A lot of Americans have been infected with SARS-CoV-2. Extrapolating from a new study suggests that the number may be as high as 60 million. Unfortunately, herd immunity is still a long way off.
What is herd immunity?
Herd immunity occurs in a population when a sufficient number of people have acquired protective immunity to a pathogen that active cases are not likely to encounter people who remain susceptible before recovering themselves.
Herd immunity can be obtained either through vaccination or natural infection. The key is that the immune response must protect against infection to a sufficient degree that a person is not able to transmit the pathogen to another person.
There’s a nice illustration on youtube:
The threshold at which a population acquires herd immunity depends on other interventions. Thus, the herd immunity threshold will be lower in a population where isolation, active case finding, and asymptomatic testing are practiced compared to one where they are not. Experts estimate herd immunity for SARS-CoV-2 to be around 70% of the population.
The new study
How close is the US to this 70% threshold? A new study by Shuchi Anand and colleagues from Stanford University allows us to estimate. Recovery from an immunizing infection like SARS-CoV-2 results in the development of antibodies that can be detected in blood serum. A person who tests positive for antibodies is said to have “seroconverted” or to be “seropositive”. The new study is the largest study of seroconversion in the US to date and was based on analysis of blood plasma from 28,503 randomly selected adults receiving dialysis in July.
Helpfully, the presentation of results used demographic characteristics of the US population and statistical techniques to standardize observed seropositivity in the sample to what it would be in the entire US adult population, meaning that the disproportionate representation of older people, men, and people living in majority Black and Hispanic neighborhoods who receive dialysis were accounted for in extrapolating to the population as a whole. The key finding was that an estimated 8,989 out of every 100,000 Americans — 9.3% — had been infected by early July.
So, how many Americans have been infected with SARS-CoV-2 now?
The paper does not actually give us quite enough information to answer this question. For instance, children are not included in their sample. But some back of the envelope calculations provide an approximation. For instance, if we assume that children have been infected at the same rate as adults and follow Anand et al. in using the 2018 American Community Survey to estimate that the American population comprises 326,274,356 people, we conclude that 29,328,801 (almost 30 million) Americans had been infected at the time of the study. What about today? The US is currently closing in on 7 million diagnosed cases compared to about half that on July 15, i.e. the epidemic is currently double what it was at the time of the study. (These numbers can be obtained from the Johns Hopkins Coronavirus Resource Center or the University of Georgia’s Covid-19 Tracker.) If we double the estimate from July, we conclude that around 60 million Americans — about 18.3% of the population — have now been infected with SARS-CoV-2.
It’s possible that this is an overestimate due to the increased availability of tests in recent months compared to early in the pandemic. However, the Atlantic’s COVID Tracking Project shows that overall test positivity — the fraction of tests administered that are found to be positive — has only dropped one percentage point from 8% to 7% from July 15 to September 25. So it seems unlikely that case ascertainment, the fraction of infections that are detected, has changed greatly.
Another complication is that it is possible there were people who acquired the infection early in the epidemic and have since lost their immunity. However, there were relatively few cases of SARS-CoV-2 in the US until late spring and immunity to coronaviruses typically lasts at least six months, so any bias due to waning immunity should be very small.
Who is getting infected?
One of the most alarming findings concerns who is getting infected. Anand et al. found that residents of non-Hispanic Black and Hispanic neighborhoods were about four times more likely to be seropositive compared with people from non-Hispanic white neighborhoods. This is unsurprising as we know that communities of color have been disproportionately affected by Covid-19, but the new study quantifies these deep health disparities for the first time.
Seropositivity was also significantly higher in cities and other high population density communities, characterized as neighborhoods in the top 20% of all neighborhoods. This probably reflects both higher transmission in such settings and the fact that the virus arrived in population centers earliest and so therefore these communities have been battling the epidemic longer.
The authors themselves summarized the main point this way:
“We confirm… a minority of the population has evidence of exposure and immune response, and a vast majority, including people at high risk for mortality (ie, the population on dialysis), remain vulnerable.”
The key takeaways are
- By July 2020, fewer than 1 in 10 Americans were infected by Sars-CoV-2; it may be double that now
- Overall, fewer than 1 in 10 Americans infected with SARS-CoV have been diagnosed
- SARS-CoV-2 infections, both diagnosed and undiagnosed, have disproportionately affected communities of color and urban areas.
That is, the US is far from herd immunity and many of the most affected are those who were the least advantaged to begin with.