Tuesday, August 4, 2020

Out of the Crisis #17: Max Henderson on Covid Act Now, exponential growth, and how to help

From the very start, the coronavirus has been a fast-moving target. The mechanisms of where and how it's moved through the country have made managing the spread challenging. Max Henderson founded Covid Act Now to help with that process by making data-driven recommendations for sheltering in place (and later, for reopening) combined with public advocacy guidance to get local government to pay attention to the real facts about spread and safety. As he told me, "The cognitive load of just understanding what the heck is going on and where we stand and what's coming next is extremely, extremely difficult." The work of Covid Act Now has made it the definitive resource for getting cities and states to take action against what he thinks of as our common enemy--the coronavirus.

As the son of immigrants from Cuba and Germany, Max has deep faith in the power of American ingenuity. All of his work is designed to help it thrive. "From the very beginning," he explained, "our goal has always been to, one, be extremely clear, two, be extremely action-oriented, and three, be additive."

We talked about contributing solutions as an antidote to despair, being the target of a disinformation campaign as a result of putting real numbers about exponential growth into the public sphere, collaborating with experts, and how not to lose sight of the big picture even when your head is in the details. 

You can listen to my conversation with Max Henderson on Apple, Google, or wherever you like to download podcasts.


In addition, there's a full transcript of the show below.

 Highlights from the show:

  • Max introduces himself (2:43)
  • Grieving during COVID-19 (3:59)
  • Feeling powerless to help others and what he did (5:18)
  • On how contributing eases despair (8:05)
  • How he started Covid Act Now (10:52)
  • Description of Covid Act Now (12:00)
  • The Covid Act Now origin story (13:00)
  • How Max modeled Covid in January (17:09)
  • The danger of not understanding exponential growth (17:47)
  • Motivated reasoning and stay at home fatigue (22:01)
  • On being the target of a disinformation campaign (24:21)
  • The power of making a personal connection over the facts (28:44)
  • How we know the models are accurate (29:45)
  • The original model showing the effects of shelter-in-place and New York as example (31:53) 
  • The difficulties of managing micro knowledge versus macro knowledge about the virus (34:30)
  • Why Max made advocacy and action part of his work from the start (37:30)
  • The importance of public opinion and political air cover (39:35)
  • Covid Act Now's Resistbot campaign (41:18)
  • On American ingenuity and being the child of Cuban and German immigrants (43:47)
  • On getting the model right and not crowding out experts (46:37)
  • Transitioning from models to metrics and what the data shows (50:26)
  • Sitting in the eye of the storm (52:40)
  • Management and moving forward (54:01)


Show-related resources:

Transcript for Out of the Crisis #16, Max Henderson and Covid Act Now

Eric Ries: This is Out of the Crisis. My name is Eric Ries. Are we really ready to reopen here in the U.S.? I don't think so. It's almost like we've forgotten how scary, dangerous, or impossible this all seemed in the early days.

One of the recurring themes in these conversations is that exponential growth is hard to understand. We're not hardwired for it as human beings. Remember when the numbers were small and people dismissed a threat? Remember how quickly things went from okay to catastrophic? We are making the same mistakes as before, looking at our lower numbers and concluding that the virus is gone and everything will be okay.

But it doesn't have to be this way. We made drastic and radical changes before and we could do it again. Remember when shelter-in-place seemed like an impossible or radical idea? Remember how quickly it went from inconceivable to inevitable?

Max Henderson is a technologist who founded a group called Covid Act Now. They had simple calculators so that anyone could see the humanitarian toll of even a single day of their life. As the name suggests, Covid Act Now is not just a model, it's not just about data. It's about the need for urgent action in an exponential crisis.

They had an incredibly effective pressure campaign that helped ordinary citizens write their Mayor, write their governor with simple metrics like, "If we don't shut down, this many people will die." They had calculators so you could see how every day of delay could cost lives. That pressure campaign was a necessary counterweight to the lobbying that many public officials were receiving about the need to keep the economy open because the economic costs were going to be high. But what Max and his team understood is that there is no economy if the people who power it are sick and dying.

Through public advocacy and an unyielding commitment to data-driven recommendations, Covid Act Now became the definitive resource for getting cities and states to shelter-in-place. Now, they're turning their attention to modeling the reopening. And what have they found? Check the data for yourself. We are not anywhere close to being ready.

Understanding exponential growth is hard, but we're going to all have to get good at it for all our sakes. Here is my conversation with Max Henderson.

Max Henderson: Hey, my name is Max Henderson and I'm the CEO and founder of Covid Act Now. Prior to Covid Act Now, I was at Google and Firebase as a senior data science and Go-To-Market leader.

Eric Ries: Max. I really want to thank you for taking time to come talk to us during what has been a really challenging time for so many people. Before we get to Covid Act Now and, of course, action has been your watchword through this whole crisis, how are you doing? How's your family? What's your quarantine setup like?

Max Henderson: Yeah, of course, thank you so much for having me, Eric. It's been an interesting time. I and my family are safe. Thank God. We are-

Eric Ries: Glad to hear that.

Max Henderson:... distributed around the country and the world. And so, I think what's missing the most is just not being able to be together during a hard time like this, and I know lots of people are going through the same thing. I actually had a death in my family during this period, and it really highlighted-

Eric Ries: Oh, I'm so sorry.

Max Henderson: Thank you. It really highlighted the fact that you can't even get together to mourn. And so, it's certainly been real. I thankfully have... Me and mine are safe, and we are certainly in a more privileged position than most. So, despite all that, I'm very thankful for the opportunity to be able to have what I have during this.

