework to correct the flaws in multiple data sources and use their advantages to let us understand the prevalence of COVID-19 in a region, a state, or an entire country."

The data source may be biased in different ways. For example, a widely cited COVID-19 statistic is the proportion of positive test results in a certain region or state. Raftery said that because access to and willingness to be tested varies from location to location, this number alone does not provide a clear indication of the COVID-19 epidemic.

Other statistical methods usually try to correct deviations in a data source to simulate the true prevalence of diseases in a region. For their method, Raftery and lead author Nicholas Irons, a doctoral student in statistics at the University of Washington, considered three factors: the number of confirmed COVID-19 cases, the number of deaths due to COVID-19, and the number of COVID-19 tests performed. Track project reports every day. In addition, they used random COVID-19 test results from residents of Indiana and Ohio as the "anchor" of their approach.

The researchers used their framework to model the COVID-19 epidemic in the United States and each state as of March 7, 2021. According to their framework, on that day, an estimated 19.7% of American residents, or about 65 million people, were infected. Raftery and Irons said this shows that without an ongoing vaccination campaign, the United States is unlikely to achieve herd immunity. In addition, the researchers found that the underestimation coefficient in the United States is 2.3, which means that only one of approximately 2.3 COVID-19 cases has been confirmed by the test. In other words, about 60% of cases are not counted at all.

According to Irons, this underestimation rate of COVID-19 also varies from state to state, and there may be multiple reasons. 

Irons said: "It may depend on the severity of the pandemic and the amount of testing in the state." "If your state has a severe pandemic but limited testing, the undercount may be very high and you will miss out Most infections are occurring. Or, you may encounter a situation where testing is common and the pandemic is not as severe. There, the underestimation rate will be lower."

In addition, due to differences in access to medical services between regions, changes in test availability, and other factors, as the pandemic develops, the undercounting factors vary from state to state or region, Lovetree said.  

Based on the true prevalence of COVID-19, Raftery and Irons calculated other useful data for each state, such as infection mortality, which is the percentage of infected people who died of COVID-19, and cumulative incidence, which is the state population with COVID-19 percentage.

Raftery said that, ideally, regular random tests of individuals will show the level of infection in a state, region, or even the whole country. But in the COVID-19 pandemic, only Indiana and Ohio have conducted random virus tests on residents, and these data sets are critical to helping researchers develop their frameworks. In the absence of extensive random testing, this new method can help officials assess the true burden of disease in this pandemic and the next.

Raftery said: "We think this tool can be useful by allowing the person in charge to more accurately understand how many people have been infected and how many people are missed by current testing and treatment efforts." 

-This press release was originally published on the University of Washington website

Tags: Coronavirus Epidemiology Industry News Model and Simulation Product Resources: Industry News

© 2021 Laboratory Manager. all rights reserved.