News: Machine learning model shows higher COVID-19 rates than reported

Researchers estimate that since the start of the pandemic, the number of COVID-19 cases in the United States is nearly three times higher than the reported confirmed number, according to a recent study published in PLOS ONE based on a machine learning algorithm.

The algorithm provides daily updated estimates of total infections to date, as well as the count for how many people are currently infected nationwide and in the 50 countries hardest hit by the pandemic. According to the model, as of February 4, 2021, more than 71 million people in the United States alone had contracted COVID-19. The publicly reported confirmed cases, however, only numbered 26.7 million, Health Analytics reported.

The disparities between confirmed cases and estimates produced by the algorithm for the other 50 countries varied. For example, on February 4, the algorithm estimated 36 million cumulative cases in Brazil, which was almost four times higher than the 9.4 million confirmed cases reported. Likewise, the algorithm estimated 14 million cases in France versus the 3.2 million confirmed, 25 million estimated versus four million confirmed in UK, and 27.6 million estimated versus 1.9 million confirmed in Mexico.

The model relies on COVID-19 death data from Johns Hopkins University and the COVID-19 Tracking Project. The number of reported deaths is thought to be more accurate than the number of lab-confirmed cases. The model assumes an infection fatality rate of 0.66%, based on an earlier study of the pandemic in China, and considers factors like the average number of days from the onset of symptoms to death or recovery.

Experts are still uncertain about the true death rate of COVID-19, so the estimates are rough, but researchers believe the model’s estimates are accurate and miss fewer cases than the confirmed ones currently used to guide public health policies.

Researchers checked their findings against existing prevalence rates found in other studies that used blood tests to check for antibodies to the virus causing COVID-19, Health Analytics reported. For more of the areas tested, the algorithm’s estimates of infections closely corresponded to the percentage of people who had tested positive for antibodies.

“Knowing the true severity in different regions will help us effectively fight against the virus spreading,” said Jungsik Noh, PhD, a UT Southwestern assistant professor in the Lyda Hill Department of Informatics and the first author of the study. “The currently infected population is the cause of future infections and deaths. Its actual size in a region is a crucial variable required when determining the severity of COVID-19 and building strategies against regional outbreaks.”

Editor’s note: To read the full study, click here. To read Health Analytics’ coverage of this story, click here. To read ACDIS’ coverage of the COVID-19 pandemic, click here.