Hacker News Reader

Comment by gbasin | Hacker News Reader

Seems like Sweden is on track to hit herd immunity long before everyone else, with minimal economic damage. Their daily new case count has been mostly plateaued for weeks, like many other places in full lockdown.

View on HN

Daily new cases reaching new highs, but definitely not exponential growth

https://www.worldometers.info/coronavirus/country/sweden/

Average time between infection and death from COVID-19 is 23 days, and 18 days average from first symptoms to deaths (1) A country can be a week or two into the overload level before it shows up in the ICUs

https://twitter.com/michaeljswalker/status/12454153391497052...

Yeah, estimates range from the current number being around 20-30% of the population. Here's a report from the Public Health Agency on the topic, based on a random sampling study: https://www.folkhalsomyndigheten.se/publicerat-material/publ...

Here's another paper from a orofessor of mathematics at Stockholm University who has appeared in the media as a subject matter expert: https://www.medrxiv.org/content/10.1101/2020.04.15.20066050v...

Meanwhile, there's a hint of a decline in the number of patients in Intensive Care in Stockholm, as seen here, although it's a still a bit early to tell: https://www.svt.se/datajournalistik/corona-i-intensivvarden/

Meanwhile, there is reason to believe that Sweden is better than other countries at attributing the cause of death correctly: https://www.nytimes.com/interactive/2020/04/21/world/coronav...

This is inaccurate. According to

https://studylib.net/coronavirus#country-se

their case count is steadily climbing, as is their CFR.

I think what he meant that the increase has platued, meaning a linear growth in infected, as opposed to exponential.

Other than that the numbers of people in intensive care and the deaths per day has been decreasing during last week, something that in my eyes are even more positive.

(data according to FOHM, Swedish CDC kinda)

Not according to the John Hopkins dashboard. Maybe that's why they're now starting to take action?

Do you have a link to the dashboard? The data I normally look at (https://www.worldometers.info/coronavirus/country/sweden/) suggests they've been sitting at just under 800 (with fluctuations) since the beginning of the month.

https://coronavirus.jhu.edu/map.html

Well, it hasn't been exploding like some critics said it would.

https://ourworldindata.org/grapher/new-covid-deaths-per-mill...

The rolling 7 day average shows they're still on a bit of an up-tick.

https://ourworldindata.org/grapher/daily-covid-deaths-per-mi...

However, if you normalize for start time they're not really doing any worse than a lot of other countries.

They have more than 5x the deaths relative to population than Norway has, and more than 7x the deaths of Finland, and it seems those would be reasonable benchmarks to compare against.

I guess everything is relative, but I'm not sure I would say the economic damage is minimal. Maybe in comparison with other countries, but I'm not sure how clear that is. Especially if other countries can go South Korean style and get back to full speed faster, Sweden could very well be left behind in terms of the economy as well.

Number of deaths so far is very high compared to the neighbouring countries, and there is still a long way to go before herd immunity.

In general, it seems to me quite insane that Sweden already in the beginning of March basically had decided to infect millions of people with a virus we only had a month of experience with or something like that.

Like every other time people talk about economic damage and "letting the cure be worse than the disease" the specific dog whistle above is talking about the impact to companies, in particular the banks which rely on rent and loan payments, which has been heavily targeted for indirect subsidy in the US - compared to talking about the damage this does to individuals and their well being, all of which can be well supported through this crisis with an approach like that taken by Denmark.

> Seems like Sweden is on track to hit herd immunity long before everyone else

There's zero evidence of that. At this time, officially less than .5% of the population has been infected, so you'd have to postulate something like 99% of cases being asymptomatic.

> Their daily new case count has been mostly plateaued for weeks

Case count depends a lot on amount of testing (In which Sweden lags behind even the US) and testing criteria. Their death count does not look great.

There is very substantial evidence of that. A random sample was taken of 700 inhabitants in Stockholm, which came in at 2.5% positive at that time. Based on that, they model that around 600 000 people have been infected.

https://www.folkhalsomyndigheten.se/publicerat-material/publ...

600000 people is 6% of the Swedish population. So Sweden would be maybe... 10% of the way to herd immunity? That does not really count as close in my book.

That estimate is exclusively for Stockholm which is ahead of the rest of Sweden. Stockholm has ~2 million inhabitants, so that’s 25%+. That said, the key aspect is of course whether the intensive care units can keep up with the needs. So far, that has been true for Stockholm, which ptobably means that the measures are adequate for Swedish circumstances. Notably, the number of cases for Stockholm is projected to have peaked around April 10th.

I stand corrected regarding their prediction. I don't understand much Swedish, so I could not follow their model. But it appears to rely on an extrapolation of data from late March. Do they include any cross-validation of the model with current data?

