Read on Twitter
Anders Tegnell and @Folkhalsomynd seem interested in working out the total number of people in Stockholm that have already been infected by #covid19 and thus may have a measure of immunity.
They've spent a lot of time and money doing various seroprevelance studies to find out
They've spent a lot of time and money doing various seroprevelance studies to find out
Sadly, the study designers have struggled with having representative samples of the population to test, so the numbers they come up with aren't necessarily reliable and reflective of the population as a whole.
So I'm going to help them out. Free of charge.
In disease epidemiology there's a metric called the Infection Fatality Rate (or sometimes, Ratio) or IFR. This is a measure of the number of people who have been infected who die.
In disease epidemiology there's a metric called the Infection Fatality Rate (or sometimes, Ratio) or IFR. This is a measure of the number of people who have been infected who die.
There's another measure, called the Case Fatality Rate, or CFR, which is the percentage of those who have been *confirmed* infected who die.
Obviously, unless everyone who has been infected is tested, the CFR is always higher than the IFR.
Obviously, unless everyone who has been infected is tested, the CFR is always higher than the IFR.
Now, if you know the IFR of a disease, you can reasonably well predict the number of people who will die if they are infected. Back in mid-February, WHO reported an estimated IFR for SARS-CoV-2 of between 0.3% and 1%
https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200219-sitrep-30-covid-19.pdf?sfvrsn=6e50645_2
https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200219-sitrep-30-covid-19.pdf?sfvrsn=6e50645_2
When countries like the UK, Sweden, and Netherlands indicated they were considering natural "herd immunity" as their strategy to combat this disease, this allowed some of us to do some simple math.
If at least 60% of the population needs to be infected to achieve herd immunity, and up to 1% of those die ... how many people are potentially going to die in order to achieve herd immunity?
It's a simple equation, no fancy SEIR modelling required. 1% of 60% of the population.
It's a simple equation, no fancy SEIR modelling required. 1% of 60% of the population.
For Sweden, that's 0.01 x 0.60 x 10.1 million people.
60 600 people. Dead. For natural herd immunity.
Needless to say, I didn't think natural herd immunity seemed a good idea! Others with more influence than me apparently disagree.
60 600 people. Dead. For natural herd immunity.
Needless to say, I didn't think natural herd immunity seemed a good idea! Others with more influence than me apparently disagree.
Now, since February we have a lot more knowledge about the disease. A recent Italian study showed the large difference in IFR in different age groups. For the elderly, exceeding 8%, for under 18, 0.002%. https://twitter.com/stats_q/status/1290330191462629388?s=20
Other studies have come up with slightly different results, and just like with age, other factors will give whatever group of people you chose to categorise together different IFRs than other groups. WHO, today, has settled on 0.6%. https://twitter.com/DrEricDing/status/1290592983197384708?s=20
So let's redo that Swedish calculation.
0.006 x 0.60 x 10.1 million people.
36 630 people. Dead. For natural herd immunity.
Better ... but honestly .... I still don't think natural herd immunity may be the best idea.
0.006 x 0.60 x 10.1 million people.
36 630 people. Dead. For natural herd immunity.
Better ... but honestly .... I still don't think natural herd immunity may be the best idea.
Back to my offer though! Helping out Dr Tegnell with his desire to know how many Swedes have been infected.
The fun thing about math is you can often do it backwards.
In this case,
Infection Fatality Rate = (# of dead / # of infected) x 100
The fun thing about math is you can often do it backwards.
In this case,
Infection Fatality Rate = (# of dead / # of infected) x 100
So, we know the average IFR now, from many studies. It's about 0.6, so we can put that in the equation.
0.6 = (# of dead / # of infected) x 100
we can then divide both sides by 100 to get -
0.006 = # of dead / # of infected
0.6 = (# of dead / # of infected) x 100
we can then divide both sides by 100 to get -
0.006 = # of dead / # of infected
Now, here in Sweden we have a reasonable estimate of the number of people who have died from Covid-19. @Folkhalsomynd currently reports it as 5747. @StatsSweden excess mortality data suggests there was many as a thousand or more back in March April who were missed though
To make the math a little easier, let's just round it at 6000 dead. Any pedants can do the math with 5747 instead if they wish.
So ... plugging the number of dead in to the equation ....
0.006 = 6000 / # of infected
So ... plugging the number of dead in to the equation ....
0.006 = 6000 / # of infected
A little more math, we have
0.006 x # of infected = 6000
thus, # of infected = 6000 / 0.006
And we have our answer.
Dr Tegnell, the # of Swedes who have been infected with SARS-CoV-2 is approximately 1,000,000!
1 million Swedes have had the virus!
0.006 x # of infected = 6000
thus, # of infected = 6000 / 0.006
And we have our answer.
Dr Tegnell, the # of Swedes who have been infected with SARS-CoV-2 is approximately 1,000,000!
1 million Swedes have had the virus!
Today, according to Reuters, Dr Tegnell said "the spread of the virus indicated very strongly that Sweden had reached relatively widespread immunity" and the Swedish strategy had worked. https://www.reuters.com/article/us-health-coronavirus-sweden-strategy-idUSKCN24M25L
Yet the numbers are clear. The percent of the population that has been infected is likely no more than about 10%.
This is nowhere near herd immunity.
This is nowhere near herd immunity.
But what about Stockholm?!?! I hear supporters already asking.
Ok, so do the math. 2500 dead. 2.3 million population.
# of infected = 2500 / 0.006 = ca 417,000 people
Or approximately 18% of the population.
Ok, so do the math. 2500 dead. 2.3 million population.
# of infected = 2500 / 0.006 = ca 417,000 people
Or approximately 18% of the population.
This is entirely consistent with the various seroprevalance studies that have been done.
But, alas, nowhere near herd immunity even for the most optimistic calculations. https://sverigesradio.se/sida/artikel.aspx?artikel=7497358
But, alas, nowhere near herd immunity even for the most optimistic calculations. https://sverigesradio.se/sida/artikel.aspx?artikel=7497358
