Carl T. Bergstrom Profile picture
#BlackLivesMatter Prof. Biol. @UW. Info flow in bio, society, & science. I love crows and ravens. he/him Book *Calling Bullshit* Aug 4th:
a king on a throne with his eyes torn out Profile picture Zen Bonobo Profile picture eDo Profile picture Antonio Ripa Profile picture ☀️💧Leon-Gerard Vandenberg 🇨🇦🇦🇺 Profile picture 72 added to My Authors
8 Aug
I want to comment briefly on why I find this particular website so infuriating and why I am confident in calling it disinformation.

Dr. Fauci and many other health officials have stressed the importance of randomized controlled trials as the gold standard.
This very slick website presents what appears to be a research report using the language of randomly controlled trials, even talking about "assigning" countries to the control group or the treatment group.
We don't say randomized controlled trial, defenders of the website protest. We say country-randomized controlled trial, and that's a different thing.

How many members of the public will know the difference? This is designed to confuse.
Read 7 tweets
7 Aug
This is an infuriating new frontier in #COVID19 #disinformation.

Here we a professional-looking website claiming to present results from a randomized controlled trial of hydrochloroquine.

No authors or affiliations are listed, just a twitter account.
The report itself is a case study in Humpty-Dumpty's "When I use a word, it means just what I choose it to mean—neither more nor less."

The fact that different countries used different treatment approaches is referred to as a randomized controlled trial. Image
I try hard to follow the advice we give in Calling Bullshit of not ascribing to malice what can be explained incompetence, but when the authors co-opt the language of randomized controlled trials and pretend countries were assigned different treatments, my self-control shatters. Image
Read 11 tweets
6 Aug
1. As you may know, I'm a strong advocate of COVID testing. For the past few months, I've been collaborating with the people at @Color Genomics to understand how workplace testing could reduce outbreak sizes.

Out today, a blog post about what we found.…
2. Our results are consistent with other work on proactive testing, such as the paper by @DanLarremore, @MichaelMina, and colleagues….

Frequent proactive testing helps, and the turnaround time of test results is critical.
3. One of the things that a stochastic network model highlights is the bimodal distribution of outbreak sizes from a single introduction. Even with R0=2.0 or higher, many fizzle immediately.
Read 8 tweets
4 Aug
Enough book bullshit; back to COVID bullshit.

There's an interesting quotation from John Ioannidis at @CNN.…
I've commented elsewhere on how epidemiology is the truly dismal science. Whenever you discover something new that would be good news all else equal, it's not as good of news as you think because now not all else is equal.
For example, if you discover that the infectious period is shorter than you thought, that might sound like good news. It would be, ex ante. But ex post, holding the epidemic trajectory constant, it means that the disease is more transmissible per contact event than you expected.
Read 8 tweets
1 Aug
One of the impediments to frequent COVID testing is the discomfort associated with the NP swabs that go deep into the nasal passages. Some find it mildly uncomfortable, others dislike it more strongly.

It doesn't have to be that way. This isn't necessary for diagnosing COVID.
A new study in the NEJM demonstrates this. Self-collected swabs (tongue, nasal, and mid-turbinate) perform very nearly as well as professionally administered NP swabs and may actually have even higher viral titres.…

(h/t @bhrenton)
I don't have a full understanding of what the hold-up has been, though I assume it largely involves regulatory approval.

Research labs have had great results with AN swabs (basically a q-tip in the end of your nose) and saliva-based tests.

I don't know about FDA authorization.
Read 4 tweets
29 Jul
It was more or less guaranteed that some colleges and universities would use the CDC's irresponsible higher education guidelines as an excuse not to conduct re-entry testing.

I've written about this elsewhere:…
The @UNC has announced that they will not be conducting re-entry testing—and indeed they use the CDC recommendations as justification. (h/t @bhrenton)
What astonishes me is the level of support among the medical and public health faculty for this plan. I've written about this as well (thread).

Read 6 tweets
28 Jul
Today has seen a remarkable push of pro-hydroxychloroquine propaganda. The America's Frontline Doctors video on Breitbart. Trump retweeting a claim that it works, now removed by Twitter. Massive social media action. Rudy (below).

I don't understand it, but it feels coordinated.
Every single issue around COVID has become politicized.