Eric Ries: Do you mind saying a little bit about how you've been managing that grieving process when we can't use the normal tools of grief and the normal rituals that help people get through a hard time?

Max Henderson: Yeah, of course. It feels post-apocalyptic, right? I lost my uncle. He died suddenly in Germany. And my family there had to essentially drive through a border control that technically wasn't letting anybody across state lines, right? So, my other aunt who lives in another place had to literally drive to a checkpoint and explain why she should be let through when the borders were closed, to be able to go and grieve. Most of us who are not within driving distance had no chance of being able to get there.

The process has been extremely challenging and interesting. You have conversations on the phone that you would want to have in person. You can't hug a person. There are no words. I'm so thankful for being able to contribute in some way because it does give me a sense of galvanization and a way of channeling that energy into trying to help others and just move the effort forward. I think if I had less to do and less ways of being additional or additive that it would have made the situation a lot harder.

Eric Ries: I want to dig into that. What's been the hardest thing for you while we've all been in quarantine during the shelter-in-place?

Max Henderson: Hardest thing? That's a great question. I think--I'm very blessed. I was able to bring some friends on quarantine with me and so, I've, at least, had some sense of community during this and-

Eric Ries: That's great.

Max Henderson:... being busy. My family and my significant other have had other folks to interact with so that they're effectively, to some degree, mourning the loss of me because I'm constantly working on other things and they support me a ton in that, so I'm very, very blessed.

I think the hardest thing has just been the desire to want to help others. I realized, this is a bit perhaps of a cop-out answer given the fact that I have found something to channel my energy into, to me being in control in a crisis-response situation has added so much psychological safety that the personal elements of being in quarantine have really fallen into the background compared to just the urgency of being able to execute on all these things we want to do.

And so, the reality is that, for me, and I don't know if this is a particularly satisfying answer to this question, but I've received so much from this opportunity to serve in the sense that it's allowed me to take the feelings that I feel and I'm sure everyone feels around lack of control, lack of clarity of the future, lack of an opportunity to participate and channel those into productive things. It's just helped my psychology so much compared to the way I know I would feel if I was just sitting on the couch wondering what's going on and what I can do and just feeling completely powerless.

Eric Ries: I'm glad you said that because it's actually a pretty consistent theme among the folks that I have talked to, who are in the fight trying to make a difference. And look, I've had my moments where I envy the people who were bored during the crisis and wonder what it would be like if I was running out of Netflix shows to watch. But actually, having now talked to a number of those people and convinced them to volunteer and to help out, it seems like it's paradoxical to me being busy and being exposed to these dark facts, especially with the work that you've been doing, having a real sense of the danger of the epidemic and the scale of it. I think there is a psychological security in it.

It's just one of those really clear moments where the people who are being the most of service, they're not actually... It's not an act of self-sacrifice. It's actually psychologically very healthy to feel like there's something you can do in a situation that for most people, there really isn't.

Max Henderson: Yeah, 100%. They say that depression is fundamentally the feeling of being unable to control your own destiny, right? And that feeling of despair is something that I certainly have had less of because I've been able to participate in this way. And to be clear, I mean, there are days where I wake up and I'm like, "Wow, another 18-hour day. I'm exhausted." The work is extremely galvanizing. And it's rare that there is an opportunity, there is something that is so pressing and emergent and relevant that is also so hard, so operationally hard, so difficult from a data perspective.

This is one of those things that no matter how deep you dig, it is truly like a problem unknowable by one person. It just grows fractally in complexity, like the data is bad, our understanding of the disease is bad, human psychology and our ability to stay in these really difficult interventions is a complex problem. And so, there's so much complexity here that you can really dive into it. And I think that has had an incredibly positive effect paradoxically, on my psyche, totally.

And it doesn't surprise me at all that other people have had the same experience. It's a lot like giving, where you think giving is going to be a thing that only benefits the receiver, but ultimately, it benefits the giver as much as the receiver.

Eric Ries: Or maybe even more.

Max Henderson: I had a very similar experience. Or maybe even more, yeah, that's right.

Eric Ries: Yeah, that's certainly been my experience. I'm incredibly grateful for the opportunities to serve even when they've been hard. And I know there's some people listening who are like, "Oh, come on," but just you'll have to take these... Here's two testimonials that that's been our experience during a time when I honestly think despair is the real enemy even more than the virus. Because we can outsmart the virus, we have the tools and the technology and the science.

But the question is, will we have the will? Will we have the ability, the social cohesion to actually take the actions that are needed? And that's what I really appreciated about Covid Act Now and the work that you've done, is you're fighting that. That's our real enemy.

Max Henderson: Yeah, that's right. I mean, people might say like, "Oh, gosh, get out of here. You're being so holier than thou," but I really think it's a much more practical thing than that. I think understanding... I'm a person who has always felt an incredible need to wrap my head around problems and just understand them. And so, it's purely... In some way, it's selfish.

By way of working on this full time, I get to have a better understanding. This whole thing started with me just being like, "Wow, I don't understand this." And there's such an overwhelming amount of information on a context that I can't imagine that many people understand it either. I'm not special. So, if I don't get it, then probably most people don't get it. How can I take all this complexity and collapse it and just make it so easy? If we can't all get on the same page about what's happening, it's impossible for us to make decisions. Like a political decision maker will not be able to say, "Hey, I'm doing this for this reason because this information and mistrust will just overwhelm that person's political capacity to make our political capital to make the decision."

People will argue about what the right solution is instead of just focusing on the solution, and the timelines are too short for that. So, it felt like, for me, the starting point was I need to understand what's going on. And then, it quickly became, "Well, if I've understood what's going on in a way that's simple, then I have a responsibility to share that with other people." And that fundamental thread is ultimately what developed into Covid Act Now.

