CDC: First virus deaths in U.S. may have been earlier
Reports of January 2020 deaths may be real or data glitches
The hunt for America’s original COVID-19 victim is still very much on.
And in recent weeks, it appeared it might be somebody new.
For almost a year now, the Feb. 6, 2020, death of San Jose’s Patricia Dowd has been considered the country’s first coronavirus fatality. More than two months after her mysterious death, tests revealed the otherwise healthy 57-year-old semiconductor company auditor had been infected with the coronavirus when she died.
As he introduced plans this January for a COVID-19 stimulus plan, President-elect Joe Biden mentioned the “beloved wife, mother, daughter, sister” who “never knew she had the virus at the time when most folks never heard about the virus.”
But in recent weeks, the Bay Area News Group found data, tucked into an avalanche of death records on both the California Department of Public Health and U.S. Centers for Disease Control
“It’s highly likely that there were quite a few cases around ... before we recognized that it was here in the United States. Could there have been an earlier death? Of course there could have, and I believe there probably was.”
— Matthew Memoli, director of the clinical studies unit at the Laboratory of Infectious Diseases at the National Institutes of Health in Bethesda, Maryland
and Prevention websites, that appeared to indicate something new: COVID-19 listed as the cause of multiple deaths in the first days of January 2020.
Was this new information, quietly released last month with no fanfare, the beginning of a new chapter in our understanding of SARS-CoV-2 — or was it a glitch in data collection that could somehow be undermining our quest for answers? After a few days of investigation, the answer appears likely to be a data glitch, but health officials acknowledge the first COVID-19 death may have been earlier than we know.
“It would be a very important story if those deaths did occur in the U.S. in January 2020,” said John Swartzberg, professor emeritus of infectious disease and vaccinology with the UC Berkeley-UCSF Joint Medical Program. “That would dramatically change the way we understand the beginning of this pandemic.”
The Bay Area News Group first discovered the possible new deaths late last week while analyzing a recently released dataset, the details of which caught California public health officials by surprise. While California has since removed the January 2020 deaths from its public website, blaming data errors, the CDC is still investigating at least five deaths nationwide that appear in its records in the first weeks of2020.
An earlier timeline would further throw into question how the U.S. and the rest of the world responded to the biggest public health disaster in a century. While definitively discovering Patient Zero in the U.S. may be elusive, it’s almost a certainty, scientists say, that the virus started its deadly journey well before the world’s current records indicate.
In late December 2019, China first announced dozens of cases of a mysterious viral pneumonia, and Chinese media reported the first death from the illness occurred on Jan. 9, 2020. By mid-January, the United States recorded its first case when a traveler tested positive after returning from Wuhan, China.
“It’s highly likely that there were quite a few cases around, at least in certain places, before we recognized that it was here in the United States,” said Matthew Memoli, director of the clinical studies unit at the Laboratory of Infectious Diseases at the National Institutes of Health in Bethesda, Maryland. “Could there have been an earlier death? Of course there could have, and I believe there probably was.”
His team ran a nationwide study of more than 10,000 people last year and found some with high levels of antibodies to the coronavirus in March 2020, meaning they would have been infected in February.
More recent studies have found evidence the virus was circulating here even earlier. The CDC tested more than 7,000 blood samples from nine states, including California, that were collected by the American Red Cross between Dec. 13, 2019, and Jan. 17, 2020, for coronavirus antibodies. The study found samples with signs of COVID-19 in all nine states although there is no indication whether any of those people got sick or died.
Earlier this year California’s state health department started reporting COVID-19 fatalities by the actual date of death, providing researchers for the first time with a more precise timeline of the virus’s deadly path in California.
The data from the California Department of Public Health showed that in 2020 from Jan. 1 to Feb. 5 — the day before Dowd’s death — 19 people appeared to have died of the virus in California, including in Kern, Imperial, Riverside, San Bernardino, Ventura, Los Angeles, San Bernardino and San Francisco counties. The records also indicated for the first time hundreds more cases of the virus in the first two months of 2020, long before it was known to be widespread.
But in response to questions about the critical new information, a spokesperson wrote in an email that “CDPH is aware of some data entry errors that resulted in incorrect dates of death being posted.” It replaced the data on Tuesday with records that only go back to March 1, 2020.
The state agency wasn’t the only source showing COVID-19 deaths in January 2020. As of early April, the federal agency that tracks all death certificates in the U.S. also was showing deaths, though fewer, in January 2020.
Provisional death data compiled by the CDC’s National Center on Health Statistics, which has compiled this type of record for over 100 years, show five deaths attributed to COVID-19 in January 2020. The site shows at least one death each in California, Illinois, Wisconsin and Tennessee.
“It’s either clearly an error or something we didn’t know about before,” Robert Anderson, chief of the Mortality Statistics Branch at NCHS, told the Bay Area News Group. The agency is working to either confirm or correct those records before its provisional death data is finalized later this year, he said.
Anderson said while the early 2020 deaths “stick out like a sore thumb,” it is most likely due to data entry errors, someone typing 2020 instead of 2021, or January instead of July in the digital vital statistics systems in which each state shares data.
Months after the virus became widespread, pathologists in many parts of the country reexamined unexplained deaths from before the pandemic by testing lab samples for signs of COVID-19, the same way Dowd’s case was confirmed in Santa Clara County.
After the shocking results from Dowd’s case, Santa Clara County Medical Examiner Michelle Jorden said her office sent a couple of dozen tissue samples to the CDC from people who died as early as late 2019. The testing turned up one positive case, but that person died March 9, so Dowd “remains our very first case in Santa Clara County,” Jorden said.
Stanford University’s Clinical Virology Lab tested samples from thousands of patients with respiratory symptoms seen at Stanford, said Benjamin Pinsky, the lab’s medical director. But samples from the last two months of 2019 revealed no evidence of COVID-19, he said. Testing of samples from the first two months of 2020 revealed just two positive samples collected on Feb. 21 and 23, 2020.
While Pinsky agreed that “it’s certainly quite possible that there were other introductions and other sites of community transmission that obviously aren’t captured by our small study” that happened earlier, he doesn’t think the virus was spreading around the community before February.
But Warner Greene, a senior investigator at the Gladstone Institutes, a San Francisco-based biomedical research institution, thinks it’s entirely possible the virus was already spreading here back then. New evidence that relies on what’s known as a molecular clock, which lets scientists track mutation rates and make estimates of when the virus entered humans, suggests the disease was spreading in China by November, Greene said, and in a global economy with lots of air travel, it could spread quickly.
Meanwhile, data errors are adding complications to an already murky story.
George Lemp, an epidemiologist and former director of the HIV/AIDS Research Program at the University of California Office of the President, encountered similar data problems such as transposed dates that hindered the understanding of the spread and toll of HIV. With a disease like the coronavirus, he said, mistakes are inevitable with myriad statistics being tracked by thousands of local, state and federal agencies.
Memoli, with the NIH, said that studying the virus helps clarify how the pandemic spread and provides clues about the future. But exactly where and how the virus struck its first fatal blows in the U.S. may remain forever a mystery.
“To some degree,” Memoli said, “the date of the first death is never going to be known.”