Hacker News Reader

Comment by alltakendamned | Hacker News Reader

"COVID is over in Sweden" is a pretty bold claim with ~400 daily infections, and increasing again as per https://experience.arcgis.com/experience/09f821667ce64bf7be6...

Just like in other EU countries, it is now the younger population which seems mostly affected, and thus the mortality is lower - which is ultimately some good news.

Summer holidays are now almost over though, so people are traveling home from vacations abroad, are going back to work, schools will start up soon and people finally get access again to health care (good luck seeing a doctor in July in Sweden). The impact remains to be seen, but it would be surprising if most other EU countries see infections increasing but not Sweden.

I remain pretty skeptical of this Swedish exceptionalism, so my guess is this blog post will not age well.

View on HN

https://www.google.com/search?q=sweden+covid+deathrate&oq=sw...

Sweden's death rate is down to zero today and has been trending down since mid April. Summer Holidays did not start until late May where deaths were already half of what they were in mid April. This post isn't aging, it's already aged.

Want to make any guesses as to what Sweden's all-cause mortality for 2020 will be?

https://www.statista.com/statistics/525353/sweden-number-of-...

https://imgur.com/9NS5LFh

Worse than 2019, which was an exceptionally good year, but most probably below the ten-year average. Might even land below the five-year average.

> Sweden's death rate is down to zero today

No, it isn't.

Sweden has fairly laggy reporting of deaths, with the deaths often taking several weeks to show up in statistics, so at any given date, deaths attributed to the last few days are always dropping to close to zero.

I looked into this a bit in mid-APRIL, when somebody claimed that deaths were dropping to zero. At the time, substantial revisions occurred for about three weeks after the date, and occasional revisions for about four weeks.

(It's entirely possible that similar phenomena occur in other countries; Sweden is the only country I looked at in detail).

Several weeks is a bit of an exaggeration. Lag in reporting has been around 4-5 days on average. Lately the lag has increased somewhat as the number of reported deaths has gone down.

Don’t take my word for it, see for yourself at https://morberg.github.io/covid-notebook/covid.html

GitHub repo here: https://github.com/morberg/covid-notebook

Average new cases per day has dropped from about 1200/day in late June to less than 300/day now. There's a slight uptick over the past week.

https://91-divoc.com/pages/covid-visualization/?chart=countr...

Downtrend in average deaths per day continues, currently at 6/day.

https://91-divoc.com/pages/covid-visualization/?chart=countr...

> There's a slight uptick over the past week.

Any uptick at all is concerning. In fact, a flattening of the decline is concerning.

The implied thesis in the linked article is that Sweden has reached a level of herd immunity due to widespread T-cell responses, and that life in the country continues as normal.

If that is the case, then we would expect the infection curve in Sweden to follow a classic logistic model -- which is more or less symmetric. A rapid rise should be accompanied by a rapid fall.

An asymmetric curve or one that has a "second wave" of any size tends to refute this theory. It would suggest that there was in fact some meaningful intervention that changed: either the initial infection was among a different group (i.e., what the US saw with infections first dominated by New York and then later by other states) or there was an effective policy response that has begun to wane.

Logistic model? Never saw anyone use that model. Do you have reference?

What one should expect, afaik, is a Gompertz curve - the derivative of which, depending on parameters, often rises quickly and drops slowly.

Michael Levitt calibrated a Gompertz model in April based on actions taken until that day, that predicted 5000 dead by 1-jul matched reality incredibly well - whereas the WHO/Ferguson model at the same time predicted 96,000 dead by 1-jul based on the same data.

> It would suggest that there was in fact some meaningful intervention that changed

Annual lockdown month happened:

https://hejsweden.com/en/swedish-holiday-dont-work-in-july/

Schools even close for ten weeks, and those will be over soon.

Simple mathematical models are not appropriate for a situation as nuanced and multivariate as this one.

Case count will oscillate up and down as restriction levels and individual comfort levels oscillate. As we approach herd immunity, the oscillations will dampen while following a downtrend.

The cases/day downtrend is still very much intact, and while it could reverse, it is premature to be concerned.

I'm quite fed up with people counting the number of cases as if it has any value, it's like they don't want to admit we were wrong for creating such a panic. Infections are meaningless.

Let's count how many people are in ER and how many die.

The conspirationist in me can't help but think that COVID has been the scapegoat for the worst financial crisis in this century.

While we look at meaningless COVID stats and debate wether we should wear a mask, we don't look at the real culprits for the current financial crisis that will take decades to heal (if it ever heals fully).

daily infections are irelevant if almost all of them are asymptomatic or extremely mild not requiring any hospitalization

Czechia has 200-300 daily infections and we have like 30 hospitalized patients also in country of 10.7M, so there is pretty much zero chance of dying from COVID, from those 390 people who dies in those 4 months (where usually dies around 40000 people from all causes) had other issues and coronavirus was unlikely cause of death

my guess is people who fell for this mass hysteria will be very quiet in year or two when we will look at yearly death stats and see mere blip because of COVID, where just cause of death may temporarily shift to COVID, but in the end total numbers of death will remain same. UK stats shows already for 6 weeks lower than 5 year average mortality and number of excess deaths is shrinking

https://twitter.com/EdConwaySky/status/1290568780251308032

> daily infections are irelevant if almost all of them are asymptomatic or extremely mild not requiring any hospitalization

In fact, more daily infections without more deaths are a good thing. It means the virus is not as dangerous as we thought and we're closer to herd immunity/the end of the pandemic.

I’m struggling to get your point about the U.K. 161 deaths below average for a week - if the trend holds for the remainder of the 22 weeks this year then that’s 3.5k deaths less which is dwarfed by the excess so far, so I can’t see that averaging out? Unless you’re point is something else?

Reply on news.ycombinator.com