Fighting Coronavirus with Data

essential travel, same with the border to Canada. Most travel from the US to Europe has been suspended. Probably even more things have changed since I recorded this podcast on Monday evening. Does any of this federal response make sense? Does the state response make sense? Are we acting based on data, logic and reason? And what should we be doing? To get a really expert perspective on this, I spoke

to doctor Farzad Moustashari during the Obama administration. Farzad was the National Coordinator for Health and Information Technology at the Department of Health and Human Services. Before that, he worked for the Centers for Disease Control in the New York Office, focusing on New York City public health. Now he's the founder and CEO of Allidate, a healthcare technology company. Farzad is one of the clearest and most rational people that I know, and I knew he would have a lot

to say on this topic. Farzad, from very early in this crisis, and I mean very early, you were loudly saying on Twitter and I was following you closely, that we didn't have a coherent national response strategy, even at the conceptual level of knowing what we were trying to do. Have your worries on that front been at all alleviated or do you still think we have a lack of coherence? Oh my god, I wish No. We don't have a plan, and there's no clear criteria on when what are we

trying to do in each community at what stage? And when are we in containment and we're trying to do contact tracing and stamp out the sparks as they're coming in and keep it out, And when are we trying to do social distancing and mitigation, And when are we going full bore on suppression and doing these extreme measures, and when are we going to get out right? So we have a plan that says a fifteen day plan

to slow the growth. What happens on day sixteen. Noah, I was hoping you were going to tell me that. I mean, you know you're asking the wrong guy. One of us is actually a public health specialist. Well, I'm telling you, on day sixteen, you know what it's gonna look like. It's going to look bad. If we had done a miraculous job of slowing down transmission, we still would be seeing mounting case of hospitalizations, ICU cases and deaths on day fifteen, purely based on the people who

are already infected. So I think we have to have a plan that says these are the measures, these are the data points that we look to to decide when we take these extreme economy crippling measures, and this is when we get out of them. And we don't have that right now. But not only do we not have the plan, we don't have the plan to get to the plan to get the data that we would need to be able to make that plan effective. Right because we're doing testing and the promise here has been testing

more tests. Everyone who wants a test can get it. And let me tell you, my friends who are in the public health world are tearing their hair out saying we don't just need more tests. We need to get actual insights from those tests. We need, for example, to know among all the people, how many tests are positive, what percent of tests are positive. But right now we're only getting the positives. Secretary Esar said, I don't know how many tests are done. We don't know how many

tests are being done. So when we see an increase in the number of positives, is that the infections getting worse or that we're testing more people. So you've described a very very grammar situation. Now you have seen public health issues from a broad range of perspectives. You've seen it as a New York City public health official, so that was the local. You've seen it from the federal level at HHS. Now you're seeing it from the private sector.

You're almost uniquely qualified, it seems to me to say what we could realistically do now, So let's be as concrete as we positive can and productive. What would be your top, say, three recommendations to our national leadership of what to do. Yeah, let me focus on the testing issue for the three recommendations, because I do think that's the biggest priority is for us to get some value out of the testing. That's beginning to roll out Number one, we need to set up what's called a zero survey.

And this is something that as an Epidemic Intelligence Service officer of the CDC station in New York City, we had this outbreak of West Nile virus that killed a bunch of people, and we said, but we don't know if it's really deadly to old people or if a lot of people get infected and only a small number die. So we need to go literally door to door to collect blood from people to test their blood to see if they've been exposed to this virus that was in

nineteen ninety nine. We need to do that in New Rochelle now. So that's number one. So that's set up a zero study, which is literally a door to door, door to door. Will you gather data from each individual person who is infected and from those who were not visibly infective and evaluate that data. That's right, and the takeaway you'll get from that is what what will you

learn from that study? Will learn of a hundred people infected with the virus, how many end up going to the emergency room, being hospitalized, being in an ICU, and being dead. Because let me tell you, that number is not two point three percent, and it's probably not one percent. It's probably smaller. The fatality rate, the infection fatality rate is probably much lower, and that's important why because what's

gonna save us is herd immunity. At the end of the day, we have to use the fact that people are immune from this, whether through vaccination or through infection. And the good news would be if there are a lot of unnoticed infections of people who are now immune and can dampen the spread of this outbreak, walk me through this. So we do this close fine grand analysis.

