On Monday the New York Times published what appeared to be an explosive finding: an internal document from the Trump Administration that forecast many more coming deaths from the coronavirus than the president has predicted publicly.
Specifically, the document included projections that in the coming weeks the rate of daily new infections in the United States will rise so precipitously that by June 1 more than 200,000 people per day will be contracting the virus — and more than 3,000 people per day will be dying. (For comparison, currently the U.S. is confirming about 27,000 new cases and 1,800 deaths each day — for a cumulative death toll of about 69,000. And on Sunday, President Trump said he expected total deaths to top out at about 100,000.)
But there’s an important caveat to this story. NPR contacted the epidemiologist who came up with the projections in the internal document, Justin Lessler of Johns Hopkins University’s Bloomberg School of Public Health. Lessler says the projections cited in the document do not represent his final forecast. Rather they were part of work that’s still very much in progress — in other words, incomplete.
“It’s as if somebody looked over my shoulder when I was halfway through putting the work together and took a picture and put the results out there,” he tells NPR.
NPR also reached out to various officials in the administration for comment.
Here are the answers we got to four key questions.
How did these projections come to be included in the internal administration document?
Lessler is on a team at Hopkins that has a contract to provide coronavirus projections to the Federal Emergency Management Agency (FEMA)’s National Response Coordination center (which, according to its website, is a “multi-agency center that coordinates overall federal support for major incidents and emergencies”). Lessler says he recently provided the draft projection graphics to colleagues at FEMA as part of his usual practice of sharing work in progress. The idea, he says, was to show them how the work was starting to take shape so they could give input.
“The understanding is that [this] was always intended to be shown to people who were fully aware that this was work in progress, not a final result,” says Lessler.
As to how and why the graphs were copied into the internal document, “I wasn’t privy to the process,” says Lessler. He says he also does not know whether the document — which is formatted like a series of slides for a presentation — was used to brief any officials in the Trump administration.
What is the administration’s response?
White House spokesman Judd Deere said that the document did not come from the White House, that it had not been presented to the coronavirus task force and that it has not gone through interagency vetting. Deere added that the projections do not reflect any modeling done by or analyzed by the task force as it worked on its recently released phased guidelines to reopen the economy.
NPR reached out to several federal agencies as well because many of the pages are stamped with logos of both the Department of Homeland Security (which oversees FEMA) and the Department of Health and Human Services or HHS (which oversees the Centers for Disease Control and Prevention or CDC). On some pages FEMA is referenced directly and on others the CDC logo is also included.
A spokesman for the CDC had no comment beyond issuing a statement that the projections were not from the CDC and directing questions to FEMA. Officials at FEMA did not respond.
Why does Lessler consider his projections incomplete? What was missing?
Lessler says that, as is usual practice among epidemiologists, he comes up with his projections by running a multitude of computer simulations for how the pandemic will play out. For each scenario he plugs in a different set of assumptions about the answers to a wide range of key variables such as how the virus is transmitted, how much the population adopts preventative measures such as social distancing, what impact those measures have and so on.
Once all the simulations are done, he then runs mathematical analyses to see which ones appear to be most predictive, and he comes up with a weighted average of all the results. But the projections that made it into the internal document published by the Times were based on only about one-third of the scenarios that Lessler will be including in the final projections.
To see an incomplete version of his work disseminated and discussed so publicly was all the more unnerving since it’s obvious from the graph that the simulations he’s run thus far are not that robust — since they fail to predict the actual number of deaths to date. “Yes, this has been my day: having Nate Silver criticize my intermediate work on Twitter,” he says, referring to the prominent blogger. “Which is,” he adds with a chuckle, “you know, a great feeling.”
So how many COVID-19 deaths would be predicted by a complete version of Lessler’s model?
Lessler says that while he has added in more simulations than the ones reflected in the internal document, he’s still finalizing his projection. And in any case, once that work is complete, Hopkins’ contract with FEMA prevents him from disclosing the results to the public. That said, Lessler cautions that the incomplete projection published in The New York Times of more than 200,000 new cases and more than 3,000 new deaths per day by June 1 is just one of many possible scenarios. And right now, “I do not know if it is likely.”