Saturday, March 28, 2020

Coronavirus modelers factor in new public health risk: Accusations their work is a hoax

The Washington Post







In the one month since the first U.S. coronavirus death, America has become a country of uncertainty.

New cases of infection and casualties continue multiplying. New York and Louisiana hospitals are grappling with a flood of patients that threatens to overwhelm their health-care systems. Meanwhile, the president and political conservatives are increasingly agitating to end drastic restrictions meant to buy time and save lives.

Running beneath it all, in a continuous loop through our national psyche, are basic questions leaders are struggling to answer: When can we safely lift these quarantines? How many people could die if we do it too early? Just how dangerous will this pandemic turn out to be? And what exactly should be our next step?

This is why epidemiology exists. Its practitioners use math and scientific principles to understand disease, project its consequences, and figure out ways to survive and overcome it. Their models are not meant to be crystal balls predicting exact numbers or dates. They forecast how diseases will spread under different conditions. And their models allow policymakers to foresee challenges, understand trend lines and make the best decisions for the public good.

But one factor many modelers failed to predict was how politicized their work would become in the era of President Trump, and how that in turn could affect their models

In recent days, a growing contingent of Trump supporters have pushed the narrative that health experts are part of a deep-state plot to hurt Trump’s reelection efforts by damaging the economy and keeping the United States shut down as long as possible. Trump himself pushed this idea in the early days of the outbreak, calling warnings on coronavirus a kind of “hoax” meant to undermine him.

The notion is deeply troubling, leading health experts say, because what the country does next and how many people die depend largely on what evidence U.S. leaders and the public use to inform their decisions. Epidemiologists worry their research — intended to avert massive deaths in situations exactly like this pandemic — will be dismissed by federal leaders when it is needed most.


The peak


St. Louis endured two catastrophic episodes during the 1918 flu pandemic. (Library of Congress)




So much of the coming months and our country’s timeline depends on the peak of the coronavirus’s spread — its steepness, length and timing.

But here’s the thing about peaks: You often can’t tell where they are until you’re already past them, on your way down the other side.

And in a country as big as the United States, the peak will be not so much a single curve as it will the sum of many curves — as the outbreak spreads to different cities and regions at different times.

Nonetheless, a new model released Thursday by the University of Washington’s School of Medicine is one of the first to forecast a national peak. It projects that the peak in daily U.S. deaths will arrive in mid-April, and the tail end of that curve, subsiding below 10 daily deaths, will arrive by the first week of June.

But that projection comes with huge caveats because of estimations and assumptions that have to be built into the calculation, given how much is still unknown about the disease covid-19.

The model — created by the university’s Institute for Health Metrics and Evaluation — assumes, for example, that all remaining states that have not enacted strict restrictions on residents will do so in the next week once they see how grave the situation is in areas like New York.

But Florida Gov. Ron DeSantis (R) has refused to issue orders for people to stay at home. Alabama’s governor has similarly resisted. And this week, Mississippi’s governor issued an order defining almost all businesses as “essential” — including auto repair, bars and restaurants.

The Washington model assumes the entire country will maintain these strict restrictions until summer. But Trump has increasingly made clear he wants to reopen parts of the country by Easter on April 12. And on Thursday, Trump unveiled a plan to identify specific counties that he thinks should reopen soon.

The University of Washington model predicts that this first wave of infections will end by summer (with subsequent waves a possibility) and that the death toll during this initial period will range from 38,000 to 162,000 — a lower projection than some earlier models. But the actual death count in coming months will largely depend on how badly hospitals are overwhelmed and whether they receive supplies like ventilators that they desperately need.

On Thursday night, Trump cast doubt on experts’ projections on those as well. “I have a feeling that a lot of the numbers that are being said in some areas are just bigger than they’re going to be,” Trump told Fox News host Sean Hannity in a phone interview. “I don’t believe you need 40,000 or 30,000 ventilators. You know, you go into major hospitals, sometimes they’ll have two ventilators, and now all of a sudden they’re saying, ‘Can we order 30,000 ventilators?’”