Eric Ries: Say a little bit about what Covid Act Now is for those that don't know.

Max Henderson: So, Covid Act Now is a tool that was trying to send one very specific message, and that is that immediate action is required in the face of exponential growth in order to prevent catastrophic outcomes, right? So, the disease grows exponentially, and human beings don't think in terms of exponentials. We think in terms of linear change, so one goes to two goes to three goes to four. The disease fundamentally doesn't work that way. The disease grows exponentially, so one goes to two goes to four goes to eight. And those numbers start small, but they eventually become extremely big.

And just being able to help people build the intuition to understand that if we don't act now, catastrophic outcomes might result was the original takeaway for me in my research, and the thing that I felt and the group of us that started this felt a calling to share with the American public.

Eric Ries: Tell us the origin. So, was there a moment for you when the gravity of the pandemic crystallized in your mind?

Max Henderson: Yeah, there was a moment. This was really interesting. There's that commencement speech by Steve Jobs where he says that you can really only connect the dots in retrospect, and that it sounds like a trope, but it's been so real for me in so many places in my life. I've always had an interest in medicine and was in EMS for a while.

Eric Ries: Explain what an EMS is for those that don't know.

Max Henderson: Oh, yeah, of course. So, EMS is emergency medical services. So, think, emergency medical technicians, ambulances. EMS folks deal with disorders and cardiovascular disorders on the regular, right? Most emergencies that people have are either breathing problems or heart problems, major injuries. And so, I've always had an interest in this sort of thing. I thought for a long time that I wanted to be a doctor, spend some time in the pharmaceutical industry as well, and got a degree in System Dynamics. I've always been interested in nonlinear systems. And so, this was a really interesting coming together of two areas of interest for me academically in an extremely morbid, but also extremely practical way.

And so, I was on sabbatical studying to get my EMS certification and came across COVID. And I remember being extremely fascinated by it from the very beginning. In January, I started building a model just to try to understand as it was starting to spread it in China, just understanding the nature of the spread. It was designed just to share with me and mine, right? It was focused on what might happen in the Bay Area where I live if it started to spread here.

And it was after some of the data started to roll in that I realized how quickly it was growing. I had this epiphany, that we were already starting to have a case or two in the Bay Area, just how quickly this thing could possibly grow, right? We're talking from the first reported case to potentially thousands of hospitalizations in only a couple weeks. I had an aha moment that this information needed to not just be kept amongst me and my community but needed to be shared as widely as possible.

Within a few days of me sharing my model out just to my network, I ended up having to create a newsletter because thousands of people had signed up for updates. And eventually, after about a week or so of sending updates, I got approached by some folks that I had never met before who said, "Hey, Max, there are lots of states that don't have an analysis like this." I mean, keep in mind, this is a spreadsheet, right? I mean, it's not some super complex... Certainly, the thing has evolved, but back then it was just one tab in a spreadsheet. "You need to make something like this for every state."

And so, I agreed that I would do that, and I partnered with my other founders, Zack, Igor, and Jonathan. We built a copy of the spreadsheet for every state in the union, and Igor built the website. We hooked it up as quickly as we could. And within 48 hours of launching the thing, nearly 10 million Americans have come to the website and seen our predictions. And from there, as these things go, every time more people saw our predictions or relied on our math, the more galvanized and the stronger of a calling I had to continue and elaborate it and make it more reliable and complex and feature-rich. And so, we've, from that early success, just continue to build.

Eric Ries: How did you make that original model? I think what's really interesting, you're talking about like it's no big deal. But most of us encountered that same set of information earlier this year, and very few of us had the insight that you did to say, "Hey, let's build a model and see what could happen in the Bay Area." But how did you actually do it? What does it mean to model a disease like this?

Max Henderson: Yeah, it's a great question. I mean, the thing is, you can always make models more complicated and all models are always going to be wrong, right? So, I had to start with that, you're never going to have a perfect model. And so, in some ways, simplicity is better because something that's simple is easier to understand.

And the reality is that the most important thing here is the nonlinear behavior of the disease. I mean, I'll start by saying that, like I said, I got a degree in System Dynamics, and so spreadsheets and models are my jam. I'm no world class nonlinear systems expert at this point.

Eric Ries: Well, but you were on sabbatical from Google.

Max Henderson: That's right. But I mean, I hadn't touched a nonlinear model probably since college, right? So, the key insight here is not some complex modeling insight, but rather that the numbers, whether it's hospitalizations, deaths, infections, etc., all double on some fixed period. So, the simplest possible analysis here really is just take one cell in a spreadsheet and then measure how many days all the systems... So, when a nonlinear system like this is in exponential growth, every single category, infections, deaths, hospitalizations, they all grow at the same rate. And that rate is some doubling period that you can measure from the empirical data.

So, if you've got a four-day doubling period, which is what it was in most places before interventions went in place, the most simple analysis here is just one column for deaths, one column for infections, one column for hospitalizations. And every four days, the number in each column doubles. And everything else is really refinements and simplifications.

Obviously, at some point, when everyone is infected, things are no longer growing exponentially so there are some... Obviously, it doesn't continue forever. But that's the fundamental analysis, and it's actually really, really simple. So, it doesn't take the... Oftentimes, I found in this and other things that I've worked on that the answer, and this is something we've tried to keep in Covid Act Now until this day, we're done when there's nothing left to take away and not when there's nothing left to add.

