Story at a glance
- Scientists used random sampling methods to compare virus prevalence against national estimates.
- The research found that nearly 60 percent of infections may go unreported.
- This is attributed largely to flawed data collection by states.
As the CDC advises fully vaccinated people to wear masks in public spaces to halt the transmission of the contagious delta variant, new research suggests that current COVID-19 infections could be undercounted by as much as 60 percent.
Published in the Proceedings of the National Academy of Sciences (PNAS), the research estimated the presence of COVID-19 infections with a statistical model that includes variables like virus fatality data, confirmed cases and the number of tests administered daily.
The researchers initially focused on virus prevalence in Indiana and Ohio using several sources of data that all estimate the actual number of infections in these states along with the fatality rate. They largely relied on random sampling that was measured against other data sources.
The results estimated that as of March 7, 2021, 19.7 percent of the U.S. population, amounting to about 65 million people, were infected with COVID-19.
Up to that date, researchers estimate that one out of every 2.3 infections were confirmed via testing, with another infection being missed completely. This implies that 60 percent of all infections in the U.S. had been unreported and unaccounted for.
“Our results indicate that a large majority of COVID infections go unreported,” the report authors conclude. “This suggests that continued mitigation and an aggressive vaccination effort are necessary to surpass the herd-immunity threshold without incurring many more deaths due to the disease.”
Current national data report that more than 34 million positive COVID-19 cases have been reported since the onset of the pandemic, a smaller figure than what the authors posit.
In explaining this discrepancy, the study authors note that daily data can be unreliable due to differences in state reporting procedures, which are best captured at a moving average to account for fluctuations and errors.
They also call into question the model used by scientists in the U.S. Centers for Disease Control and Prevention, which is a multiplier model. The authors claims that this model bases its estimates on nationally reported laboratory-confirmed cases, which do not represent the population adequately.
“Reliable estimates of population prevalence are necessary for understanding the spread of the virus and the effectiveness of mitigation strategies,” the authors note.
These results put greater emphasis on the need to both continue building immunity against the virus, namely with a vaccination, but also underscore the importance of continuing wearing masks in public as the virus spreads potentially wider than anticipated.