April 9, 10, 11: 200, 148, 240 new cases were reported in Stockholm

April 21, 22, 23: 163, 289, 281 new cases were reported (and might still be subject to upward revision)

As we both agree, reported cases is a fairly uncertain metric, but I don't see a peak on April 10. Do you?

https://experience.arcgis.com/experience/09f821667ce64bf7be6...

The model is calibrated based on the study from late march, which randomly sampled 700 individuals in Stockholm. From that they can infer the number of infected people at that point in time. They use two separate variables to track unknown infections and known infections (hospitalizations), so from that study and hospital data they get a baseline. They can then verify how the model holds up for the group of hospitalized cases.

There are charts towards the end of the paper which shows the real data plotted compared to the predictions, with confidence intervalls. The data is obviously very noisy, but it seems to track reasonably well, and is expected to taper of only slowly. The expected decline is not by any means drastic.

Swedish officials do believe that something like 95% of their coronavirus cases aren't reflected in the official count. Most countries are undercounting by at least an order of magnitude, and as you say Sweden isn't doing a ton of testing, so this makes sense.

> so you'd have to postulate something like 99% of cases being asymptomatic.

No, that's not how it works. There are plenty of people who have had and do have symptoms but have not been counted. Seeking care and getting counted is a last resort here.

One caution on this; they've a pretty low rate of testing, certainly lower than most of the rest of Western Europe.

And that seems to anger a lot of people, as if they are invested in the month they've spent inside. People are strange ...

Indeed, we should try to learn from them instead! https://twitter.com/garybasin/status/1252721673566117895?s=2...

I'm not sure that's fair. I've only seen anger from the same set of people who write angry rants about how everyone wants to kill grandma. Most opinions range from curiosity how conventional wisdom can be so different in Sweden to quiet resignation that the rest of us might have made the problem significantly worse.

Herd immunity without a vaccine isn’t a strategy. It’s just justifying doing nothing.

Not really, if age separation can be achieved efficiently, herd immunity wouldn't be impossible.

The problem is that you really need an advanced country to organize separating people by age (making young people get through the virus without passing it to older people).

I would have significantly more respect for the Swedish approach if they had protected their elderly and other vulnerable communities. The fact is that the weak do not account for an unavoidable minority of the deaths.

The Swedish approach may get them out the other end before any country aside from China and Vietnam, but the Swedish execution did not save lives by protecting the vulnerable.

I agree with you totally. In Hungary our first hospitalization was because of a guy who lives in UK and travelled to Hungary to visit his parents. Of course it was his father who went to hospital, not him.

That was the intention, and the actions where designed for it. However, the care of elderly are handled by the municipalities, who didn’t have any readiness for handling it properly. Hopefully some of those mistakes can be avoided outside of Stockholm by more extensive testing of personell.

Judging by the downvotes on your comment, you must not have gotten the infographic explainer:

https://babylonbee.com/news/infographic-common-covid-19-talk...

Being on the other side of the opinion spectrum, I see what looks like wishful thinking on the part of the "Herd Immunity" adherents, with a lot of goal post moving.

Two weeks ago, many of them lionized Dr. Knut Wittkowski. But now that his forecast of 10000 (sic!) deaths in the US has proven somewhat optimistic, I haven't seen any followups where he explains what he got wrong in his model, and what adjustments to it he's making.

Richard Epstein is at least making revisions to his estimate, but unfortunately, he's making them retroactively as well: https://www.bradford-delong.com/2020/04/the-hoover-institute...

A few weeks ago, the UK briefly flirted with a herd immunity strategy. Would anybody today argue that these days were wisely spent?

Now, the benchmark du jour is Sweden. Will Sweden still be the benchmark a month from now? Would be nice if it were, but I'm not confident.

It seems like there are two possible outcomes:

1. We realize that the shutdown had a small effect, because a) the virus is already widespread due to easy transmission and b) the mortality rate is much lower than previously thought. Thus, we don't need to continue the shutdown.

2. The virus turns out to be as bad as feared, and things get much worse in Sweden (and elsewhere). The shutdown needs to continue indefinitely.

I'm leaning toward #1, but I think it's still too early to rule out #2. If we are living in the first scenario, I wonder how long it will take for the general public to realize that's the case.

I do think that a combination of political jockeying and mass hysteria have placed us into the position we're in now. If that's true, then the worst thing about this virus will certainly be the economic impact which has a number of far reaching effects.

Here's an actual science-based report detailing how we can start to reopen with proper testing and tracing. I think Germany is doing something along these lines?

https://www.nga.org/wp-content/uploads/2020/04/NGA-Report.pd...

It's by the national governors association in the US and seems pretty well done.

New York City is increasingly providing a lower bound for mortality rate in a population with relatively high infection rates, and it's not all that low. You would have to postulate some mechanism like Herd Immunity Through Blondness to expect a substantially better outcome for Sweden.

Reply on news.ycombinator.com