So, we can always make the thing more complex. And the model is now this complicated Python thing that takes a 96-core machine 12 hours to run, but the fundamental dynamics are very much the same. And that is the scary thing about the fall that we face, right? Especially now that we've delayed a potential spike here, I think we're forgetting the fundamental nonlinear dynamics at work. And if we forget that, it's going to be very much to our detriment.

Eric Ries: It's driving me crazy. I was just talking to a very smart person who is saying, "Well, because here in the Bay Area, we've got the disease under control. Therefore, hospitalizations are down. And just looking at the linear data, therefore, it will be safe to open back up and therefore, there won't be danger, exponential-type danger anymore. We're past that phase." And it was a completely reasonable inference. I completely understood where they were coming from, but it's dead wrong. And if people draw that conclusion from the fact that these early interventions worked, aren't we in a lot of danger?

Max Henderson: Absolutely. Yeah. I mean, I actually heard one of the leading epidemiologists in Germany call this the prevention paradox, and I loved it. The idea that because through positive action, you stopped the bad thing from happening. The bad thing either was never going to happen or has gone forever now and can never come back, right? And effectively what we've done is buy ourselves time. That's it.

Every person who hasn't yet been infected is a potential person who can be infected once non-linear growth comes back. And when it comes back, it'll start off slow. At first, it'll seem like everything is fine when we're past the worst of it. Because when that one goes to two, and two goes to four, and four goes to eight, the numbers are small. But when you're at 500 and it goes to 1,000 in four days, and then it goes to 2,000 in four more days, and then 4,000 in four more days like we saw in New York, that's when things really become completely insane. And the scary thing is we all have to check our bias on this all the time because it's just not a way... It's not intuitive. It's never going to be intuitive.

And so, you really do have to look past the intuition of this thing. And even I, looking at the math every day, sometimes I feel it. I feel the, "Oh gosh. We've been doing this for two months, and it doesn't feel like it can continue and nothing bad has happened. So, was this all maybe an overreaction?" I know for certain the answer is no, right? There's no question in my mind whatsoever. But I think feeling that and knowing that that emotion is there is a super powerful thing because we're all feeling it, right? Everyone's tired. Everyone is uncertain. And if we-

Eric Ries: And motivated reasoning starts to kick in.

Max Henderson: Yeah, that's right. Motivated reasoning starts to kick in. And it's so subtle, you don't even notice it's happening.

Eric Ries: I have friends who wanted to have a party, and they were trying to convince themselves that it was safe to have the party and they were going to socially distance at the party. They had a whole plan for why it was going to be safe. But I noticed in the planning of this party that not only were they trying to convince themselves that it was safe, but it started to spill over into them making justifications of the pandemic wasn't actually that dangerous.

I don't think they were consciously aware of the fact... The fact that they were just tired of being stuck at home and missed each other and wanted to have this party was starting to affect their ability to reason objectively about the data that they were seeing from the outside world, never mind the fact that the data is grim and it's hard to face it head on.

It's funny you talked about the paradox of prevention. I actually wrote a blog post called The Curse of Prevention, but not recently, back in 2009. I had to look it up. It's that old. Where I was grappling with this, because in engineering, this comes up all the time, because how do you know when it's worth it to pay the cost of a mitigation of something that may never appear? And there's this political, we should call it what it is, this political impulse to use the curse of prevention to attack something that you don't want to have happen. And you can use it both ways.

You can say, “You're alarmist." The thing wasn't really true that we had to do the lockdown and criticize that way. You can use it the other way. You can make up potential dangerous things that could force someone to do a prevention that they don't really want to do, which has been the criticism on the other side. You've had the distinction of even being called fake news, which is one of the highest honors you can achieve in our current crazy times.

Talk a little bit about what that's been like to be on the receiving end of this idea that you're somehow manipulating the data or trying to serve some... I'm not even sure what your secret nefarious agenda even would be, but whatever it is, how do we know that this data is accurate and what's it been like to be the subject of this disinformation campaign?

Max Henderson: Yeah, it's been surreal, right? Because never having been the subject of something like this before, you'd think, "Oh, well, maybe there's some nuance or confusion." Where is the line between stretching the truth and just making something up? I can confirm there are people out there who are just willing to make things up.

Eric Ries: Just arguing in bad faith.

Max Henderson: Yeah, yeah. Arguing totally in bad faith.

Eric Ries: 100%.

Max Henderson: It's really, really interesting. Gosh, how can I describe it? Yeah, surreal is probably the best way to describe it. I mean, the funny thing is that’s the one major piece that's ever gotten put out on us, and we certainly get a lot of hate mail and things like that and way more positive encouragement thankfully. Everyone gets their share of malicious actors or people acting in bad faith.

The one article that really got written that I think wasn't most in bad faith also contained something in it that was like, "And these guys are predicting 13,000 hospitalizations in New York in two weeks. Can you believe it? And the number ended up being like 19,000, so there's some... It didn't age very well.

Eric Ries: Yeah. And I'm sure they didn't print a retraction and apology.

Max Henderson: No.

Eric Ries: But they're just onto the next conspiracy theory.

Max Henderson: They certainly did not. Yeah, nobody ever went back and, in fact, check that. Honestly, I don't know what our nefarious ends would be. But more than anything, it makes me sad that we are... At least there are some portions of our national discourse that are so divorced from reality that we can't help. Almost my reaction to these things has been like, "Gosh. Well, I wish I could just sit down with you and show you that I'm a human and that we're the same," and just show you what I see. And rather than argue with you, just try to bring you in and show you like, "Well, here's the challenge that I see. Let's just logic through this together."

And how do you see us solving this problem? What is it here that you think is fake? What is it here that you think my motivation is? Or is it just because like I'm some faceless person somewhere on the internet that it feels like this is the right way to handle the situation?