It tells us with much more accuracy than we presently know of the number of people who are exposed to the virus, how many will be hospitalized, and how many will die. Then with that information we can make a better prediction about at what point we can start relying on people who are immune to start getting back into

the world. Is that right? And then we need an antibody test to test if people had been exposed, because there are lots of people out there on your hypothesis who've been exposed and haven't gotten sick and now aren't going to get the virus again. Assuming that it works like other viruses and not like the common cold where

you can keep on getting it. That's right. So this would give us the data which would then move us in the direction of enabling what what's the picture of the world where we've got this data and where we have an antibody test and we can say, Okay, you know, Noah's been exposed, but he didn't get sick, so he can now go out there and do what. If I'm a doctor, I can go back to work as a doctor. If I'm running an ordinary shop, can I go back and work in my ordinary shop now? Because I'm not

going to infect anybody exactly. The first use of this is honestly to inform our models of the world. If we're going to say that this thing is going to go on until thirty percent of the population or twenty percent of the population is infected, well, how many ICU beds is that it's a very different story if every ten people who get infected one of them needs an ICU bed versus if it's one hundred, versus if it's

a thousand. So the first thing it informs is the state of the situation we're in right now where we desperately need to know and do not know what the impact of this is going to be on our healthcare resources and facilities in the search capacity because we don't know the ratio between infected and the cases. This is super helpful. So basically number one priority is you can't plan if you don't know what actually is going to

happen in the world. And this information is so basic to figuring out what's going to happen that we can't do intelligent planning really without it. Correct And I was talking to a modeler from a university near you who was saying, I don't know that the future could go you know, many many different directions, And I said, what is the piece of data you need to make your models have smaller variants in terms of the outcomes. And she said, what I need more than anything else is

I need to know the percent infected. So okay, let's do that. The other application of it is what you said, which is and some have posited this, well, maybe we could have, you know, green bracelets for people who are already immune and they could end up helping run the society while the rest of us are in lockdown. I don't know about that use of the antibody testing, but let's start with the epidemiologic uses. So that's number one. Number two is we need to know within a given

city whether we're seeing widespread disease outbreak or not. And right now, in the absence of any guidance, in the absence of data, individual governors and mayors and others have made individual decisions, and I'm telling in some places it was too late, and I can also tell you in some places it's too early. And this is the problem with the germ of truth that the kind of cynics are having out there, of like, oh, this is much ado and we're overreacting. Well, in some cities, maybe we are,

but we don't know which. And so we need to have a systematic way of using the tests that we have and using the information we've already collected to be able to know is this virus spreading. Is it at the point where there's sparks we can stamp out with contact tracing, or it's too late to start to stamp out sparks. The whole house is on fire and you need to just turn the hose on and slow it down and make it go a little less fast. And how would we find out this information and number two.

If number one is door to door study, number two is just massive testing. I take it. Actually, it's not the number of tests, it's how you do the tests. Okay, tell me more about that. How do you do the tests? Yeah? The big problem is that we have two different public health reporting systems in this country. And if you just think about it, it kind of makes sense. Noah. Right, you go into the doctor's office and they draw your

blood and they send it to the lab. The lab then gets a positive result and they report it to the public health authorities. Right, that is the laboratory arm of public health reporting. What information does the lab have about you? Almost none? I take it. I just know that's your blood sample. Yeah. They know your name, and they know your day of birth, and maybe you're addressed, maybe not, depending Right. They don't know your symptoms, they

don't know your exposures. They don't know if you're hospitalized or going to be hospitalized, And which is why the CDC, in the Morbidity Immortality Weekly Report, the flagship publication of the CDC, had their first case report of forty two hundred plus cases positive cases in the United States. They said, we do not know the hospitalization status of half of them, So we didn't even know the age of ten percent

of the positive cases. So giving more testing done in lab core and quest and hospital labs that end up flooding the public health system with cases that we know nothing about is not helpful. It's not getting more testing out there, it's we want to have tests done where the laboratory results are tied to the key clinical and epidemologic data for us to make sense of it. So that key data is is this person part of a known cluster? What is their exposure? Did they travel? Do

they know someone? Do they who has it? That? We should ask people basically on a form at the same time as they're having their blood drawn, that forms should be filled out. That's the simplest form of this, right, And we should also ask them, oh, do you have any symptoms? When did those symptoms start. With those two pieces of data and the person's age in county, I can now construct an epicurve and I can tell you with those pieces of data, is the outbreak in this

city getting better or worse? But I need both parts of that data. I need the clinical and epidemiologic risk factor data and I need the lab data so where can we get both of those pieces of data. We have to set up sentinel surveillance sites where at the cost of getting the lab test, you also will have

to fill out the form. So this is where not just blasting the tests out there, but actually setting up some planful places where in an emergency room, every person who comes in with fever cough is going to get tested. Or in a doctor's office we set up doctors offices, we set up sentinel testing sites, or at a drive through clinic a drive through testing site, we make sure that we collect both pieces of information. That's how this

