I Collaborated With An Epidemiologist!
Added 2020-03-25 20:39:57 +0000 UTCIf you need to reduce or cancel your pledge, I understand. Lots of folks are out of work right now. Don't worry about me – I have enough in savings, thanks to your support over the years. <3 See you on the other side of this pandemic!
. . .
(reading time: 6 minutes)
What happens after the lockdown?
Even if you haven't heard of the March 16th Imperial College Report [PDF], you may have heard of its pessimistic findings. To recap, for the US/UK:
1) Merely "mitigating" the curve still means exceeding hospital capacity by thirty-fold
2) So, we need to suppress the curve with a lockdown
3) But, if we let up before a vaccine is available (~18 months out), even a single domestic/imported case means a massive outbreak again. A lockdown isn't a cure, it's just a restart.
So, as far as the report analyzed, our only options are:
- Let millions die
- Lockdown for most of 18 months
Well, crap.
Unless... there's a policy the report didn't analyze... and that several East Asian countries are already using to successfully contain COVID-19 while keeping normal life going... and that the rest of us can do after we use the lockdown to get our "restart"!
Is there? There is!
It's called "testing & tracing".
I collaborated with Marcel Salathé, a Swiss epidemiologist who's been cited by 5,000+ papers, on visualizations to explain "pre-symptomatic transmission" and why "contact tracing" is needed to stay #OneStepAhead of COVID-19.
👉 Please read+share Marcel's Twitter thread, with my visualizations!
👉 (also here's my Twitter thread)
Here's just the visualizations:
📣 Please let me know what you honestly think – what concerns/questions do you have that we didn't fully address? We'll be making a video next, so we'd like it to include your concerns.
P.S:
(Looking at the Twitter replies, the big concern is "it's too late for US/Europe to copy East Asia". I think for the video we'll emphasize that the lockdown is "just" a restart, but a restart will let us copy East Asia. Though I'll admit I don't know exactly how many cases/tests-per-capita we'd need first. Will ask Marcel when our time zones overlap.)
P.P.S:
(Disappointingly, I didn't see much concern over privacy. But since I care about privacy, let me emphasize: we can make a privacy-preserving contact tracing app. Singapore already made one and open-sourced it, so that folks can 1) verify it indeed protects privacy, and 2) the world can get on contact tracing apps, ASAP)
. . .
To avoid unwittingly spreading mis-info, I waited to make a COVID-19 thing until I could collaborate with a real epidemiologist. But now, I want to say something personal:
For the first time in days, I feel hope.
Knowing that multiple highly-cited, infectious-disease epidemiologists – Marcel Salathé, Marc Lipsitch, Trevor Bedford – support the "restart with a 1–2 month lockdown then copy East Asia" plan... that makes me 90% confident this isn't just wishful thinking.
Because: an 18 month lockdown isn't just a hit to folks' financial stability. It's a hit to our mental health. Social isolation is one of the big risk factors for major depression, and the elderly are already the most at-risk for suicide. I don't think they could take 18 months of this. I don't think I could take 18 months of this.
So, I'm heartened there's a science-based alternative – already used by South Korea/Singapore/etc – that's not "millions dead" or "1½ years of isolation".
But:
Most people haven't heard of this idea yet.
From what Marcel told me, most policymakers don't understand this idea yet. (They didn't even understand exponential growth until Italy!?)
This is a problem, because a few weeks into a lockdown, new confirmed cases will drop dramatically. That's good! But policymakers & the public may get cocky, remove the lockdown – thinking it's over when it's just a restart – and cases will spike again.
UNLESS: we start talking about our strategy after the lockdown.
So, please start talking about staying #OneStepAhead, with testing & tracing! (I know it's tacky to ask for shares, so share Marcel's thread instead of mine.)
Lockdowns are "just" a restart, but a fresh start's what we need. While we reboot, let's build loads of easy-access tests & have folks download privacy-preserving contact tracing apps –– and we do have the resources to do this –– in order to do what we should've from the start.
I hope this gave you hope. Determination. The more intensely we do the reboot now, the quicker we can switch to East Asia's "test & trace", and life can get back to sorta-normal.
In Singapore: kids still go to school, malls still hawk their wares, couples still dine out at bougie restaurants. Sure, the restaurants are weirdly quiet, everyone's wearing masks, and international travel is restricted. But life, mostly, goes on.
That's what's keeping me sane. Knowing that – with smart strategy & regular folk working together – in 2 months I can go out to a café, meet a local queer cutie, and get hot chocolate with them.
That's worth hanging on for.
“He who has a why to live for can bear almost any how.”
~ some German dude
Let me know your honest feedback re: the visualizations in the comments!
