Read on Twitter
THREAD
This article is a challenging read. I had a go at replicating its findings for Ireland to assess the effectiveness of the two lockdown stages in March.
I'm quite surprise by what I've found.
1/ https://www.thepublicdiscourse.com/2020/04/62572/
This article is a challenging read. I had a go at replicating its findings for Ireland to assess the effectiveness of the two lockdown stages in March.
I'm quite surprise by what I've found.
1/ https://www.thepublicdiscourse.com/2020/04/62572/
Ireland went into lockdown in two stages.
The "first measures" were on March 12th with school closures, social distancing and a ban on large gatherings.
2/ https://www.rte.ie/news/coronavirus/2020/0312/1121849-taoiseach-full-statement-coronavirus-ireland/
The "first measures" were on March 12th with school closures, social distancing and a ban on large gatherings.
2/ https://www.rte.ie/news/coronavirus/2020/0312/1121849-taoiseach-full-statement-coronavirus-ireland/
Then on March 28th, we began the "full lockdown", with non-essential workplaces shut and the 2km rule.
3/ https://www.rte.ie/news/2020/0327/1126911-ireland-restrictions-covid19/
3/ https://www.rte.ie/news/2020/0327/1126911-ireland-restrictions-covid19/
The key point in the "Lockdowns Don’t Work" article is the delay between transmission, symptoms and death.
On average:
Transmission = day 0
Symptoms = day 5
Death = day 20
4/
On average:
Transmission = day 0
Symptoms = day 5
Death = day 20
4/
So the March 28th lockdown should have led to a peak in deaths taking place over 20 days later, any date after April 17th.
That's pretty much what the headline data shows. April 20th saw the largest number of new deaths (77).
5/
That's pretty much what the headline data shows. April 20th saw the largest number of new deaths (77).
5/
However, we know that the date of death being announced is several days *after* the death actually took place.
The HSE has given some indication of this difference, but not a regular dataset.
6/
The HSE has given some indication of this difference, but not a regular dataset.
6/
The best source we have on this is http://RIP.ie deaths, as published by @seamuscoffey.
This data shows a peak in deaths around April 17th
7/ https://twitter.com/seamuscoffey/status/1254106762233470978
This data shows a peak in deaths around April 17th
7/ https://twitter.com/seamuscoffey/status/1254106762233470978
However, there is a lag between deaths taking place and a posting on http://RIP.ie . Typically about 2 days.
So the peak is more likely around April 15th.
The problem is, that's 2 days before the March 28th "full lockdown" should have had an effect.
8/
So the peak is more likely around April 15th.
The problem is, that's 2 days before the March 28th "full lockdown" should have had an effect.
8/
Bear in mind again that this is the analysis in the original article.
1) Lockdown begins
2) Peak deaths happens
3) Lockdown effectiveness kicks in (but deaths have already peaked, suggesting measure *before* lockdown were the main reason transmission stopped).
9/
1) Lockdown begins
2) Peak deaths happens
3) Lockdown effectiveness kicks in (but deaths have already peaked, suggesting measure *before* lockdown were the main reason transmission stopped).
9/
This is my replication of the same chart for Ireland.
It shows a similar trend.
The full lockdown wasn't the main cause for peak deaths!
10/
It shows a similar trend.
The full lockdown wasn't the main cause for peak deaths!
10/
The implications of this are profound.
It suggests that the social distancing alone (between March 12th and 28th) was the main driver of #FlattenTheCurve.
11/
It suggests that the social distancing alone (between March 12th and 28th) was the main driver of #FlattenTheCurve.
11/
This doesn't mean that the March 28th lockdown was ineffective. It had its own role in saving lives.
But, it appears we may have hit peak deaths anyway without it.
12/
But, it appears we may have hit peak deaths anyway without it.
12/
Here's the data in a table to give another view.
13/
13/
This analysis gives me greater confidence that we can lift on May 5th.
We could potentially return to the days of March 12th-28th. Social distancing still in place, but 2km lifted and shops back open!
14/14
We could potentially return to the days of March 12th-28th. Social distancing still in place, but 2km lifted and shops back open!
14/14