On the attack

 
Deborah Birx, White House coronavirus response coordinator, speaks at a Thursday news briefing

The attacks from Trump supporters against epidemiologists ratcheted up several bars on Thursday as pundits on the political right took aim at one of the world’s leading epidemiologists, Neil Ferguson of Imperial College in Britain.

Ferguson had co-written a paper this month estimating 510,000 deaths in England and 2.2 million in the United States — if those countries did not take drastic actions. The paper’s conclusions were so chilling that they launched leaders in both countries into action. The next day, Trump abruptly stopped encouraging Americans to go on with their lives and began urging them instead to work from home and not meet in groups of more than 10.

Since then, Trump has flipped back to wanting workers back at their jobs — framing it as a decision between saving the U.S. economy or a handful of lives. And his supporters have followed, attacking Ferguson online.

After Ferguson gave new testimony to British officials Wednesday, they hailed it as evidence that Ferguson and other experts were overselling the coronavirus threat. Fox News host Laura Ingraham wrongly stated that in his testimony Ferguson’s projection had been “corrected.” The chyron on her show Thursday night stated, “Faulty models may be skewing COVID-19 data.”

“Today — this is what our instinct was — the lead researcher did an about-face on those terrifying projections, the very projections that drove so much of our response,” Ingraham claimed on her show.

A Wall Street Journal columnist wrote that the revision “raises serious questions about the radical countermeasures inspired by public-health experts like Mr. Ferguson.” Even one of Trump’s coronavirus task force coordinator, Deborah Birx, seemed to lean into the questioning of Ferguson. “I’m sure many of you saw the recent report out of the U.K. about them adjusting,” Birx said. “If you remember, that was the report that said there would be 500,000 deaths in the U.K. and 2.2 million deaths in the United States.

They’ve adjusted that number in the U.K. to 20,000. So half a million to 20,000. We’re looking into this in great detail to understand that adjustment.”

But in fact, Ferguson had not revised his projections in his testimony, which he made clear in interviews and Twitter. His earlier study had made clear the estimate of 500,000 deaths in Britain and 2.2 million in the United States projected what could happen if both took absolutely no action against the coronavirus. The new estimate of 20,000 deaths in Britain was a projected result now that Britain had implemented strict restrictions, which this week came to include a full lockdown.

But the argument over models in some ways is beside the point, said Natalie Dean, a biostatistician at the University of Florida. “The models are planning tools, but it doesn’t take a genius to look at what’s happening in Italy and realize that we’re on the same trajectory,” said Dean, who is working on coronavirus vaccine evaluation with the World Health Organization. “That should be enough to tell us we need to be doing more in reaction.”

A warning from pandemics past

One clear warning from epidemiology of past pandemics is the danger of lifting restrictions — as Trump wants to do in two weeks — too soon.

A seminal 2007 paper shows what happened in several U.S. cities when they eased restrictions too soon during the 1918 flu pandemic. Those cities believed they were on the other side of the peak, and, like the United States today, had residents agitating about the economy and for relaxing restrictions.

Once they lifted them, however, the trajectory of those cities soon turned into a double-humped curve with two peaks instead of one.


This chart shows the double-humped peaks of deaths during the 1918 flu in St. Louis. The city imposed strict restrictions early on but loosened them under pressure from its citizens, only to see deaths jump again. (Courtesy of JAMA and Howard Markel)



Two peaks means you get the overwhelmed hospitals, the death and destruction, without that flattening benefit people were trying so hard to achieve with arduous restrictions.

“Knowing when to release the throttle is hard. There’s is no button that says push me now,” said Howard Markel, a historian and physician at the University of Michigan who co-wrote the 2007 paper along with a top CDC official, Martin Cetron. “But the trick is to be patient, not to jump the gun. Otherwise, all that happens is you get more cases, more deaths and everything you worked so hard for with those restrictions just goes to waste.”

One of the perpetual frustrations of trying to prevent disease rather than curing it is that it’s often difficult for the public to appreciate the disasters you help them avoid.

“The problem is there’s no metric for prevention. How many cases you avoid. How many lives you save,” Markel said. “That’s why it’s so hard to stay the course but so important, too.”