Read on Twitter
Did you know that *study after study* finds most people don't seem to transmit COVID at all? That a small percent is responsible for almost all infections? That R is not that informative? My new piece on why this may be key to controlling the pandemic. https://www.theatlantic.com/health/archive/2020/09/k-overlooked-variable-driving-pandemic/616548/
COVID has a very skewed distribution in its transmission patterns. You can't beat back a threat if you don't take into account how it operates, and this skew—its overdispersion— should affect the way we do *everything* from tracing to testing. That's what countries like Japan do.
Folks asking me if this skew—that a few account for most infections while most don't transmit—is due to the people & what it means. That's all in the article.
It's not just a mathematical curiosity; it has big practical consequences. Also in the article. https://www.theatlantic.com/health/archive/2020/09/k-overlooked-variable-driving-pandemic/616548/
It's not just a mathematical curiosity; it has big practical consequences. Also in the article. https://www.theatlantic.com/health/archive/2020/09/k-overlooked-variable-driving-pandemic/616548/
Wow. New paper from India, tracing ~500K contacts for ~85,000 cases. They did not find a single positive contact for 70.7% of infected people. Zero! About 5% accounted for 80% of transmission—not in my piece because it's out today! That's overdispersion. https://science.sciencemag.org/content/early/2020/09/29/science.abd7672.full
Look, the science doesn't fit into the lockdown/anti-lockdown fights. It's not how much we restrict, but whether we target the right things. Overdispersion means the right measures take us a long way, but also that we can't relax: It can quickly flip back. https://www.theatlantic.com/health/archive/2020/09/k-overlooked-variable-driving-pandemic/616548/
See this thread for some great early pieces (some I link to in my piece) explaining the importance of superspreading and overdispersion, and if you want, also take a look at Japan or Hong Kong early epi documents. We're so slow to learn from the evidence. https://twitter.com/foxjust/status/1311352169896071169
Also, if you suspect that I went to the trouble of writing ~5000 words on overdispersion so I could type "the mean is not the message", I admit nothing. https://www.theatlantic.com/health/archive/2020/09/k-overlooked-variable-driving-pandemic/616548/
Described as dense but incredibly interesting=life goal
https://twitter.com/Justin_Ling/status/1311336869968445442
https://twitter.com/Justin_Ling/status/1311336869968445442
Then again, the number of people who assume that I just "discovered" skewed distributions or the Pareto principle this month has been pretty impressive, too. Always makes me wonder if men get the article they just wrote explained back to them this much?
https://twitter.com/RobotArchie/status/1311456542047326214?
https://twitter.com/RobotArchie/status/1311456542047326214?
This example is exactly why we should do more backward tracing for a superspreading pandemic. It means, mathematically, “who infected you” can be a more productive question then “who did you contact since becoming infected”—forward tracing. https://www.theatlantic.com/health/archive/2020/09/k-overlooked-variable-driving-pandemic/616548/ ht @reillyhay https://twitter.com/kj_seung/status/1311443956656492548
The why of superspreading isn't fully settled. I do have more about it in my article on the importance of ventilation from July. Crowded, poorly-ventilated indoors is a common factor, though individual variation is also plausible. https://www.theatlantic.com/health/archive/2020/07/why-arent-we-talking-more-about-airborne-transmission/614737/ https://twitter.com/christateston/status/1311653698490294272
I'm getting many questions on superspreading. My July article on ventilation has details on what we know of the science, but there *is* a tl;dr: Let's focus on Japan's 3C: avoid closed (indoor) spaces with crowds, especially at close range and/or unmasked. https://www.theatlantic.com/health/archive/2020/07/why-arent-we-talking-more-about-airborne-transmission/614737/
Much gratitude to @svscarpino, @AdamJKucharski, @mugecevik, @dylanhmorris & @nataliexdean who are quoted and to many others whose papers I read, and to Dr. Hitoshi Oshitani who patiently answers my Japan questions and to @miekocakes who has been a tireless translator & bridge!
Also, my sense from years of writing on the messy, complicated topic of technology and society, and now about the pandemic science/sociology intersection: There is a real demand for factual, dense and long pieces that don't simplify or cut corners, but do try to stay readable.
It's not that I or anyone else can get everything right all the time (nobody can; anyone thinking out loud on complex topics will miss things or get things wrong) but that we can and we should try to dive into it all, with all the complexity, and treat the public as partners. 
Yes! As I explain in my piece, the polarization around Sweden is not useful because Sweden is a middling country—not that lax but also not great. It's not an example of success OR failure of herd immunity or lockdowns. Japan, though, is *very* interesting. https://twitter.com/chip_uni/status/1311727369086922757
As explained in my overdispersion piece, if there are two transmission events (both the President and the First Lady) there are likely more. Also see my earlier piece about airborne transmission: indoors, six feet isn't enough and the *talker* can infect. https://www.theatlantic.com/health/archive/2020/07/why-arent-we-talking-more-about-airborne-transmission/614737/
This pathogen can be surprisingly non-contagious most of the time... But when it strikes, it's often in clusters. So they may well have a cluster on their hands. https://twitter.com/nytimes/status/1312024557356806145
The cost of not following the science. The CDC still has not (re)posted updated airborne transmission guidelines, and WHO still has not updated their documents for the public. DISTANCE IS NOT PROTECTIVE ENOUGH INDOORS. Ventilate, filter, mask up. https://twitter.com/jonlevyBU/status/1311976244553101313
Yes. The authorities should never have let anyone—no matter their position—tour hospitals unmasked in violation of their own rules. The debate commission should have stood its ground. The WHO and the CDC should have listened to the scientists. So it goes. https://twitter.com/spin4srv/status/1312027812551970818
Yes, the Trump campaign was flouting the guidelines. But the current guidelines we do have from the CDC and the WHO are not enough to protect Biden. Such failure puts at risk people who do follow the rules. Indoors, distance alone IS NOT enough to prevent airborne transmission.
It looks increasingly like a cluster. The key question will be if they all got infected from the same source/event. With all this, it would be irresponsible not to quarantine almost everyone in the White House and the Senate. Tests can need days to detect. https://twitter.com/SenMikeLee/status/1312055257992753152
No word on ventilation, no mention that distance isn't protective enough indoors, and no enforcing of masks. This is why the lack of CDC guidelines on airborne transmission matter. WHO docs still say no masks necessary indoors if separated by three feet! https://twitter.com/HenryJGomez/status/1312058477758603265
Indoors. Unmasked. Talking. https://twitter.com/senmikelee/status/1311064032300929038
Cluster looking more and more likely. https://twitter.com/jackmjenkins/status/1312068942157025280
Cluster. K. Overdispersion. Airborne. Backward tracing. (Words and concepts we need the talk more about.) https://twitter.com/brianstelter/status/1312079653486632961
Yes, finally! For details on why this is important, see my article on aerosol/airborne transmission. https://amp.theatlantic.com/amp/article/614737/ https://twitter.com/demfromct/status/1312131078157524992
Cluster. It was bound to happen. https://twitter.com/costareports/status/1312085226370334720
A superspreading event. They have to trace this back asap. https://twitter.com/KellyannePolls/status/1312214949658152960
There was an indoor component to the event. But we definitely need more information. More in the cluster. https://twitter.com/howardkurtz/status/1312200574067068930
One more senator. This is how a cluster works, and these people are tested heavily. Imagine this cluster without access to testing, igniting more clusters. That’s the catastrophic process my article has been trying to explain, and the one we should gear up to avoid. https://twitter.com/edokeefe/status/1312378700818722822
