A pair of studies conducted by researchers in Santa Clara and Los Angeles counties drew national attention this week with a finding that COVID-19 is much more widespread — and far less deadly — than public health officials had assumed.
But the studies, which suggested that the fatality rate of COVID-19 could be less than 0.2%, immediately faced a backlash from other scientists, as well as words of caution from public health administrators, who indicated that they will not reopen the economy in a meaningful way any time soon.
The Stanford study, which was led by Assistant Professor Eran Bendavid, asked more than 3,000 volunteers for help in answering a critical question: "How many people in Santa Clara County have been infected by COVID-19?"
Using finger-prick tests and kits that resemble pregnancy tests, the team took blood samples at three drive-thru sites on April 3 and 4, crunched the numbers and published the early draft of the results on a preprint server, giving other researchers a chance to review the results before they are formally published.
In analyzing the blood samples, Bendavid's team found that 50 of 3,330 tested positive for COVID-10, a rate of 1.5%. After adjusting the results to account for precision of tests and sample bias, the study concluded between 48,000 and 81,000 county residents have been infected, a rate of between 2.49% and 4.16%. This led the team to conclude that the number of COVID-19 infections in Santa Clara County is 50 to 80 times higher than the number of confirmed cases.
The Stanford study calculated the mortality rate of COVID-19 as between 0.12% and 0.2% when one considers the true number of infections. That's far lower than the roughly 4.8% rate one would get by only looking at the county's confirmed cases, which stood at 1,987 as of Thursday afternoon.
The study in Los Angeles, which was undertaken by the University of Southern California and the Los Angeles Department of Public Health, reached a similar conclusion. Led by Neeraj Sood, professor of public policy at USC Price School for Public Policy, the study concluded that between 2.8% and 5.6% of that county's adult population — or between 221,000 and 442,000 adults — have antibodies to the virus.
As the Los Angeles tests were kicking off on April 10, Sood, who collaborated with Stanford researchers on both studies, suggested earlier this month that a far lower fatality rate of COVID-19 would mean that some of the shelter-in-place orders currently in effect may not be necessary.
The Stanford and USC teams are part of a global push to obtain information about COVID-19 through antibody tests — a tool that experts say is critical for learning how the virus spreads and lifting the social-distancing orders that continue to ravage local economies. In Germany and Italy, national governments have made widespread testing a key strategy for easing out of the economic shutdown. Assessing the percentage of the region that has already been infected can help researchers determine how close the area is to herd immunity (a point at which about 60% of the population is immune), which regions have been hit hardest and which measures are going to be most successful in preventing the spread of infections.
"We're starting to get a better picture," said Jay Bhattacharya, professor of medicine at Stanford University and a co-author of both the Stanford and USC studies. "It's clear that it's spread more in some places than in others, and it's clear it's more deadly in some places than others."
As the shutdown stretches into its second month, seroprevalence studies are becoming increasingly common, both in the Bay Area and around the world. Just this week, a team from the University of California, San Francisco embarked on an effort to provide antibody tests to all residents of Bolinas, an unincorporated community in Marin County. It plans to follow suit on April 25 with a four-day test of about 5,700 residents in San Francisco's Mission District.
Much like its counterparts at Stanford and USC, the UCSF team is trying to address the lack of data about the spread of COVID-19. It is doing so by administering both a swab test for the virus' DNA, which shows whether an individual is currently infected, and a blood-sample test for antibodies, which indicate past infection.
"All our public health decisions, including when it will be possible to relax regional and statewide shelter-in-place orders, are driven by rough assumptions about how the virus behaves based on very limited data," Bryan Greenhouse, an associate professor of medicine at UCSF, said in a news release.
Studying in detail how the virus spreads in Bolinas and in the Mission will "give us crucial data points that we can extrapolate to better predict how to control the virus in similar communities nationwide," he added.
On the national level, the National Institutes of Health is trying to recruit 10,000 people for its own study of antibodies. Dr. Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, said the study will "give us a clearer picture of the true magnitude of the COVID-19 pandemic in the United States by telling us how many people in different communities have been infected without knowing it because they had a very mild, undocumented illness or did not access testing while they were sick.
"These crucial data will help us measure the impact of our public health efforts now and guide our COVID-19 response moving forward," Fauci said in a statement.
Dr. Sara Cody, Santa Clara County's health official, on Tuesday called the recent seroprevalence tests "exciting" and said that they confirm that "we have many, many, many cases that we didn't pick up."
The point was further underscored by Cody's announcement Tuesday that the county's medical examiner confirmed two deaths from COVID-19 that occurred on Feb. 6 and Feb. 17. The confirmation, based on tissue samples from the deceased, means that the virus was spreading in the community far before the health department became aware of that fact. Prior to the confirmation, county leaders had assumed that the first death associated with COVID-19 occurred on March 9.
But while the new research efforts promise to shed light on the prevalence of COVID-19, public health officials are indicating that they will need far more information before they can think about lifting the restrictions. Cody suggested during her Tuesday update to the Board of Supervisors this week that social-distancing rules are unlikely to be relaxed any time soon.
"We know the level of immunity in our county is in single digits," Cody told the Board of Supervisors on Tuesday. "We are far, far, far from herd immunity and not likely to get there until we have a vaccine."
She also pointed to uncertainty about the large number of antibody tests out on the market, all of which have different characteristics. She cautioned that the detection of antibodies does not necessarily mean that the person is immune to the coronavirus.
"In order to know whether the antibody test indicates that you're protected, there are many more studies that need to be done because basically you have to make sure that the antibody is a marker that you have enough protections such that if you got infected with the virus again you would not get sick," Cody said at the meeting. "Right now, it's not clear whether the antibody tests that are out there provide that."
