Sweden: the undeserved battle ground of the pandemic mess. Thankfully we do have at least one country that did things differently to help us learn what really matters in our policy responses. Folks love to cherry-pick pairwise comparisons to support their views. Maybe tho ... 1/
... policy actions don’t matter much at all as far as the virus’ trajectory. Which if true would mean we ruined economies, mental health and so much more for little if any benefit. So I decided to do some analysis using OWID and Oxford Lockdown Stingency data to look at this 2/
I plotted daily Cov19 deaths per Million using a 7 day ma on a log scale to show growth rate. And then I shifted this curve 25 days earlier to approximate deaths by day of infection. I also plotted lockdown stringency to look for any correlation to virus trajectory. 3/
Let’s start w Sweden. What do we see? Infections as estimated by deaths from 25 days later started growing at a true exponential 25% rate (compounding at 1.25%^t). But we see a break in rate rather soon at a daily level of 4D/M. Long before the weak 46 stingency nonlockdown. 4/ 
We also see that by the time the nonlockdown hit, deaths had peaked around 10/M and started declining. Did their nonlockdown even matter beyond possibly early mild measures? Let’ s look around the EU 5/
Here’s the UK. A very similar curve to Sweden. Same initial 25% exp growth, higher breakpoint at 5/M, and higher peak at 14/M. Growth rate nosing over before more stringent 76 lockdown. Looks like that wasn’t useful... now let’s look at the worst hit...6/
Belgium. Here we see a much worse 35% exp growth rate, a similar breakpoint at 5/M but a massive peak at 26 deaths/M. The growth rate is certainly declining pre lock but the “momentum” took longer to stop. Next, we look at the Sweden-deniers’ favorite compare, Norway. 7/
Surely Norway did a great job locking down and squashing the curve- or did they? Norway has admitted they regret their short but stringent lockdown. Here we see that Norway barely had a curve to squash! Relatively low 15% initial exp growth, breakpoint at <1 D/M, peak 1.5 D/M. 8/
And if the Norwegian lockdown did anything it was the early mild Sweden-level stuff. So maybe now we start to see that the virus was very different in different countries. But not due to political policies. One more country for now. Germany! 9/
Germany is typically used as an exemplar. And look at the data- they look like Norway. Relatively low 20% initial exp growth. Low 1.8 D/M breakpoint. Low max of 2.8 D/M. And the max is hit after the most stringent lockdown. 10/
What does all this tell us? To me it says that each country’s fate was sealed long before any actions were taken. Some countries got it bad, others just didn’t. To attribute this to policy actions would appear ridiculous. Comparing Sweden to Norway is also ridiculous. 11/
Look how similar Sweden, UK and Belgium curves are. And how different Norway and Germany are. For sure Sweden UK Belgium are similar industrially and travel hub-wise. Norway is relatively isolated. Germany is tougher to explain. 12/
Countries with high initial growth had high peaks. Countries with low initial growth had low peaks and much better outcomes. This was pre determined before any lockdowns. Stringent lockdowns appear useless. Mild measures may help, no surprise. 13/
Bottom line: those countries that for whatever reason were hit by a large viral load before anyone was aware it was spreading got it bad (NYC prob fits here). Others were more lucky. But then the virus largely did what it was going to do based on the foothold it had 14/
Could my hypothesis be wrong? Of course! Tell me why it is or isn’t. Unlike certain groups I am looking for disconfirming evidence and will accept it. I have many more of these graphs. And I have yet to find any disconfirming evidence myself. /end
