Thread by @ASlavitt

COVID Update May 4: We’ve been lied to about the potential death toll in this country & it matters. 1/

The warping of the conversation about death tolls is problematic. People who warn of dire consequences get labeled fear mongers or worse. Some are accused of wanting to be right at all costs.

All this wrangling cheapens our losses. 2/

It wasn’t a month ago that the IHME, touted by Trump, released a 60,000 estimated death forecast. 3/

I participated in a conversation with the lead modeler and reported what I heard. At the time, even with the hope that Americans would #StayHome, the reasonable estimates of most models were 100,000-250,000, in some cases higher. In some cases much higher. 4/

IHME’s 60,000 declaration was a bomb let off by the White House. It sent a clear message— this virus is overblown hype. We got this. 60,000 is a number Americans can “tolerate.” It sounds vaguely like a bad flu. 5/

The president was quick to come out & claim success for this new number, even though his team had told him 100k-250k. We could have had over a million deaths, but now it will be 50-60k. 6/

For many governors & Trump supporters, it was all they needed to hear. 7/

In my estimation something the president gets wrong is the public’s tolerance for bad news. Bad news served up straight. Hope and plans are good but only when they are rooted in reality. 60k was another version of “gone by April” or “15 going to zero.” 8/

One reason I avoid criticizing these models is the people who put them together are doing something unknowable, have different purposes, will be wrong be definition & exponential growth means wide confidence intervals depending on assumptions. 9/

As soon as Trump unleashed the 60,000 number my Twitter mentions filled up “I told you sos” from “you were wrong” when I shared that scientists in March were telling me they could foresee a death toll over 1 million if we didn’t dramatically reduce the infection rate. 10/

Making sure that didn’t happen is the primary reason I, along with 15 other bipartisan health care & scientific leaders launched the #StayHome #StayHomeSaveLives initiative. Because even with bad news, if you have the truth, you can act. 11/

(The other reason, of course, was to to steal people’s freedom.) 😏
12/

So back in April I felt if we’re going to be overrun by a “new” model, I wanted to know what went into this model so I listened & put what I heard in an April 9 Twitter thread summarized in the attached. I will repeat what I heard here. 13/

medium.com/@ASlavitt/maki…

60,000 model assumes:
1-states without social distancing don’t have any outbreaks
2-no inter-state/international travel
3-The infection rate R0 goes below 1
4-There is no let up of social distancing
5-No state opens until they have <10 cases
6- It ends by August 14/

At the time the model was issued, as I probably too politely pointed out in the medium post, it was already wrong.

Wrong in New York’s results. Wrong in assuming no cases it couldn’t see. Wrong in not accounting for a second wave. Wrong in assuming there would be testing. 15/

States like South Dakota, Nebraska & Iowa started having hotspots, travel continued (& continues) to many of the world’s hot spots, the infection rate outside of places like NY and WA are growing, social distancing let up w summer in many states & on & on 16/

Is this just hindsight? Not to anyone with eyes. These were bad assumptions & they were too rigorously defended. And by the president, not rigorously examined.

Forget the modeler. Blame the White House for announcing a number that they couldn’t even escape April with. 17/

Today IHME is roughly doubling their death toll estimate to 135,000 by August. 18/

In a sign that the modelers either have a great sense of humor or don’t understand cause & effect, they blamed the increase on people being lax about social distancing standards.

I wonder if their model can diagnose what model White House used to downplay the virus.19/

Why do I say all these bogus numbers matter? Because releasing a model with these clearly unrealistic & favorable assumptions misled people. And with an infectious disease, one misled person can be enough. 20/

Most of us can’t swallow how fast infections spread with exponential math. How social distancing— until we have reliable testing & tracing & good masks for all— is the only way to save lives. That’s why modelers should pay careful attention that their black boxes be misused. 21/

Trump at a Fox News town hall this weekend said the number of deaths could now be 75,000 to 100,000.

As we approach 70,000 deaths, with 1700 dying every day, he’s not leveling with us, so much as he’s helping his supporters become the last to know. 22/

In fact, there was a skirmish today as the NYT published that a White House report said the death toll will reach 3000/day by the end of May. The WH disowned it & pushed back. Not a forecast, just a model. Can’t disagree with the president. 23/

Another model from FEMA which forecasts 200,000 new cases/day by the end of month was uncovered. If that’s even a remote possibility, it’s not responsible for the president not to be warning the country. 24/

The president needs to consider for this instance erring on the side of caution, risk being too conservative, make Americans over-prepared. He has too often roundly refuted experts the last three years. His base will listen only to him. 25/

Twitter being Twitter, I can expect some people to sharpen their teeth & argue points of the model. That’s fine. I expect that. But it’s so beyond the point which is to help guide us to safe smart actions. 26/

COVID-19 seems to be a uniquely tough American foe. The rest of the world is figuring it out. The Czech Republic did it with masks. China with isolation. Germany with testing. Hong Kong with experience. New Zealand with alerts. Greece with discipline. 27/

The cost of these lessons is already too high. But it is not beyond our power to change it. But I’m afraid to change this, we do have to first face it. /end

PS I already regret the bad joke told in the middle of this thread. Mostly because it wasn’t as funny as it was when I thought it.