A lot of people are misrepresenting/misunderstanding this Lancet systematic review into school closures for #COVID19 so I thought I go over it a little bit 1/npic.twitter.com/f1VlmcvYzu
You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
6/n What did they find? - 6 studies evaluating school closures for SARS - 2 modeling studies for SARS - 1 interview study for SARS - 5 preprints of school closures in China for COVID19 - 1 preprint of other human coronaviruses - 1 Imperial College modelpic.twitter.com/u8WealVXZD
7/n Confusingly this is both a very broad pool of evidence (lots of information) and also not very helpful in giving us evidence (limited in scope)
The challenge of causal inferences in epidemiology 
8/n For example, one preprint looked at Hong Kong, and found that school closures in combination with other measures reduced the R(eff) below 1
Great
Except, hard to know what role school closures themselves played
pic.twitter.com/kI1YL7ZS6z
9/n Some published work on SARS examined school closures and had mixed findings 

Maybe a benefit, but maybe not - hard to say!pic.twitter.com/3NdAF8A7yB
10/n Another study looked at coronaviruses during snow days
in Seattle in 2019 and found a reduction in other coronavirus infections of ~6%
Not shabby, but hard to compare to our current situationpic.twitter.com/uz0rP55mZM
11/n Which brings us to the modelling studies This is actually where things get much more complexpic.twitter.com/2woeBo1hix
12/n The problem with some modelling studies is that they aren't necessarily *evidence* in the traditional sense What you can find is that the information you are looking for is ASSUMED in the model, rather than demonstrated by it
13/n
This is actually exactly what we find when we look at the Imperial model
They ASSUMED that school closures would reduce transmission by a set amount, and then predicted how many deaths this would eliminate if that were truepic.twitter.com/GPPY5hvNet
14/n Other models similarly ASSUMED things about SARS and COVID-19, and similarly concluded that school closures might be effective if these things were true But again, these are not demonstrating efficacy, they are putting a number on closures IF they workpic.twitter.com/pP5HXxd4Vk
15/n The flipside to this is that we know very well the negative social implications of school closures
They are not great for a whole host of reasonspic.twitter.com/DJ3xkTQDVP
16/n Which brings us to the MAIN FINDING of this systematic review
There isn't much data 

We really don't know how great a policy school closures may be, especially considering the societal impact of the policypic.twitter.com/NtCIFZ8FPb
17/n As a package intervention, it seems that school closures may be effective, but we have no real idea what role they played in reducing transmissionpic.twitter.com/Ab3ozYsFff
18/n On the other hand, there isn't even much insight into non-closure options for schools
No one has so far looked into how we might socially distance children without closing schools entirely
Is it possible
No ideapic.twitter.com/YGZO8YGjrd
19/n In addition to this, the only modelling study to look at school closures separately, which also assumed that they were effective, found a "relatively marginal" impact from the interventionpic.twitter.com/n9dp3wHUJK
20/n We also don't know to what extent non-school social gatherings play a role in transmission, and whether these may lower the impact of school closures as wellpic.twitter.com/sCEXqbgzRf
21/n It's also worth noting that the 2-4% figure that you may've heard on the news or from certain celebrity doctors is from a model ASSUMING that school closures do indeed work
22/n Which brings us to the conclusions of this policy-focused review What should countries do? Well, Taiwan and China may be a guide - school closures early, but one of the first measures to be scaled backpic.twitter.com/6un7nUszJ1
23/n However, given that we aren't sure how effective school closures may have been - potentially quite effective - we should consider OTHER ways to minimize transmission while opening thempic.twitter.com/TtcJIF66P9
24/n It's unlikely that we can keep schools closed forever, but there may be ways to reduce harms without the enormous cost of closing them completelypic.twitter.com/yN1MGAviMw
25/n So...do school closures work to prevent the spread of #COVID19?
WE DON'T KNOW
They probably help, but maybe not that much and have some definite harms
26/n And while dumping on Dr. Oz is fun, he's not entirely misrepresenting the evidence when he says that school closures may have an overall detrimental effect during this pandemic
27/n (Although, to be fair, he says it in a MUCH more absurd way. Fuck, who calls 2-3% of deaths trivial? Ugh)
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.