is going to get done. And right now I have heard roughly nobody create an actual funded plan to resource the development of dedicated testing sites that collect the information at scale sufficient to answer these questions? Why far as odd? Why is it the case that if something is as straightforward as you're describing it as being the sentinel sites, and I take it it's called sentinel because it gives you an early warning of what's going on, or in this case and not so early warning. Why is it

the case that no one is proposing that? And if I could make the question even a little meaner, you know, you were National Coordinator for Health IT for the Federal Department of Health and Human Services in the Obama administration. Why was this not part of what your team or the broader HHS community was trying to have in a contingency plan for the day that you knew perfectly well would some day come where a crisis like this would break out. Because we as humans lurch from panic to

panic in periods of complacency. That's what we do. We all do that, and there are some more extreme examples of where we let complacency take root. But I don't think anyone is blameless in forgetting you just forget what it feels like to be in this moment, Like we should make a list of the shit we're going to fix during the period of complacency between panic and panic. Like we should make that list, and we should now, and we should just stick to it for God's sake

and get it done. What's the barrier though, to simply a national edict from CDC that says, Hey, everybody in the country who's testing you must simultaneously fill out this form which we're posting online right now, and you must ask the patient about the progress of his or her symptoms. I mean, it sounds like of all the interventions we've you know, we can imagine that sounds like a pretty inexpensive one, except for the coordination of the data, which

I recognize would take some work. So look, the US system really does delegate public health to stay local officials. The CDC is an incredibly powerful institution, but mostly through guidance, yes, funding, but expertise, and ultimately they need to be the ones who are front and center, who are speaking with the voice of evidence based public health to the American people about what the strategy should be. And let me ask you, when was the last time the CDC was at the

podium at the Coronavirus Task Force. It's been some days. It's been many, many, many days, so we have not heard from and shook it. But to be fair, the CDC doesn't have to be at the I mean, that has some symbolic meeting, but the CDC doesn't have to be at the podium to issue a guidance on this, especially if it sees itself as, among other things, the

coordinator of National Data. I mean, if we had the head of the CDC here and asked her, you know, why haven't you done this, what would she be saying, I don't know. I don't know, Noah. And to me, one of my proudest career experiences was being at this CDC. It's a fantastic institution with thousands of incredible experts, and I just do not understand why they have not been frem and center and leading in the way that they know how to in this experience. I just I'm baffled,

and I don't have a good answer for you. We'll be back in just a moment. You've given us one in two suggestions, super clear, what's your third biggest recommendation? So the third piece of this is a system that I did play some part in really designing or creating some twenty years ago, which has now become commonplace practice

in public health, which is called syndromic surveillance. And this is saying remember I talked about how long it takes and the data problems of getting a lap specimen confirmed

with say coronavirus. The idea here was, well, people go to live their lives, and they register in the emergency room, and there's a piece of data collected for that, and they go buy medications and the phazinc at the pharmacy, and it goes deep at the counter, and you could gather up all those little bits and drabs of the exhaust of administrative data that governs our lives, and you can actually put it to purpose, putting your finger on the pulse of the city's health in real time and detect.

At that time, we were thinking bioterrorism. Now we're thinking coronavirus pandemic. And it turns out we spend hundreds of millions of dollars. And as part of the health information technology transformation that I helped push, we required hospitals to report every emergency room visit to these state public health systems in syndromes where you could group them and say, does the person come in have a GI syndrome or a respiratory syndrome or a flu syndrome? And so we

have the system. You don't have to build it. Now, you don't have to recreate it. We've spent a lot of money and resolved all the governance issues in state and local blah blah blah, and we're not using it. And again you're gonna asked me, why aren't we using it? I don't know, I don't know. I do not know. But the only place that has made that information publicly available is New York City. It's literally the website that we built fifteen years ago still works, and you can

go on that website. You can google ep query Queer y Syndrome Surveillance, and you can go there and you can click on the box that says influenza like illness or respiratory and you can see the percent of all emergency room visits daily up until I think Friday. They have data in there now and you can look at daily rates of emergency room visits in every emergency room in New York City, what percent of them were for

respiratory syndrome or flu like syndrome. And what you will see is that what has happened in the past two weeks has never happened in New York City before. I've been looking at this data for twenty years. Never ever have I seen a spike in illness that sharp, that steep, that fast. Four thousand, six hundred cases of respiratory illness or influenza like illness presented to emergency rooms in New