Stay safe & sane, folks. 💖
. . .
P.S: I made a song adaptation of the WHO handwashing guidelines! 🎵
P.P.S: For the sake of intellectual honesty, I want to note 2 things I was wrong about in my last post: 1) the "flatten the curve" message is mathematically incorrect, and 2) face masks are way more complicated than I thought.
P.P.S: One day after I tweeted a GIF of a COVID-19 sim I was making, Washington Post published an amazing COVID-19 sim. It's the most-shared WashPo article of all time, even tweeted by Obama. I am sliiiiightly jealous. But the author's a fan of the explorables I've made, funded by you folks, so I guess we can take some pride in that.
Comments
Thanks Voyage! Privacy & our rights is a damn good concern, and one I care deeply about – and that's why I'm glad the group I'm working with asked me to explain how we can do contact tracing that's *mathematically, cryptographically* guaranteed to preserve privacy! Like even if the govt or a hacker group seized the server & stole all the data, *no privacy would be lost*. (My biggest concern now is if people *don't know about that*, it'll be the 2 scenarios you just described: Covid-19 remains uncontrolled, or the surveillance state gets bigger. It doesn't have to be *either*.) Patreon post with an explanation-comic very soon!
Nicky Case
2020-04-10 20:05:43 +0000 UTC* knowing that you might finally have a strategy for after the lockdown... * it fills you with DETERMINATION. More seriously though... I'm afraid of various possible misuses of tracking basically anyone, be it during or after the lockdown. I mean, maybe I'm too pessimistic about the human race, but I'm afraid that we have two global scenarios: either a nation CONTROLS its inhabitants with an Orwellian surveillance state, or humans don't take the application seriously and can't auto-regulate and continue to spread the virus. I'm afraid of a case where there would be no middle ground because of a lack of self-discipline of most humans... (great figures and great links though, it's very clear and definitely one step ahead! In the end I'm more worried about, sociopsychology or something? The feelings that order regulated with a tracking app WILL cause abuses, and that otherwise people just love taking risks for subjectively good reasons but that would objectively put others or themselves in danger...)
Voyage Goya
2020-04-02 23:07:56 +0000 UTCHey Nicky, I'm worried that the visualization only covers one of the two cases where there's a false negative on an individual carrying the disease. The visualization covered the asymptomatic carrier. However, a situation where someone is unaware they were in contact with someone showing symptoms seems likely (e.g. the president being told by reporters that he was hanging out with someone who tested positive). Access to test remains limited in most of the US (including hot spots for the outbreak like the Bay Area). Many people who are symptomatic are unlikely to know. Similarly, information about whether people were at events with those who have been positively tested continues to be limited. So, I'm wondering why your recommendation is suggesting people wait to know if someone is sick before quarantining rather than suggesting people self-quarantine regardless of if they have explicit confirmation of being in contact with the disease. Either way, the visual is easy for me to read & I appreciate you sharing something at this time :)
Michael Merchant
2020-03-28 20:34:45 +0000 UTCOne other random feedback is the CC0 licensing. My default is CC-BY-SA-NC, but if you're hoping that lots papers and other Social Media outlets will share the images, then CC-BY ensures that they actually include your name and the doctor's name. Likewise, I would add your Twitter handle or Patreon link to the images
Gaëtan Perrault
2020-03-27 06:00:25 +0000 UTCHey Nicky, some notes on details before these go viral on Social Media. - Image 1 has a lot of text. You have 5 visual sections. Sections 3,4,5 all have really long complex sentences. It feels like you are writing video voiceover instead of infographic text. - You have "no symptoms yet", but maybe just "no symptoms". That way you cover asymptomatic carriers and also the notion that some people just have the virus. - I think picture 3 is just better than picture 2, though it may be less Social Media friendly. You may want to give them both headers such that people could rationally share 2 OR 3 because they basically have the same effect.
Gaëtan Perrault
2020-03-27 05:55:15 +0000 UTCHey Nikki, I really like your graphics. I agree with Ross that these graphics are wrong in a subtle but important way wrt asymptomatic carriers. But I also understand the raw value of this visualization and how much good it does. Getting a visualization like this that includes asymptomatic carriers would be far more complex and would likely only appeal to nerds like you, me, Ross and SImon. People who probably don't need any convincing to stay home anyways. Ultimately, with some tweaks from other commenters, I think this is worth sharing. If people want to quibble about details you just point them at the research papers or wherever people are having these nerdy discussions.