Eric Ries: Believe me, having done that all with such people, I mean, write a book and they'll see this experience and see what it's like. I think I can safely predict that not a single one of those people would be willing to sit down with you precisely because I think, at some level, at some visceral level, they understand that it would work. This bad faith nonsense, it is only effective if you can hide behind a veil of pseudo anonymity and you can really other the person that you're talking to. And if we actually had the connection and trust and love, right, the fundamental human connection that is needed, it's not possible to sustain that kind of ideology. And therefore, those things are actually dangerous to it.

Max Henderson: Yeah, I try. I certainly don't have the time to do this with all of them, but I try as much as I can when people post a troll comment or something or send me like a really, totally, unacceptably vitriolic email like, "Hey, you guys are the worst. You're trying to destroy the country," or even worse things that I won't even mention.

Eric Ries: We will not be linking in the show notes, but people can find them if they would like.

Max Henderson: Yeah. A lot of them are coming privately directly to me, I try in almost every case to react with exactly the opposite. I mean, some of them, I just can't engage with, but I try to respond with exactly the opposite language and say like, "Hey, thank you for bringing this up. You're so right to be skeptical. There's so much bad information out there. Here's what I see. And really, I see this situation as a once in all of our lifetime opportunity to save lives.

I love this country as much as anybody else does, and I'm trying to do the right thing. These are the facts as I see them and I'm willing to have a debate with you about what you see as being problematic here." And 90% of the time, people don't take me up on it. But about 10% of the time people do actually like, "Wow, I never expected an answer to this email. I'm sorry. I see what you're talking about. I have these two other questions." And then, it ends up turning out to be a productive discussion.

Eric Ries: That's outstanding.

Max Henderson: Those things are making me, honestly, they make me happier than almost anything else when I can actually make a connection with a person who is coming from such a place of mistrust and, I mean, I don't want to use the word "ignorance" because it makes it sound like I am somehow the keeper of knowledge, but mistrust and negativity. They can turn around to something as positive as a real personal connection and an alignment on some kind of objective reality.

Eric Ries: So, how do we know the models are accurate?

Max Henderson: It's a good question. We're working. So, I think first of all, we have to really understand what accuracy in the context of modeling means, right? All models are wrong, some models are useful. And the goal of modeling really more than calling an exact number X number of days, weeks, months out is to understand the variety of positive outcomes, find the catastrophic ones and figure out how to prune those, figure out how to make sure that those catastrophic outcomes never happen, and build intuition for how our actions can change the future.

You don't have to be right on every single right layer. There's so many ways a model can be right or wrong. It can get the actual number of hospitalizations or deaths or infections wrong, or the relative ratios of those things to each other, or exactly when they happen or what the shape of the curve is exactly. So, there are just so many different ways that things can be wrong.

And then, there's also the fact that this is a system where our expectations of the future change the future, right? If I expect that a huge spike of infection is coming and we shut down, the huge spike of infections doesn't come or at least it gets delayed like the situation we're in now. And so, in a world like that, right and wrong are a little bit more complicated than just like, "Could you call the exact number?" So, we think about this in a couple different ways.

One is we think about getting the shape of the curve right and understanding how the scenarios differ relative to each other. So, if I do X versus doing Y, what is the difference that I can expect on average between those two different outcomes, even if the exact numbers are going to be slightly different? The second is how do my actions change reality? Our original tool showed you what happened if we went to shelter-in-place and what would happen if we didn't. Even if those curves don't perfectly mirror reality, the relative difference between them is incredibly important for building intuition.

Eric Ries: Actually, if you don't mind, talk about what the model said at that time, because I think many of us have gotten used to the idea that we're going to do shelter-in-place and we forget that it was like five minutes ago, in historical time, that that was an extremely controversial idea.

Max Henderson: Yeah, I mean, when we first started talking about shelter-in-place, it wasn't even clear that this was within the realm of possibility, right? I mean, when we put the tool up, again, we were not telling anyone to do anything necessarily. We were just pointing out the speed at which a decision needed to be made because these were the relative possible outcomes that we were dealing with. But I mean, we were talking, at that time, about hundreds of thousands of hospitalizations in just a few short weeks.

I mean, we saw New York was the last to act. And so, New York really provides one of the few real examples of how bad it could get, and given it's an urban area, but even in the state, how bad it can get, 0.1% of the entire population of New York state has died from COVID, not 0.1% of infected. It's literally 0.1% of the entire population of the state. And last we checked, the antibody testing to see how many people had been infected came out to somewhere between 10% and 20% depending on where you were in the state. So, you're talking... This could have been five times as bad at least.

And so, the models were showing something actually quite similar to that. It was basically somewhere around 1% to 3% fatality rate for anyone who is infected, and probably about 70% or so of people infected before herd immunity would be reached. And so, you're talking about 1% to 2% or so in that range of fatalities in this country. So, at a population of 300 and 30 million, you're talking about somewhere between 3 million and 6 million people.

We've learned a lot more about the disease since then, but the numbers actually have not shifted that much, right, now that more data has come in about infection rates and we actually have some antibody tests. Keep in mind, at this time, we didn't have any. So, we didn't know how many people were actually getting sick. We just knew how many people were getting sick enough to go to the hospital.

The numbers have come down a bit. Our estimates of death rate have been cut roughly by a factor of two. But I mean, even a 1% death rate or a half a percent death rate in the United States is still somewhere between 1.5 million and 3.5 million people.

Eric Ries: Still a catastrophic loss.

Max Henderson: Still a catastrophic outcome, and that hasn't changed. That is our latest thinking. That is our more optimistic thinking.