Fascinating how arbitrary the political alignment is in some cases.

Back in 2019 you might have been able to predict some things: GOP wouldn't want to limit church services and Democrats would want free health care....
But I think you'd have been very hard pressed to predict which side would be pro-hydrochloroquine or even pro-mask.

For me this severe politicization of the entirely arbitrary highlights the ingroup-outgroup dynamics of the entire epidemic.
Read 4 tweets
26 Jul
It's late July. SARS-CoV-2 has been in the USA for over six months. In the past week, the FDA issued the first two emergency use authorizations for pooled testing: for pools of four w/ symptoms, and pools of five without.….…
Given the logistical overhead associated with pooling, this is next to useless.

The Quest EUA is only for suspected cases. At least the LabCorp is for proactive testing of asymptomatics.

But what are we doing? Screening can be low-sensitivity as long as it's high throughput.
As I've written elsewhere, test results are taking 5-10 days nationwide. Companies that want to enter with new testing technology get hung up for months in the FDA approval process—and this is just the rapid emergency use process!
Read 6 tweets
24 Jul
1. I am optimistic about the prospects for an effective Covid vaccine within a reasonable timeframe.

Through phase 2 of numerous trials, nothing has gone spectacularly wrong.

(I've written about a caveat here: .)
2. BUT an effective vaccine in 2021 is not an absolutely sure thing.

And we're not really talking about this all that much.

We're thinking about how to get through the next six-to-twelve months until we can be rescued by an immunological deux ex machina.
3. I think this is a mistake.

*If* we fail to develop an effective vaccine in a timely manner, this will have a catastrophic effect on the prospect of economic recovery.

This @NBCNews story begins to explore the issue.…
Read 10 tweets
22 Jul
A very short thread on the concept of pooled / batch / group / Dorfman testing.

Suppose that you have an assay, say, an RT-PCR test for COVID, and you have some number n of people you want to test.

Normally it will take you n tests to do that.
But if the disease is uncommon, you can arrange people into groups and pool their samples. Most groups will test negative, clearing everyone in the group with a single test.

If a group tests positive, you test stored samples from each individual. This uses fewer tests.
Going back and re-testing can be an undesirable logistical hassle. Double-pooling designs can eliminate the need for this. They can also improve efficiency in other ways.

A bit of theory:
Read 4 tweets
21 Jul
1. I've written elsewhere about three different roles for COVID testing:

But whether you are testing for individual health, surveillance, or mitigation, speed is of the essence.
2. When testing for individual health, every person with symptoms deserves to be able to be tested in a timely fashion, and to receive results promptly. Any functioning health care system would offer its members as much.
3. When testing for surveillance, we want to know what is happening now, not what was happening last week or the week before. Every day's delay in returning results is a day that we lose in being able to respond to changing prevalences of disease.
Read 15 tweets
19 Jul
As we struggle with the difficult decisions about reopening K-12 schools, one thing I've been unsure about is the degree of transmission by children.

A new study from Korea provides high-quality data about this question based on a large sample size.…
It is not good news, by and large. Older children (10-19) appear to spread disease at rates comparable to adults, though younger children may be less likely to transmit.

The @Nytimes provides context.…
This is not wholly unexpected—I was never persuaded by the small studies suggesting reduced transmission in this age group–but it's not going to help one bit as we try to keep the pandemic under control while returning to some sense of normalcy.
Read 18 tweets
15 Jul
Imagine having such a severe case of Dunning-Kruger that you write something like this in @USAToday and attempt back it up with a mere 269 words that would constitute a low-quality effort even for your average Covid-denialist twitter troll.
I get it that the White House would like to discredit Tony Fauci.

But my goodness they're really not sending their best people. Or if they are....yikes.
There's also a question of journalistic ethics.

I understand that the OpEd page follows different rules, but when should a newspaper allow a government figure to publish blatant misinformation around a national crisis?
Read 5 tweets
14 Jul
This is literally what the CDC is intended to do.…
But the CDC has been cut entirely out of the loop.
The idea of bringing in the National Guard to do the CDC's job is bizarre but was floated in an earlier draft letter:
Read 5 tweets
13 Jul
1. This thread about an important paper from @jonassjuul, K. Græsbøll, L. Christiansen, and @suneman will be more technical than usual.