Gaëtan Perrault
2020-03-27 05:42:41 +0000 UTCOnly if we get enough testing! https://www.linkedin.com/posts/ajamesphillips_corona-qiagen-activity-6648943480796172288-w0P7
AP
2020-03-27 01:33:17 +0000 UTCYou're right. If everyone could isolate themselves in small groups for say 3 to 4 weeks. The whole thing would be over in a month. But most people don't have food for that long, they might have other medical needs. Some people need to main power stations, plant fields, repair homes etc etc. So the reality is you can't have a perfect quarantine. Ok so now we've got people moving around and mingling, spreading the disease still even after "locking down" it's not a full quarantine so what do you do? Well "aggressive" contact tracing... and pre-symptom targeted quarantine. The "One step ahead"
AP
2020-03-27 01:05:59 +0000 UTCHey Yves, I completely empathise. I would say though I think you already have the answer: "I feel like I have no control over what's really happening" in part true, because we can only do what we can do. But also I'd say if we can paint a better picture of the world, people will form up behind it and start pushing. But it has to be realistic. I'm going to message Nicky now to ask if I can help. I've cofounded HelpfulEngineering.org which now has 13,000+ volunteers and some of them have already deployed real solutions to care workers and others. We just have to keep pushing, keep trying (and making sure we get enough sleep and look after ourselves so we can look after others :) ). Regarding profit over lives at the administration level... the two are actually pretty closely linked so I'd be less concerned... but I have started firing shots at our inability to scale up testing. As Nicky pointed out a key part of the exit plan is testing. But we need a lot of testing and it looks like we'll still be well short come May (Qiagen said 333,000 per day by July... which is still *real* low vs what we need globally). Please check out, like and share this article to start pushing for scale up of testing reagents (this is the main bottle neck): https://www.linkedin.com/posts/ajamesphillips_corona-qiagen-activity-6648943480796172288-w0P7
AP
2020-03-27 00:41:22 +0000 UTCHi Leslie! Ack yes that's actually really good visual design feedback. For the next visualizations we make (and we'll make more!) I'll implement your feedback! (Like, maybe the "wrong" cut is more faded out, to show how it's behind? Dunno) Thanks! ^_^
Nicky Case
2020-03-26 19:01:31 +0000 UTCGood question! Most countries are doing isolating people who *are* showing symptoms, and some more are also isolating people who are *in the same house* as someone showing symptoms...but contact tracing is different in that it asks to self-isolate folks who were in close contact with an infected person *before* they showed symptoms! That's not what US/Europe is doing. Yes, it's a quarantine, but not "just" the usual quarantine – HOW/WHEN/WHO you quarantine is the big difference – especially since an estimated 45% of infections are caused by folks who *were not symptomatic* yet. (src: https://github.com/BDI-pathogens/covid-19_instant_tracing/blob/master/Manuscript%20-%20Modelling%20instantaneous%20digital%20contact%20tracing.pdf )
Nicky Case
2020-03-26 18:48:24 +0000 UTCbtw the figure in your second link comes from this article: https://medium.com/@andreasbackhausab/coronavirus-why-its-so-deadly-in-italy-c4200a15a7bf
Simon Morrow
2020-03-26 17:40:36 +0000 UTCIsn't that just a quarantine?
Rev Storm
2020-03-26 15:49:18 +0000 UTCI hope it's okay to be a little negative despite this instantly making me feel a lot more informed. I'm glad to hear all of this, and as usual your diagrams make me feel more confident in what I know, because it's logically expressed in a way I easily understand. (Oh God I keep hitting enter before I'm done. Whoops. Will edit this.) On the other hand... 18 months of social isolation. I didn't even know that that was being considered--at least not seriously; I'd heard people discussing it but much more often people talk about the "three months" it took East Asian countries to get this 'under control'. So I was kind of already stuck on the "please let this be over in a couple of months" train... And admittedly I feel depressed no matter what I see now. Our president (in the US) doesn't give half a crap for human lives. His lackeys care more about stocks and 'business' than the well-being of the most vulnerable among us. It's hard to feel like anything good can happen anymore, because nobody in our government will let it happen. And this while we have the Democratic primaries going on... it's just painful. I feel like nothing can really make me feel better about the situation, because it all seems so hopeless. What do you do when you feel that way? Is there any way to get past Trump and actually instate these policies? I mean, sure, I RT things on Twitter, but I feel like I have no control over what's really happening, and people are still being forced to work.. I do want to reiterate that this is a really good post and I'm always glad to hear from you! I am also so happy to see information backed by experts conveyed so clearly. So there's my one question; how do we know that the right thing to do will get done? I'm not sure if you can address that; any thoughts appreciated.
yves.