Eric Ries: It seems like that, in some ways, because we're having such a hard time forming a national consensus about the facts as we learn more about the microstructure of the epidemic and the disease, and the effects of vitamin D, and the antibody tests, and all this micro knowledge, we lose sight of the fact that the macro facts have been established and are pretty well understood and have been pretty stable through the whole thing, namely this is a highly deadly, highly contagious, exponentially growing epidemic. And nothing can or will change that until a vaccine is developed.

Max Henderson: That's right. That's right. This was one of the reasons that we created and continue to work on Covid Act Now is that the reality is that we are doing a really bad job of paying attention to the forest and not the trees, right? Our national discourse is largely occupied by either, "Here's the number of deaths there were yesterday, or here's the number of hospitalizations there were last week. Or hey, someone said that chloroquine is a great treatment for..." We're dominated either by facts out of context or anecdata, and so it becomes incredibly... The cognitive load of just understanding what the heck is going on and where we stand and what's coming next is extremely, extremely difficult.

Using a weather analogy, instead of saying like, "This is what the temperature is going to be. Oh, you should bring an umbrella because there's a 50% chance of rain," well, right now in terms of COVID communication is, "Well, it was 75 degrees three days ago at 5:00. It has rained at least three times this year," which is not at all useful for understanding what you actually have to do.

Eric Ries: And who could really trust those umbrella manufacturers anyway?

Max Henderson: Right. Exactly. Maybe, "I've never seen rain, so maybe rain is not a thing. And I don't even need an umbrella."

Eric Ries: Yeah, distraction, delay, obfuscations act as information.

Max Henderson: Exactly. And when you factor in the fact that there, obviously, massive economic, public health, and mental health concerns even to the extreme measures we're taking, it becomes a massively complicated thing to understand.

Eric Ries: I kept thinking about-- in the early days of the crisis, especially in that wave when shelter-in-place was coming into effect, I tried to imagine what life must be like for the principal policymakers, especially mayors and governors, who must have been hearing from endless business lobbyists. None of whom I'm sure were saying we need to shelter-in-place. I'm sure they were all saying, "I need you to understand the catastrophic impact, the actions you're talking about taking will have on my industry and you better believe will remember who blah, blah, blah." The usual lobbying, self-interested playbook, that must have been so loud in the ears of policymakers. And one of the very few counterweights to that was the campaign that you were running with Covid Act Now.

So, talk about why you decided to take action and advocacy orientation versus just having like a neutral model at the beginning. What did you do and why was it important to you that policymakers have somebody telling them the actions that they could and must take?

Max Henderson: Yeah, great question. So, first of all, let me say I have... I mean, mistakes have been made and will be made, but I have just an immense respect for every policymaker and decision-maker in this situation. I mean, talk about the worst possible situation you could possibly be in as a decision-maker, you are responsible for every life lost and you're responsible for every job lost. And no one is ever going to know how it would be if you had acted differently. And so, you are the risk vessel. You are the person responsible no matter what, and that's a lose-lose.

So, from the very beginning, our goal has always been to, one, be extremely clear, two, be extremely action-oriented, and three, be additive. So, the goal here is not to say you have to do X or you have to do Y, but rather provide information that allows decision-makers not just to... They're suffering from the same information overload all the rest of us are. So, distill the information down to something so clear that the action is obvious, right?

The next thing to do, forget 10 levels down the decision tree and all the detail, just look at the situation in its most basic clear form and make the decision that obviously needs to be made no matter how hard it is. That, in and of itself, is a difficult thing. It's a really, really hard thing to do and something that we've been focused on as being the primary difference between us and all the other models out there, right, because there are now dozens of them, and they all do great work. Some of them are private and some of them are public. But the reality is that we don't necessarily need more complexity. We need less complexity and more clarity.

The other thing that I would add here is for decision-makers specifically, I think we overlook how important public opinion and political air cover is to making the right choices. In this country, we believe in freedom. You mentioned like a shutdown would nearly have been unmentionable or wasn't unmentionable when we started this thing, and if you're willing to convince the American public that something like this is necessary, and I'm convinced it was and continues to be necessary, then you need to clearly explain why that's the case. And you can't just hand over a model that only a PhD expert can interpret and say, "Well, we're just doing this. You got to trust me." That's just not the level of national discourse in this country.

And so, we wanted to create a tool that wasn't just telling decision-makers how to act, but that was simple enough that the general public could see it, understand it, and provide air cover for their leaders to do what was required. And one of the things that worries me the most right now is that as we shift into this prevention paradox, "Oh, well, maybe this was all an overreaction, or even if it wasn't, we're past the worst and it's time to go back to normal."

If we don't have similarly clear graphics to explain why either we're not there yet or what it will take to get there, then we're going to lose control as a nation. We're going to lose control over our decision-making here because ultimately, we're going to go back to something much more reactionary and less first principles-driven because it's just impossible for everyone to agree. And so, what ends up happening is decision-makers are forced to just take the average of public opinion and do that, right?

Eric Ries: Talk about the campaign you did with Resistbot to have people write to policymakers and try to motivate them. I was really proud to see that come together.

Max Henderson: Yeah, of course. So, the campaign we put together is we essentially sent out text messages to several million people who were Resistbot members to essentially let them know, on a push basis, what was going on in their state to actually look at the numbers and to come to their own conclusions about what action needed to be taken, and then make it really easy for them to communicate with their electeds about what their opinion was, right?

So, essentially taking the website and instead of forcing you to find it, sending it to people and just laying the information out and saying, "Look, here's a couple different scenarios. You can interact with them and see what the outcomes would be. And if you're convinced that action is needed, you should let your electeds know." It's that kind of thing, and I'm super proud that we did this. That, I think, made probably a bigger difference than modeling, right?