York City last Thursday, a year ago. That day it was sixteen hundred, almost a threefold increase in those visits. It is an incredibly powerful tool for seeing what is going on in the community. And is it actually don't tell me that we have it? How many cases tell me? Is it causing enough illness in the community to make a difference to be seen in the data? And we have it in more than just New York City. We

could look at it potentially in every state. And for reasons that I do not understand that data is not currently the centerpiece of our surveillance and response to this outbreak. What am I not asking you about that you see over the horizon going back to the national level as a potential problem that we haven't yet flagged. And I'm asking you that not because of your expertise only, but because you flagged a lot of the problems that we've

been seeing earlier than other people did. So when you look now two weeks or three weeks or even a few months down the road, what do you see as the most serious problems that are also not being discussed. I'm really interested in this confluence of politics and policy and data around when we go to these extreme measures and when we come out, and particularly if we're not able to mobilize suppression effectively enough that we can go

back to reclaim containment. That's what we have to be able to do to get out of this crisis without twenty thirty forty fifty percent of the population infected. Is we have to reclaim containment. We have to put out the fire and then really assemble crack teams of public health workers who can go around, stamp and out sparks

much better than we've done before. And if we can't do that, then we will be continually faced over the next eighteen months until a vaccine hopefully hopefully is developed, where we're going to be facing economic ruination and trying to decide make those hard trade offs between how much can we ease up and then see more people dying and then push back down again, And every policymaker is going to be having to make that, Every elected official is going to be making that decision based on their

own environment. So I hope that we can reclaim containment. I really really do. But if not, I think we're in for eighteen months of what I fear will be somewhat haphazard decision making around when to close, when to open, when to reclose, when to reopen back and forth. As far as said before I let you go, I do. A lot of people are wondering is there any hope here? You know? Is it all doom and gloom? What are your thoughts on that? Earlier I was much more freaked

out when no one was talking about it. It was just freaking me out. And now I'm actually much less freaked now that everyone's talking about it, because what I am seeing is, even in the absence of a plan, even in the absence of a strategy, even in the absence of data, I'm seeing massive behavior change in society, each person, each company, each school, each mayor deciding for themselves, each person deciding for themselves that they're gonna live life

a little bit differently. I'm not seeing very much handshaking right now. I'm not going to any conferences that airports are deserted. Like. This stuff doesn't have to be perfect to work, and I think it's working. We don't know if it's working. We won't probably for several weeks at least under the best of circumstances. But I'm optimistic that it's working because the average number of contacts just has

to come down. That's all we're trying to do to go from an effective reproductive number of two point five are not down to and are effective of less than one. Well, what that means is that if you had ten contacts a week before, you want to get down a four on average. If you can do that, will beat this thing. Right, The number of new infections that each person causes will be less than one, and this thing will extinguish on

its own. If before, on average, you went to the gym five days a week, and now you go no more than two. If everybody did that, this thing would snuff out. And I think some people are not doing it. Other people are doing it to a great much greater extent, and on average, I really do think all of us, acting individually, are making a difference. So keep doing it, America. Despite all of the stuff I talked about, at the bottom line, what matters is can we change our habits?

And that for me, the bright glimmer of hope here is Japan. Actually, because Japan did not pump out a ton of testing, but that what they did do is they embraced their sense of responsibility to each other. And I think that is in some ways more feasible for us to embrace than you know, contact tracing tens of

thousands of people in New York City every day. Well, if Americans can pull together by staying not together, then maybe they can accomplish exactly what you're what you're talking about, Brazza, thank you for helping us not go completely off the rails, but simultaneously, thanks for the clarity and honesty and directness of your analysis. Thank you. Noah, Well, there you have.

Farzad Mustashari, whose whole career has been trying to leverage data for public health, is very worried that we do not have the kind of data that we need and that it's not entirely clear we can get it without a substantial change in policy. That said, he does not

think that the world is over. And it's significant to my mind that somebody who was in his own terms, freaking out about this a month ago is now calmer than he was and does believe that our efforts at social distancing may be having good effects, imprecise and imperfect

though they are, so it's a mixed picture. We could be doing a lot better, we could be doing this a lot more rationally, but we're not facing in his view, the kind of existential threat that we cannot defeat based on the social distancing techniques that are presently being used. Until next time, be safe, take care of yourselves, maintain that distance. Deep Background is brought to you by Pushkin Industries. Our producer is Lydia Jeane Caught with research help from Zooequin.

Mastering is by Jason Gambrell and Martinezalez. Our showrunner is Sophie mcibbon. Our theme music is composed by Luis gera special thanks to the Pushkin Brass, Malcolm Gladwell, Jacob Weisberg and Mia Lovel. I'm Noah Feldt. I also write a regular column from Bloomberg Opinion, which you can find at Bloomberg dot com slash felt. To discover Bloomberg's original slate of podcasts, go to Bloomberg dot com slash Podcasts. You can follow me on Twitter at Noah rfeld This is Deep Background.