About 9% of Americans exposed to COVID-19 by midsummer. That's a long way from herd immunity.

By the end of July, about 9% of American adults had been exposed to the coronavirus that causes COVID-19, according to a new study of dialysis patients, the largest yet looking for evidence of the disease in people's blood.

The infection rates varied from essentially zero in some states that avoided infection by midsummer, to more than one-third of residents in parts of New York hard-hit in the spring.

The upshot is the American public is a long way from achieving "herd immunity" – having enough infections to prevent further spread of the virus.

Infection rates are so patchy that even if some areas have had high infection rates, people there are unlikely to be protected because others will bring the virus in from elsewhere, said William Hanage, an epidemiologist at the Harvard T.H. Chan School of Public Health.

Hanage is also concerned communities not yet hit by COVID-19 will feel a false sense of security.

"We expect small-town America not to be in the first surge," he said, but over time, as people move around more and it gets introduced multiple times, more communities will have outbreaks.

Dr. George Rutherford, an epidemiologist and biostatistician at the University of California, San Francisco, was surprised the national infection rate appeared so high. He would have expected closer to 3%-4%. But whatever the number, he said, the U.S. is nowhere close to an end to this pandemic.

"The only way we're going to get to herd immunity, unless you're in a very closed community like a prison, is for everybody to get vaccinated," Rutherford said.

Hanage said the study also confirms what he has long thought: that the outbreak in the spring was enormously underestimated, "ridiculously so," and started earlier and infected more young people than have been diagnosed.

The new study, published in The Lancet, is in line with previous, smaller studies, and also showed areas with high numbers of Black and non-white Latino residents had higher infection rates than mostly white communities.

Stephen Morse, an epidemiologist at the Mailman School of Public Health at Columbia University, said the study is an important contribution at a time when there isn't much good information about who has been infected with the virus.

"The sad truth is that we don’t have reliable national data," Morse said via email. "Testing for public health surveillance is an enormous gap right now. We test mostly sick people, and others who are somehow able to get a test, and access can vary tremendously even in the same city. This isn’t very helpful for public health."

To understand an epidemic and help slow it, epidemiologists need lots of information at as local a level as possible, he said. But for most of the United States, "we have very little knowledge of what the virus is doing in that community."

The Stanford University researchers, who led the new study, looked for antibodies to COVID-19 in the blood of people who receive dialysis in clinics or at home. When someone mounts an immune response to an infection, their body produces antibodies that can be read to determine their infection history.

Dialysis patients, whose kidneys are failing, get routine blood tests, so researchers were able to use existing blood samples. The patients were treated at 1,300 dialysis centers in 46 U.S. states, representing about one-third of counties.

Dialysis patients tend to be older than the general population, and more diverse, as Blacks and Latinos are more likely to suffer from kidney disease than whites.

Dr. Shuchi Anand, a kidney specialist at Stanford and the paper's first author, said she thinks the dialysis patients provide a useful and interesting – though imperfect – snapshot of the COVID-19 outbreak.

The patients might have been more careful to avoid infection but also more likely to suffer severe consequences if they were infected, said Dr. Glenn Chertow, the paper's senior author and a Stanford nephrologist.

"It should be a pretty good estimate of the population in the United States with a little bit of tweaking around the edges," he said.

The data was most precise for Texas, New York and California, all of which had very large outbreaks, said a third author, biostatistician Maria Montez-Rath. Roughly one-third of New York patients had antibodies, compared to 3.6% of Texans and 3.8% of Californians.

The team also plans to continue looking at samples from dialysis patients to watch COVID-19 over time, Anand said. "It will give us a good sense of the epidemic as it evolves,” she said.

Contact Karen Weintraub at kweintraub@usatoday.com

Health and patient safety coverage at USA TODAY is made possible in part by a grant from the Masimo Foundation for Ethics, Innovation and Competition in Healthcare. The Masimo Foundation does not provide editorial input.

This article originally appeared on USA TODAY: Herd immunity study: US infection rate too low to prevent COVID spread