2020-03-26 07:04:35 +0000 UTCFor reference, I'm basing the 1/5 number off this Forbes article which links out to other studies https://www.forbes.com/sites/brucelee/2020/03/18/what-percentage-have-covid-19-coronavirus-but-do-not-know-it/#240b70c77e90 And sorry this is a Discord URL, but I found this comparison of the age distributions of Covid-19 cases in Italy, where they are only testing people who show symptoms, and S. Korea, which has broad testing. https://cdn.discordapp.com/attachments/386227514680410113/688588087428776042/ETFXGXeXYAUrxUv.png So I'm thinking a lot of the low numbers we see on symptomless cases right now might just be because of how locations are testing people? I'm finding conflicting studies, but the conflicts all seem to trend higher the more broadly testing is done.
Ross James
2020-03-26 06:02:37 +0000 UTCThank you for fighting the good fight 💖
Brittany B.
2020-03-25 23:54:27 +0000 UTCNicky, Hello! The pics are very clear and easy to understand. Here are a couple questions it leaves me with: 1. How effectively can we really identify and contain the “close contact” people? If we’re relying completely on an app to do it, does that mean anyone who can’t afford their own smartphone is SOL? 2. How many people get sick just from touching something a sick person touched, even if they never got near the sick person? 3. In this system do they recommend that the people who are most likely to die if they get sick still self-isolate as much as possible? I know you probably don’t have all the answers, just some things to possibly ask Marcel. Also it’s really encouraging to see some solution-focused work when there’s so much out there about everything that’s going poorly. <3 Simon P.S. as much as I liked the WaPo sim, one thing it doesn’t do which your sims often to do is have some interactivity. imo it’s really powerful to turn the dials on the variables yourself and try every possible scenario that you’re curious about.
Simon Morrow
2020-03-25 23:01:18 +0000 UTCHey Nicki! I love how the visualization explains the concept. The only thing I'd suggest visually is to distinguish the two cut lines from each other. One means success, the other means failure. At first glance they appear equal and it's the text that does the work of explaining the difference. Maybe through color or iconography can be clear at a glance what the goal is. That's my two cents. Thank you for your hard work on translating the abstract to concrete :)
Leslie
2020-03-25 22:45:35 +0000 UTCThanks for your in-depth honest feedback, Ross! Those are all very good concerns, and I'll pass them over to the epidemiologist. (For "infectious but *never* shows symptoms", as far as I know right now, that's estimated to be relatively rare: about 5% of all infections caused by never-shows-symptoms people. Source, this (not yet peer-reviewed!) preprint paper from University of Oxford, see Figure 2: https://raw.githubusercontent.com/BDI-pathogens/covid-19_instant_tracing/master/Manuscript%20-%20Modelling%20instantaneous%20digital%20contact%20tracing.pdf TBH, 5% was way higher than I thought. But the paper analyzes that, even if we totally miss the infectious-but-no-symptoms people, a moderate uptake of an instant-notification contact tracing app could still contain a COVID-19 outbreak) > relying on individual people to take the initiative to self-isolate before they're sick when we can't guarantee sick leave protection for them Absolutely crucial concern, thank you. That's outside the scope of science & epidemiology, more politics. But yeah, my personal opinion, policymakers gotta subsidize/cover sick leave for even suspected COVID19 cases. Would be a lot less economically costly than bringing down the whole workforce or the healthcare system But yeah, given how messy policy's been recently, "just make policymakers do the thing" is a lot easier said than done :/
Nicky Case
2020-03-25 22:24:53 +0000 UTCHi Nicky! We are SO on the same page! I too was feeling hopeless until I came across the incredible article Coronavirus: The Hammer and the Dance - https://medium.com/@tomaspueyo/coronavirus-the-hammer-and-the-dance-be9337092b56 which explained that complete lockdown does not have to continue for months (which would be economically devastating) but that we can actually implement strategies like South Korea etc with contact tracing and aggressive testing. Keep at it! I was thinking of contacting you to make a visualization of it, and you are already ON it! The visualizations above are exactly what is needed :)
Adam Edell
2020-03-25 22:12:31 +0000 UTCA big problem I see with this is just how many symptomless infected there are of Covid - not just 'don't show symptoms yet', but go through the full two week contagious period without ever showing any. About 1/5 people who get it won't think to go into self-isolation until the people *they* infect start showing symptoms. Because of that, and the contagiousness of it, and the fact that we'd start relying on individual people to take the initiative to self-isolate before they're sick when we can't guarantee sick leave protection for them, really seems to underscore where the thread leaves off - how do you apply this technically? EDIT: Sorry, to clarify, this was really good and I enjoyed it a lot and the visualizations were great. I'm just looking at a Business Insider article one tab over telling me that McDonalds is refusing to give paid sick leave to employees
Ross James
2020-03-25 21:43:41 +0000 UTC