The paper addresses how to depict ranges of outcomes in simulations of COVID outbreaks (or other stochastic processes).
2. As such it is of critical importance to people modeling the pandemic. But it's also important for anyone who wants to read and interpret models that others have created, because it warns against a trap that would be easy to fall into.
3. Disease outbreaks are fundamentally stochastic processes. The same disease, introduced into the same population, might infect a large number of people one time, and disappear quickly another time, based on the luck of the draw.
Read 17 tweets
10 Jul
I don't want to succumb to the temptation of hyperbole, but watching the Trump administration's response to the COVID pandemic I can't help but think about the Three Years of Famine in China from 1959-1961. I'm far from expert. My understand is based largely on one college class.
That said, my sense of what happened is that Chairman Mao became enamored of some crackpot agricultural ideas from Trofim Lysenko (yes, that Lysenko) and Terentiy Maltsev.

These involved high-density planting and deep ploughing, both of which led to crop failures.
At the same time, the central government mounted an aggressive Smash Sparrows Campaign. Millions of birds were killed as supposed agricultural pests, leading to an explosion of the far more harmful insect pests that they would have fed upon.
Read 9 tweets
10 Jul
Sixteen hours ago the South China Morning Post published a story about what is purported to be a new, deadlier-than-COVID pneumonia in Kazakhstan. The source seems to be a statement from the Chinese embassy there. Details are sparse and skepticism merited.…
CNN has now picked up the story is essentially repeating it without any further information. There is nothing about how COVID has been ruled out, nor much by way of epidemiological detail. Until we know a lot more, COVID seems the most likely explanation.…
One report is that this is based on a translation error. Seems more likely to me than a new, even worse pneumonia.

We've got known dangers to worry about right now. I'm not going to worry about unlikely ones until we've got a lot more information.

Read 6 tweets
10 Jul
I do stupid shit in my dreams.

Last night I dreamed I was trying to steal a pizza from a sleeping grizzly bear. The bear woke, and very kindly asked for the pizza back. (It talked.) I should have just returned the pizza but instead I cursed the bear out and insulted its mother.
The bear started moving toward me, and I should have run. But instead I thought maybe I could still steal the pizza if I played my cards right. To make a distraction, I started smashing all the furniture in its bear house. The bear was pissed and started to growl as it advanced.
At this point I was in big trouble. I should have set down the pizza and begged for my life. But no. The magical thinking kicked in and I decided to up the ante. I opened the box and threw the hot pizza right in the bear's face. The bear swatted the pizza away with a huge paw.
Read 5 tweets
8 Jul
It is heartbreaking to watch @CDCDirector Robert Redfield turn a great public health institution @CDCgov into a political propaganda tool. I feel dreadful for the thousands of talented, devoted professionals under his direction.

Today President Trump tweeted this:
In response, @CDCDirector Redfield undermined his own agency's position, promising to go back and issue new, politically expedient recommendations.

This is the exact opposite of public health leadership.…
The would be craven cowardice from any agency director, but from the @CDCDirector with thousands of lives at stake, this obsequiousness servility is criminal.

This is the same man who lectured us on personal responsibility three days ago.

Read 25 tweets
6 Jul
I don't want this pandemic to be a disaster. I don't want to see hundreds of thousands lives lost in the USA. I don't want to see hospitals overrun in a second wave. I don't want to see businesses closed, livelihoods lost. I don't want any of this. I'm not rooting for the virus.
This is not a career opportunity for me. In my line of work, no one cares how many twitter followers I have. I haven't submitted a single paper (the currency of our realm) in 2020—I've been too busy doing public communication and response planning around COVID.
Read 6 tweets
5 Jul
@CDCDirector I wholeheartedly agree regarding masks.

But while we're taking about personality responsibility, @CDCDirector, some of us have more personal authority than others.

I have a few suggestions about how you can take personality responsibility for the agency you direct.
@CDCDirector 1) You deleted this life-saving language from the @CDC guidelines because of pressure from the White House.

Restore it.…
@CDCDirector @cdc 2) Your agency provided COVID modeling parameters to federal agencies and scientific researchers. You set the IFR at less than half consensus estimates, and as of yesterday refused to explain why while stonewalling FOIA requests.

Explain yourself.…
Read 4 tweets