Thread by @WhitneyEpi

Reading this article (in which I'm quoted near the end), I realized why I have such a strong visceral reaction to some of the negative coverage of @ProfEmilyOster's work on #COVID19 and schools...prospect.org/coronavirus/wh… 1/

..The treatment of @ProfEmilyOster reminds of 2016 press coverage of @HillaryClinton! An outsized focus on that woman's flaws had big long-term consequences.(I'm not a media studies or communications person, so I may be totally off, but I swear there's a PhD thesis in here! 2/

On to my promised tweets on #epitwitter data about #SARSCoV2/#COVID19 among children and particularly in school settings. It's less comprehensive than I hoped, but I'm trying not to let the perfect be the enemy of the good... 3/

First, I disagree with the premise of many articles that we don't know enough to start making some conclusions about schools and risk. In particular, @DFTBubbles updates its amazing comprehensive literature syntheses on a regular basis 4/

The studies with the strongest scientific designs are consistent in a few respects: young kids (<10 yo) seem less likely than older children & adults to get & transmit #SARSCoV2 even in settings of close contact like households. It's not zero risk, but much less than adults 5/

This age difference is important for thinking about public health policy -- and the transmission risks we expect in a given room of 5 year olds vs. a room of 17 yos vs. a bar full of 20-somethings 6/

One of the critiques of this literature is that much from non-US settings with lower case counts. This reminds me of the reverse exceptionalism where folks in US thought, "China & Italy can't happen here." Or rural folks thinking, "NYC won't happen here"... 7/

Just because data from a different setting doesn't mean it has no application. As epidemiologists, we're trained to identify critical biases to decide what applies & what doesn't. We may disagree, but I take well-done studies as evidence even if from a different setting 8/

Second, I'm not here to tell you the science is totally clear because it's not. But most results in epidemiology not definitive. People criticized @ProfEmilyOster's prospective cohort study of schools for being non-representative and representing only a fraction of schools...9/

Look, I'm a social epidemiologist researching the #gynhealth of racially, socioeconomically, & gender-minority marginalized populations. I care A LOT about who's left out of #epitwitter data, but idea that we can't draw conclusions from unrepresentative cohort studies is wild 10/

Our field would collapse if that were true! I criticize overreliance on cohort studies that draw from non-representative samples of select volunteer populations, but that data also tell us important things. It can suffer from selection bias, limits to generalizability, etc..11/

But prospective studies are also super important for avoiding recall bias, some measurement issues, getting incorrect rates bc of missing denominators. Bc much #SARSCoV2 surveillance limited to case counts, prospective studies are vitally important for triangulation! 12/

Another critique is reliance on voluntary testing and reporting. Here is where the bias of my personal experience comes in... I have two kids in daycare and many friends with kids in daycare. In my friend group, kids in daycare are among the *MOST* tested people...13/

Bc #COVID19 symptoms in kids are so non-specific & similar to other illnesses, every slight fever or runny nose could mean being excluded from daycare for 14 days. Almost every daycare kid in my small circle has gotten a #SARSCoV2 PCR test to show daycare that COVID neg 14/

This is anecdata that is probably not true in many other settings, but it does make me less worried than some others about rampant outcome misclassification and missing hidden outbreaks in @ProfEmilyOster's data set (this is my non-scientific leaning) 15/

Third, I get frustrated with the whole framing of the the #COVID19 schools debate. Here, I wrote about one set of issues that I have with the framing of the conversation

https://twitter.com/WhitneyEpi/status/1320164073812692993?s=20

16/

But another framing issue with the #SARSCoV2 schools science is that it's framed in terms of a very limited menus of options. One critique of @ProfEmilyOster's study is that low transmission occurring in well-resourced private schools taking lots of precautions... 17/

But to me, that's kind of the point! With enough resources and precautions, there's evidence that in-person schooling can be safe. That's important! That's the standard we should be demanding is met for ALL our students & teachers. And only way teachers will feel safe 18/

I know this is easy for me to say in my college town in a state with a Democratic governor and @ncdhhs leader who taking #COVID19 seriously. I'm not in the trenches with folks with FL, AL, UT, TX, ND right now. But we CAN demand more and better resources. 19/

Finally, I first started following @ProfEmilyOster's #COVID19 work not because of her school cohort but because of this post from her blog: emilyoster.substack.com/p/grandparents… ...20/

This post continues to be the clearest articulation of how I think we should approach #SARSCov2 questions about in-person schooling: What are we considering & what's the alternative? If we are delaying decisionmaking until X, will anything be fundamentally different at time X?21/

Since spring, it's been clear that the #COVID19 pandemic would be 1-2 years, even under =best of scenarios. So my question has been, Will you focus on finding a safe way to have in-person schooling, or will we have schools closed for 1-2 years? 22/

https://twitter.com/WhitneyEpi/status/1252571781413703682?s=20

And, if we choose longterm closure, are we preparing for the long-term consequences? I don't feel that I see those conversations happening. If this is our course, it's imperative that we as a country start preparing now for that long-term fall out of this course. 23/