The model is a tool. It's a tool to an end. And that end is ultimately to get us as a country to make faster and better decisions. We could create the fanciest model in the world and be extremely proud of it on an academic basis, but if it doesn't actually create change, then it's a waste of time, in my opinion.

Eric Ries: Quick shout out to Dustin Moskovitz who underwrote the cost of the Resistbot campaign at a critical time. What role do you think that campaign and Covid Act Now played in convincing policymakers to take that drastic action to save lives?

Max Henderson: Wow, what a question. I'd like to believe, I mean, having communicated directly with probably half the states in the union, either in the public health department or in the governor's office, my sense is I believe I played a non-trivial role. I mean, at the end of the day, I'm much less worried about the role that we individually play and more worried about us getting to the right answer.

And the only reason I even care about the role we play is purely just to make sure that we are adding to the signal and not adding to the noise. We're actually doing something that has any effect whatsoever and we're not making the situation worse in some way. Because it is possible to have negative impact, right? It's possible to-

Eric Ries: Absolutely.

Max Henderson: I like to stay curious about our impact, but the truth is we're all on the same team together, right? I mean, my view here is this is... Depending on how you look at it, this is the first time that we've had a real enemy on our soil, and the American ingenuity can overcome anything. I mean, I'm convinced of that through and through, right? I mean, I'm a first-generation immigrant. My parents came here for a reason. As long as we are all on the same page, we can do anything. And so, if I can help, if our organization can help get us all on the same page, even just a little bit, that is work worth doing.

Eric Ries: Where did your parents come from?

Max Henderson: My father is Cuban. He fled the Castro regime in the late 1950s, early 1960s, and actually fought in the Bay of Pigs invasion. So, he was part of the entire Cuban missile crisis. I wrote a book about it. Mom is German immigrant, and she also came over here around the same time in the '60s. They met in Miami which is, I guess, where a Cuban and a German would be.

Eric Ries: I guess that that makes a lot of sense. And I appreciate you--there's nothing... We often see that immigrants and the children of immigrants have the most optimistic view of what America is capable of. And I appreciate you tying it to that patriotic sensation that we all ought to have, that we have a common enemy. It's really, in some ways, our truly global enemy, and that coordination, cooperation, solidarity, those are going to be essential ingredients to combating.

Max Henderson: That's right. And that's why I see there being a real missed opportunity to come together here, right? And I realized that either out of ignorance or out of malice, there were those that want to divide us. But the reality is that this is a unique opportunity for all of us to come together across the world, but even more so inside this country to find something that all of us have a common stake in, right?

I mean, all of us have parents and grandparents and people we care about that we don't want to lose but more importantly, whether you care about National Defense, or whether you care about protecting this country and what it stands for, or whether you care about the public health outcome, the reality is that all roads lead to Rome here. And so, if you choose to see this thing as the first successful invasion of the United State by some foreign enemy, then feel free to see it that way. But when you... I can't see how anybody looks at this and says like, "Oh, it's not a big deal, or we should just ignore it." That, I really cannot understand.

Eric Ries: You talked about how important it is to make sure that you're adding to the signal and not the noise. Can you talk a little bit about some of the guardrails you've put in place as this has grown to make sure that the data is accurate, obviously, within the parameters you described, but also that you are integrated with the scientific and public health communities? And this is not just tech... I think of the caricature of ignorant tech people trying to metal in scientific matters we don't understand. Just talk a little bit about the seriousness with which you've taken the need to get the model right.

Max Henderson: There are two things here, one is getting the model right and the other is not crowding out. We're not adding so much, just adding so many more voices to the discussion that the true experts cannot be heard, right? And we think about them both separately.

So, getting the model right is actually almost, in some ways, the more straightforward thing because there's a lot of prior art in epidemiological models that we can reach into. So, our model, we ended up moving from a spreadsheet model to a model based on one created by Alison Hill at Harvard that is open source. We ended up taking that thing and modifying it to have a couple more features, but ultimately, fundamentally operate in the same way.

We've partnered with Stanford and Georgetown epidemiological and public health folks there as well as other advisors like Nirav Shah, who is ex-Health Commissioner in New York and also a lecturer at Stanford and ex COO at Kaiser Permanente, to help provide us the guidance to make sure that we're sending the right public health message, the model has the right inputs, the output is believable. And so, our approach there has really leverage existing well-respected prior art and bring advisors into the fold so that we have...

We're not making these decisions on our own, right? The value that we add is the engineering work, being able to scale the thing up, communicate it clearly, essentially taking this great academic work and turning it into a consumer product that is reminiscent of the best in class tech products out there. That is the value that we're adding, but the guts are all public health and epidemiological people. So, that part is pretty straightforward because it's just like, "Hey, let's go find the experts and let's just amplify their thing."

I think that the much harder thing here is figuring out if you’re additional. Are you crowding out the experts by just adding more noise? And are there others who are sending pathological messages that you have a responsibility to answer to or to amplify the right message in the face of? And this is a much more difficult question. I think we've ultimately come to the conclusion because we're working with experts who tell us, we rely on them to tell us whether they believe our message to be additional, and we ask the question, "Are we actually adding to the message here? Are we saying something that others aren't already saying? Or are we crowding them out?" And we've come to the conclusion that, again, where we really add the most value is making the message approachable.

And that is the thing that it seems nobody quite had the ability to do, and it's not because we're special, it's just the multi-disciplinary nature of our organization. We brought together epidemiological experts who really understand the science, public health experts who really understand what the public message for that science needs to be, and tech folks who understand how to take a message and make it super approachable, super easy to interact with, super polished, and just very confident and inspiring. It's that combination of things that has really made us additional. The mixture really is greater than the sum of its parts.