'Perfect' for task, or a 'screwup'?
While results from both the Stanford and USC studies are preliminary and are now undergoing peer review, the reaction from the scientific community has been swift and, in a few cases, brutal.
Bhattacharya told this news organization Thursday that the team has received more than 1,000 constructive comments on the study, which continues to get refined.
But Andrew Gelman, a professor of statistics and political science at Columbia University, suggested in an online forum that the Stanford authors "owe us all an apology." After detailing his concerns about test accuracy and selection bias, Gelman concluded that the study's findings are a "product of a statistical error."
"I don't think (the) authors need to apologize just because they screwed up. I think they need to apologize because these were avoidable screwups. They're the kind of screwups that happen if you want to leap out with an exciting finding and you don't look too carefully at what you might have done wrong," Gelman wrote.
Other researchers also weighed in and criticized the study for the way it selected its participants (through Facebook ads) and for using tests that were not FDA approved. Dr. Eric Topol, a professor of molecular medicine at Scripps Research, suggested that the two studies have given "a false sense that this is not a bad virus at all."
"It's bad math, bad tests and bad outcomes for the confusion that it engenders," Topol told the Associated Press after the preprint was released.
Bendavid said the research team analyzed the test kits from Premier Biotech before the experiment to determine the percentage of positive cases that the tests miss. By testing blood samples from Stanford Hospital patients that were shown to be positive through a DNA test, as well as samples that were known to be negative because they were taken before the pandemic, the researchers concluded that the test sensitivity is about 91.8%.
Bendavid told this news organization that this was factored into the group's calculation.
"They may be somewhat less precise than laboratory-based tests, but for the purposes of what we have done — for trying to get a sense of prevalence of antibodies to coronavirus in our community — they are perfect," Bendavid said last week.
Part of the skepticism from the broader community stems from the positions that the researchers have taken in the past. In late March, Bendavid and Bhattacharya coauthored an opinion piece in the Wall Street Journal positing that the virus is far less deadly than many experts suggest.
"If it is true that the novel coronavirus would kill millions without shelter-in-place orders and quarantines, then the extraordinary measures being carried out in cities and states around the country are surely justified," Bendavid and Bhattacharya wrote. "But there's little evidence to confirm that premise — and projections of the death toll could plausibly be orders of magnitude too high."
Another member of the Stanford team, John Ioannidis, wrote an article in March suggesting that the true fatality rate of COVID-19 could be 0.05%, lower than the seasonal influenza. If that is the real rate, "Locking down the world with potentially tremendous social and financial consequences may be totally irrational.
"It's like an elephant being attacked by a house cat. Frustrated and trying to avoid the cat, the elephant accidentally jumps off a cliff and dies," Ioannidis wrote in Stat, a health news website.
Sood also said earlier this month that if tests show that the death rate of COVID-19 is indeed nine or 10 times that of the flu, people should indeed stay at home and practice "very strong physical distancing," according to a news release from USC.
But if the true mortality rate is far lower than the flu, he wrote, "then we don't need to be doing that."
(Infectious-disease expert Fauci said in March that the seasonal flu has a mortality rate of 0.1%. However, it should be noted that more than 100 million Americans get the flu vaccine annually, according to the CDC, and no such vaccine exists for the coronavirus.)
But researchers also emphasized that the mortality rate of the virus is not the only factor that should be considered by public health officials as they consider their next steps in managing the pandemic. The findings in Los Angeles showing that about 4% of the population is infected mean that "we are very early in the epidemic and many more people in Los Angeles County could potentially be infected."
"And as the number of infections arise, so will the number of deaths, the number of hospitalizations and the number of ICU admissions," Sood said in a news conference Monday.
Bhattacharya shared that view. The biology of the virus, he told this news organization, is not the only factor that determines the fatality rate. Another key factor is the setting in which patients that get viral pneumonia are treated.
"It's a very important input into the decision, but it's not the only thing," Bhattachary said of the death rate. "How likely is it, if we lift the caps, that we will overwhelm the hospitals? That's a really important question that needs to be addressed."
Given the outstanding questions, county and state leaders are continuing to take a cautious approach on social-distancing orders. Santa Clara County Executive Jeff Smith pointed to the recent discovery of February deaths from COVID-19 as another reason to proceed slowly.
"It's all consistent with what we know," Smith said. "There's a larger population in the community that is infected than we are currently identifying with testing, and a significant portion of that population is asymptomatic as far as COVID — another reason to recognize the great risk we have by relaxing the shelter-in-place too early."
And while some governors are already restarting their states' economies, Gov. Gavin Newsom said California's stay-at-home order would only be modified based on numerous indicators. These include expanded testing and contact tracing, the ability of hospitals to handle surges and the development of treatments.
In his Wednesday address, Newsom highlighted the Stanford and the USC studies, noting that they have "generated a little bit of controversy about what they are and what they're not."
But he also said that the state is preparing to greatly expand its use of serological tests to manage the COVID-19 crisis, noting that the state has agreed to buy 1.5 million such tests from Abbott Laboratory.
"While it (serological testing) can be extraordinarily helpful in terms of guiding our framework and indicators for reopening, it is not at this moment a panacea, but it could be profoundly significant," Newsom said. "And we will certainly do our part, in addition to procuring PCR (diagnostic DNA) tests, we will do the same to provide serology tests to people all over the state," Newsom said.
Find comprehensive coverage on the Midpeninsula's response to the new coronavirus by Palo Alto Online, the Mountain View Voice and the Almanac here.