Thread by @KasperKepp

With many empirical studies now showing more modest effects of lockdowns vs. other NPIs, voluntary behavior, and other confounders, the original model concluding large effects is increasingly criticized, incl. a comment in the journal that published it.

https://twitter.com/the_brumby/status/1349478824606502912

1) "we suggest that the model, and its conclusion that all NPIs apart from lockdown have been of low effectiveness, should be treated with caution with regard to policy-making decisions."
nature.com/articles/s4158…

2) "Such modelling efforts have deemed lockdown to account for 81% of the reduction in R0, contributing to government policies. Here, we show that these conclusions are unsupported and that policies therefore should not be based on these studies."
doi.org/10.1101/2020.0…

3) "The results included in the Nature paper seem to suffer from serious selective reporting, providing the most favorable estimates for lockdown benefit"
doi.org/10.1101/2020.0…

4) "infections were in decline before full UK lock-down (24 March 2020) [...] An analysis of UK data using the model of Flaxman et al. (2020) gives the same result under relaxation of its prior assumptions on R"
arxiv.org/abs/2005.02090…

5) "we are concerned that these studies may substantially overstate the role of government-mandated NPI’s in reducing disease transmission due to an omitted variable bias." nber.org/system/files/w…

6) "Although we do not have the code that has been used in the study by Ferguson et al, we know about the values of the R0 and fatality rate used there, which is excessive."
journals.sagepub.com/doi/pdf/10.117…

7) "Flaxman et al. concluded that [...] lockdowns in particular - have had a large effect on reducing transmission’.[...] a model that has better fit to the data [...] reduces the estimate of ‘counterfactual’ deaths [...] from 3.2 million to 262,000"
medrxiv.org/content/10.110…

8)"These are strong assumptions indeed but woefully misleading, the problem being that they regard interventions, lockdowns in this instance, as fixed treatments in fixed applications hav-ing fixed effects."
journals.sagepub.com/doi/pdf/10.117…

9) "It follows from the above that that study provides no information whatsoever as to the actual contribution from all NPI combined to the reduction in transmission"
nicholaslewis.org/wp-content/upl…

10) "Despite including details of the contagion and response options, their model is several degrees of abstraction away from what is warranted by the situation."
static1.squarespace.com/static/5b68a4e…

Lockdowns work but question is how much vs. milder NPIs and voluntary behavior. Social distancing can be reached by many means.

The model reshaped our thinking in 2020 of lockdown as a superior tool. The 10 papers/reports above suggest that we should rethink this view in 2021.

@VPrasadMDMPH
@AgnesWold
@the_brumby
@MartinKulldorff
@carlheneghan
@gummibear737
@mgmgomes1
@NickHudsonCT
@MLevitt_NP2013
@jhnhellstrom
@dockaurG