On Thursday, a Florida health official told a local news station that a young man who was listed as a COVID-19 victim had no underlying conditions.
The answer surprised reporters,
who probed for additional information.
“He died in a motorcycle accident,” Dr. Raul Pino clarified. “You could actually argue that it could have been the COVID-19 that caused him to crash. I don’t know the conclusion of that one.”
The anecdote is a ridiculous example of a real controversy that has inspired some colorful memes: what should define a COVID-19 death?
While the question is important, such incidents may be just the tip of the proverbial iceberg regarding the unreliability of COVID-19 data.
In May, a public radio station in Miami broke what soon became a national story. The US Centers for Disease Control and Prevention (CDC) had been conflating antibody and viral testing, obscuring key metrics lawmakers use to determine if they should reopen their respective economies.
The story was soon picked up by NPR, who spoke to an epidemiologist who condemned the practice.
“Reporting both serology and viral tests under the same category is not appropriate, as these two types of tests are very different and tell us different things,” Dr. Jennifer Nuzzo of the Johns Hopkins Center for Health Security told NPR.
The Atlantic soon followed with an article that explained the agency was painting an inaccurate picture of the state of the pandemic. The practice, the writers said, was making it difficult to tell if more people were actually sick or had merely acquired antibodies from fighting off the virus.
Public health experts were not impressed.
“How could the CDC make that mistake? This is a mess,” said Ashish Jha, the K. T. Li Professor of Global Health at Harvard and director of the Harvard Global Health Institute.
In some ways the “mess” was no surprise. Two weeks earlier, Dr. Deborah Leah Birx, the White House’s coronavirus task force response coordinator, reportedly ripped the agency in a meeting, saying “there is nothing from the CDC that I can trust.”
Birx’s concerns about the CDC’s data did not alleviate concerns of data manipulation. The New York Times speculated that perhaps the agency had sought to “bolster the testing numbers for political purposes.” TheTexas Observer wondered if the state was “inflating its COVID testing numbers by including antibody tests.”
Considering President Trump’s sometimes comically inaccurate boasts about America’s testing prowess, perhaps such questions were not unjustified. The many people who spoke to the Times said the answer was simpler, attributing the flawed system to “confusion and fatigue in overworked state and local health departments.”
If data manipulation had been the motive, the architects of the ploy were in for a rude awakening. Testing numbers did soar, but so did case numbers; the surge in late June and throughout July spawned new fears of a second wave and more lockdowns and more charges that America was botching the pandemic. (The surge was the result of both increased testing, including antibody testing, as well as a resurgence of the virus.)
Coronavirus Epidemic in America? Can We Trust the Data? How Much Does it Cost to Get a COVID-19 Test?
Tensions between the White House and its own agency boiled over last week when the Trump Administration stripped the CDC of its role in collecting data on COVID-19 hospitalizations.
A Data Fiasco of Historical Proportions?
It’s hard to read the drama, incompetence, and confusion without thinking about Dr. John Ioannidis, the C.F. Rehnborg Chair in Disease Prevention at Stanford University.
In a March 17 STAT article, Ioannidis warned the world was looking at what could turn out to be a “once-in-a-century evidence fiasco.” He worried central planners were making sweeping and reflexive changes without sufficient data.
Locking people up without knowing the fatality risk of COVID-19 could have severe social and financial consequences that could be totally irrational, Ioannidis warned.
“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,” said Ioannidis, one of the most-cited scientists in the world.
In one sense, Ioannidis has already been proven right. The models on which lockdowns were initiated have already proven astronomically wrong. But that was hardly the only example.
Every day it seems there’s another story about reporting flaws or mixups.
Tuesday it was a lab in Connecticut where researchers said they discovered a flaw in a testing system for the virus. The flaw resulted in 90 people receiving false positives. That may not sound like many, but researchers said the test is used by labs across America.
A few days earlier, it was announced that Texas had removed 3,484 cases from its positive Covid-19 case count because the San Antonio Health Department was reporting “probable” cases. None of the people had actually tested positive for COVID-19.
We don’t know how many new cases are probable cases and not positive cases, but we know it’s a lot. That’s because in April, the CDC changed its reporting to include people who had not tested positive for the virus but might have it. (The CDC’s criteria for what qualifies as a probable case are more than a little confusing.)
As the Associated Press noted, the change was made with the understanding that “deaths could soon jump because federal health officials will now count illnesses that are not confirmed by lab testing.”
The One Thing Everyone Agrees On
COVID-19 has been far from the deadliest virus in modern history, but it has been the most divisive. The public, politicians, policy experts, and public health officials have disagreed on how deadly it is and how best to contain it.
But the one thing everyone seems to agree on is the numbers we have—fatalities and cases—are way wrong. A new CDC report estimates COVID-19 rates about 10 times higher than reported. Ioannidis put the figure even higher, estimating weeks ago that as many 300 million people had already been infected globally.
Deaths are more complicated.
The New York Times says COVID-19 deaths have been massively undercounted. Dr. Ashish Jha, speaking to Lawrence O’Donnell on MSNBC, agreed, saying most experts agreed there is a “substantial undercount.”
Others, including nearly one-third of Americans according to a recent survey, believe that the COVID-19 death toll is inflated. This includes physicians who say medical professionals are being pressured by hospital administrators to add coronavirus to death sheets.
Writing at the American Mind, Angelo Codevilla recently argued if the CDC had used the same criterion for the SARS virus as COVID-19—primarily “severe acute respiratory distress syndrome”—total COVID fatalities in the US would have been 16,000 through June.
Nobody knows the true count, of course. But the one thing left and right seem to agree on is the data we have are junk. And yet the lesson we keep hearing is “trust the experts.”
“Follow the science. Listen to the experts. Do what they tell you,” Joe Biden said in April.
But thinkers as diverse as Matthew Yglesias at Vox to author Matt Ridley have pointed out the dangers of blindly following “the experts,” especially when they’ve shown themselves to be spectacularly wrong from the very beginning on the COVID-19 pandemic.
“It’s dangerous to rely too much on models (which lead politicians to) lock down society and destroy people’s livelihood,” Ridley recently told John Stossel. “Danger lies both ways.”
Ridley has a point. The experts can’t agree on their own numbers or even clearly answer if a man who died in a motorcycle accident while infected should be labeled as a COVID-19 death.
In light of this, perhaps it’s time for the experts to exercise some humility and begin offering guidance to individuals instead of advocating collective blunt force.
What Should Define a COVID-19 Death? The Unreliability of COVID-19 Data. Dr. Ioannidis
Fair Use Notice: Images/logos/graphics on this page contains some copyrighted material whose use has not been authorized by the copyright owners. We believe that this not-for-profit, educational, and/or criticism or commentary use on the Web constitutes a fair use of the copyrighted material (as provided for in section 107 of the US Copyright Law).