Eric Ries: So, Max, tell us what the models are saying now. Where are we in this crisis?

Max Henderson: It's a great question. So, we've recently transitioned much more from models to metrics. And the model was really, really useful when we were trying to understand how non-linear behavior like explosive exponential growth was going to impact us in the short term because it's so counterintuitive. Right now, we're in a state where we've at least delayed, not avoided, but delayed the worst.

We've transitioned to a much more of a metrics approach. The difference here being rather than trying to understand what the future is going to be like, I'm trying to set goals and I'm trying to manage against goals, right? So, in this case, reopening. When am I ready and what does a good reopening look like versus a bad reopening?

There are a couple things that are clear. We are still maintaining the models because once we get through reopening, there's obviously risk of a second spike. And so, having the models in order as well. So, there's a couple really clear takeaways here. The first is we have not prevented. We have delayed what is potentially the inevitable unless we work on a holistic containment strategy.

The second is that the lockdown containment strategy, although it is extremely effective, and the best thing about it is that you can do it really quickly, is I think we all know is fundamentally unsustainable and we need to transition out of it as soon as possible. Because unlike some of the criticisms that are levied, I think all of us are intimately acquainted with the economic and human cost of being on lockdown. Nobody thinks that this is a free lunch. So, we need to transition away from this as quickly as possible. And the way to do that is through some combination, a slow-controlled reopening where we carefully and conservatively test how much disease growth comes back as we introduce various different parts of the... Or reopen various different parts of the economy, coupled with testing and tracing, right? So, coupled with both the opportunity.

Essentially, we have to know where the disease is spreading and how fast it is spreading at all time. And the way that we do that is we test as many people as possible to understand where the disease is, and then we trace all of their contacts and we test them as well. And anybody who's positive, we isolate them such that they cannot re-infect people. This way, we keep each infection from infecting less than one other person and things start to trend to zero.

I think our biggest risk here is that we believe we're outside the storm, but we're really in the eye of the storm right now. I'm from Florida, so you're going to hear hurricane analogies for me a lot. We're in the eye of the storm. And if we lose, if we take our eye off the ball, we will have as big of a spike as what we were predicting before this whole thing started. But we will be coming at it from a much weaker economic position, a much weaker emotional position because I know we're all tired.

And so, the main thing to pay attention to here is we are not out of the woods. And if we're not careful when we drop the ball halfway, we are in for as bad or potentially even worse of an outcome than before. So, careful management here is key.

Eric Ries: Where do you think we go from here? How do we get out of the crisis?

Max Henderson: That is a great question. Nobody has a crystal ball, but my view here is that... So, containment actually worked a lot better than we originally anticipated. There was a question at the very beginning when we were building these models, "Well, what is social distancing really going to buy us? Is it going to slow the disease, but it'll stay growing exponentially? Is it going to flatten? Or are we actually going to be able to make the disease shrink just by social distancing?" And the reality is that... At least our team was surprised.

We essentially set up two confidence intervals for our model, best case and worst case. And we ended up basically right on the button for the best-case scenario. So, from our perspective, this lockdown, as extreme as it was, actually has been much more effective than it could have been based on the data we had when we started. And that, to me, is really encouraging.

So, I think it's completely achievable for some combination of policies, lightweight social distancing policies, policies like masks for all that are really focused on preventing transmission in close quarters, and a combination of testing and tracing to allow us to bring the economy back online while at the same time slowing or completely stopping disease spread. I think it is possible.

The thing that scares me the most is that rather than approach this the way that we have past crises as an American people to say like, "We can do anything. We're going to figure this out, right? There's no challenge too big," it seems like we're copping out to some degree and saying like, "Oh, well, that seems unachievable. We can't do that, so we're not even going to try," right? And so, that is the thing that gives me the most pause through all this.

The way that we choose to frame these problems are going to have such a huge impact on what we end up doing. So, we can think of this as unemployment and people losing their jobs and the economy collapsing or we can think about this as a national effort where these people are heroes for slowing the spread of a deadly pathogen and we're going to treat them as such, right?

Eric Ries: Like a 21st century WPA?

Max Henderson: Yeah, exactly. This can either be like an aspirational, "We are coming together," or it can be, "Oh, my God, things are falling apart." And the reality is that that framing is super, super important. Similarly, workers on the frontlines, they can either be people who are unfortunate enough to have to return to work or they can be the heroes going to battle for us that we take care of with free health care and free testing and hazard pay. We can do this, right?

Eric Ries: Protective equipment.

Max Henderson: Right, exactly. That framing is super, super important. Similarly, any national challenge is either an opportunity to come together and rise to the occasion or a cause for despair. And I think it is important for us to realize that we have control over these narratives and we have control over these outcomes.

Eric Ries: Max, thanks for taking time to share the story and thank you for your work as an antidote to despair in these really challenging times. It's meant a lot to me, and I know to millions of others.

Max Henderson: My pleasure, Eric. Thank you for having me. It's been great.

Eric Ries: This has been Out of the Crisis. I'm Eric Ries. Out of the Crisis is produced by Ben Ehrlich, edited by Jacob Tender and Shawn Maguire, music composed and performed by Cody Martin, hosted by Breaker. For more information on the COVID-19 crisis and ways you can help, visit helpwithcovid.com. If you are working on a project related to the pandemic, please reach out to me on Twitter. I'm @E-R-I-C-R-I-E-S. Thanks for listening.

blog comments powered by Disqus