In some parts of the U.S., COVID-19 cases may be six to 24 times higher than reported figures, according to new antibody data from the CDC.

The data covers more than 16,000 blood samples collected from patients as part of routine care in 10 cities and states between March 23 and May 12. The CDC shared preliminary antibody data for six sites in June. Findings for all 10 sites — Connecticut, Louisiana, Minnesota, Missouri, New York City, Philadelphia, San Francisco, South Florida, Utah and Western Washington state — were published in JAMA July 21. 

The estimated proportion of people with COVID-19 antibodies ranged from 1 percent in the San Francisco Bay area (as of April 27) to 6.9 percent in New York City (as of April 1). 

The data suggests that many people who did not have symptoms or seek medical care were unknowingly spreading the virus in their communities, the CDC said. 

The analysis further solidifies the finding that current COVID-19 case numbers reflect a severe undercount. For example, an estimated 171,000 people in Missouri had been exposed to COVID-19 as of May 30, 13 times the number of reported cases in the state, according to the Johns Hopkins University Covid-19 dashboard

“For most sites, it is likely that greater than 10 times more SARS-CoV-2 infections occurred than the number of reported COVID-19 cases,” CDC researchers said. 

The data also shows that New York City, the hardest hit city in the analysis, is still far from achieving herd immunity.