How Scientists Create Models For Disease, And What The Latest Ones Say

Apr 8, 2020
Originally published on April 8, 2020 8:47 pm
Copyright 2020 NPR. To see more, visit https://www.npr.org.

ARI SHAPIRO, HOST:

Coronavirus infections and deaths are still rising fast in many parts of the country, leaving many people to wonder, how bad will this get? And then after the peak, how long will it take to get back to some version of normalcy, and what would that normalcy even look like?

Joining us to talk through all of this is NPR's global health correspondent Nurith Aizenman.

Hi, Nurith.

NURITH AIZENMAN, BYLINE: Hi, Ari.

SHAPIRO: So you've been talking to public health experts who are trying to predict where the outbreak is headed. How do they come up with these projections?

AIZENMAN: The first thing to say is these are computer simulations that are only as good as the data that's plugged into them. For instance, a lot of them consider how many infections or deaths there've been in a particular state, how that's changed over time. But we might not have enough tests to accurately say what the case count or even the death count has been.

Also, each model's based on all sorts of assumptions about how this spreads, how much of an impact social distancing has. And scientists are still gathering information about that and updating their formulas every day. So the modelers would be the first ones to warn us these simulations are not forecasts. Their purpose is just to help guide our strategy.

SHAPIRO: OK. So with all of those caveats in mind, what are some of the models showing right now?

AIZENMAN: The Trump administration has released a bottom-line estimate. You've probably heard that best-case scenario - 100,000 to 240,000 people die. But officials on the coronavirus task force haven't revealed how they came up with that calculation and the timing.

So I'll tell you about one outside model that the task force has at least frequently cited. It's been put together by Chris Murray at the University of Washington's Institute for Health Metrics and Evaluation. And he and his team project that if statewide social distancing rules are put in place in all states that don't currently have them and kept in place through June 1, nationwide, the U.S. will see the highest daily death toll on April 12. Then deaths will start to decline each day, such that by mid-June, this wave of infections will be quashed, and somewhere between 31,000 to 127,000 people will have died.

This model also finds, I should note, that a number of states, including New York, Louisiana, Montana, Colorado, Washington, D.C. - that they have already now hit their peak number of daily deaths. But I should also add that another important model has a different finding.

SHAPIRO: All right. Well, what does that other important model show?

AIZENMAN: Right. So this other model, which the administration has also consulted, was put together by a team at Columbia University that includes Jeffrey Shaman. And he tells me that their projections only look at what happens six weeks out, but they find that a lot of those states I just mentioned won't hit their peak number of daily deaths until at least mid-May, maybe longer. And even if those states take much more drastic steps to social distance, it'll be more like late this month before daily deaths peak there.

SHAPIRO: OK. So it sounds like it's really tough to pin down when this wave of infections will pass. But whenever it does, society needs to somehow reopen. How are these people thinking about what needs to happen then?

AIZENMAN: Yeah. So the challenge to that is that the rough estimates are that even if we succeed in quashing this current wave of infections, the vast majority of people in the United States will not have been infected, which means that until there's a vaccine, we're still vulnerable. So now there's all sorts of ideas being floated and studied about how you could open up the economy a bit while still tamping down how much people interact face to face. I mean, these are ideas like maybe people who are older than 45 or 50 are still asked to stay home, and everyone else works; or maybe everyone whose last name begins with A through M works on even days; everyone whose name begins with N through Z works on odd days; or we all work for four days, and then we go back to massive social distancing for 10 days.

And also, we're going to need a really robust system for quickly detecting any flare-ups, testing people with symptoms, identifying their contacts, maybe even coming up with housing to put those contacts into during quarantine so they don't infect...

SHAPIRO: Yeah.

AIZENMAN: ...Their families. And there's no sign the U.S. is anywhere near setting all that up.

SHAPIRO: I guess it's some consolation that the whole world is having to answer this - these questions, so at least we can learn from each other.

AIZENMAN: Absolutely.

SHAPIRO: That's NPR's Nurith Aizenman.

Thank you very much, Nurith.

AIZENMAN: Good to do it, Ari.

(SOUNDBITE OF MUSIC) Transcript provided by NPR, Copyright NPR.