It is estimated that 14.3% of the US population had antibodies against COVID-19 in mid-November 2020, suggesting that the virus has infected many more people than reported, but still not enough to get close to the ratio needed to herd immunity, according to a study published today in Open JAMA network.
In the cross-sectional study, researchers from the study sponsors Pfizer and Merck analyzed data from random community seroprevalence surveys and five regional and national surveys from the Centers for Disease Control and Prevention (CDC) to estimate multipliers of infection underreporting. Seroprevalence surveys reveal the proportion of a population that has antibodies against a particular disease, such as COVID-19.
After adjusting for underreporting using validated multipliers, the analysis revealed an estimated median of 46,910,006 infections with SARS-CoV-2, the virus that causes COVID-19; 28,122,752 symptomatic infections; 956,174 hospitalizations; and 304,915 deaths from April to mid-November.
According to these figures, 14.3% of Americans had been infected with the new coronavirus on November 15, 8.6% had symptomatic infections, with an infection-hospitalization rate of 2.0% and a lethality rate for people with symptoms of 1.1%.
In contrast, the CDC reported 10,846,373 cases of COVID-19 and 244,810 deaths in the same period, with 1,037,962 cases recorded only in the last 7 days of that period (an average of 148,280 new cases reported daily).
Estimated 35% of unreported COVID deaths
The authors noted that seroprevalence surveys are essential to monitor progress towards herd immunity, which occurs when about 60% of the population has antibodies that can help delay transmission in the community. In comparison, the number of reported cases is underestimated due to many symptomatic people who do not seek medical exams or care and about 40% who have no symptoms.
“The population of the United States is still a long way from collective immunity, even with millions of new infections every week,” the researchers wrote. “The estimated number of deaths from COVID-19 is also notably higher than the deaths reported in the U.S. as of November 15, 2020, supporting the conclusion that approximately 35% of deaths from COVID-19 are not reported.”
They added that several methodological issues need to be considered when using seroprevalence surveys to derive underreporting multipliers in order to adjust disease surveillance data to estimate disease burden. These problems include the time from infection to the development of antibodies, decrease in antibodies and reporting of confirmed infections. Multipliers also change over time, with changes in the proportion of the population that has been tested, diagnosed and reported.
“A sensitivity analysis of CDC seroprevalence surveys suggested that using the number of cases reported at the end of the survey period provides a useful estimate of underreporting multipliers, particularly at the beginning of the pandemic,” said the authors. “Therefore, the number of cases reported on the last day of the seroprevalence survey was used to estimate the COVID-19 disease burden in this study.”
The researchers concluded that surveillance should continue to measure progress towards herd immunity. “Additional seroprevalence research is needed to monitor the pandemic, even after the development of safe and effective vaccines,” they said.