Hacker News Reader

Comment by fullshark | Hacker News Reader

Financial times has been doing this exact analysis country by country / city by city:

https://www.ft.com/content/a2901ce8-5eb7-4633-b89c-cbdf5b386...

They estimate about a 20% increase in deaths so far this year for Sweden.

Some people have argued that hospitals are skewing the numbers to inflate the number of people that are dying of Covid-19 vs other causes, or that these are people that would died anyways. Neither of these arguments are supported by the FT data. To be fair, it's not clear how to quantify people that might be dying due to lockdown-related causes (eg. suicide, untreated health issues), but this appears to be smaller than Covid-19 deaths.

Edit: I'm curious why this is getting downvoted.

I didn't downvote, but I can see why.

It seems like you didn't read the FT graph. It's showing the total death rates in a graph, and during initial spike in COVID, Sweden showed 20% increase during the peak. By definition, there's no "skewing" possible, since it's all death - unless you're claiming hospitals are making up death.

If your main point is that reported death due to COVID might be inflated, sure, that's possible but that doesn't mean there's no increase due to COVID related death. If you want to point out the reported number due to COVID might be slightly higher than it really is, you should at least acknowledge that there is meaningful and substantial increase in death due to COVID as there absolutely is. By omitting that, your post can't be distinguished from someone posting with a conspiracy mindset and ignoring actual data on substantial increase in death.

Clearly, I must not have articulated myself well enough, since you took away the exact opposite of my point: I have heard the suggestion of inflated death numbers within the US in casual conversation multiple times and I frequently refer to this specific article as a counterargument.

Fwiw - I have been a subscriber to the FT for ages and I've followed that chart (which is free to all) since the beginning.

The inflated death numbers thing is Fox News disinformation.

I think you're post was pretty clear. "Some people argue that X, but the data doesn't support X." Sometimes there can be a tendency for people to skim posts and try to quickly box people into Team A or Team B. It's not unlikely that some people read your first sentence (or even the first half of your first sentence) and pegged you for having a particular position without paying attention to the whole thing.

> Some people have argued that hospitals are skewing the numbers to inflate the number of people that are dying of Covid-19 vs other causes

This is a common theory in the US as well, where the accusation of "they are doing it for money" is most repeated. It's ridiculous and untrue, but that's the logic here.

What could even a theoretical justification for hospitals inflating numbers in a fully government-run system ?

I'll be briefer than the other reply.

The FT excess mortality graph doesn't care whether anything is flagged as a COVID death or not. If there's a lethal pandemic, you don't need to rely on doctors classifying deaths, you can see it in the total number of deaths from all causes.

That's exactly what they are saying.

I think you're being downvoted because it can be confusing sometimes when someone replies to a post in agreement, it's much more natural on the internet for replies to be a disagreement. This makes it really easy to read your first sentence as something you are proposing rather than referencing, which makes the whole post rather confusing if not read more carefully.

Then the FT estimate is wrong, because the actual statistics in "deaths so far this year for Sweden" are what's linked to and you can clearly see that they're currently below prior years, not above. A short wave in April of excess deaths hasn't changed much on a longer term basis, because there was no winter flu death season this year.

Perhaps you meant "deaths assigned to COVID". The FT work has a few issues. One is that they assume any mismatch in COVID vs excess death means undercounting. This is based on the faulty assumption that lockdowns don't kill anyone, but in the UK the hospitals were emptied out, people were afraid to go to A&E. Admissions halved almost overnight and took months to return to prior levels over the summer, then they dropped again due to the second lockdown. Actual death certificate data in the first wave showed about half of all excess death didn't mention COVID anywhere. The FT's assumption is strange and unsupported by any data.

I admit I am confused by your comment and what point you are making. I"m talking about the second graph I linked to, AFAIK FT shows deaths per week from all causes and compares to historical average. Then it basically does an integral of the deaths per week in 2020 compared to the historical average to see the excess deaths (it does appear to only consider the outbreak / times when deaths per week is above historical average though). That at least is what is labeled as the shaded red areas, I admit i didn't read every footnote.

What assumption are they making here that is weird? Unless there are fake deaths I'm not sure why this analysis is faulty, this analysis doesn't care if something was labeled a COVID death or not.

Shading excess deaths isn't weird, that's fine. However it's misleading because they don't compute the reverse integral. This article does:

https://www.aier.org/article/swedens-dry-tinder-accounts-for...

If you take into account the amount by which Sweden's excess deaths were lower in the runup to COVID than expected, and in particular, lower than its neighbours, then Sweden starts to look more like its immediate bordering countries and in particular, 2020 looks like a normal year that simply caught up with an otherwise very non-deadly 2018-2019.

The weird assumption they're making is that excess deaths == COVID deaths. It's stated half way down.

Yeah its true they only look above the curve, the excess deaths this year is probably closer to +15% based on that time series (eyeballing the first few months and being conservative). In any case this argument seems to be that because Sweden had a larger share of vulnerable population in the run up to the outbreak, you can't count those deaths as COVID deaths, which I don't agree with. It just means they had a more vulnerable population (dry tinder so to speak) and was possibly disproportionally impacted by COVID as a result. Obviously some % would have died even without COVID, possibly above the historical average but we have no way of knowing that and I've personally decided on excess deaths as the best though imperfect metric to measure COVID's impact. It is possible in 3 years we look back 4 years and find no excess deaths over that time period, which means basically COVID's major impact was shortening the lifespans of our oldest cohort of citizens, which is undeniably tragic but perhaps we acknowledge some things we did to stop that from happening were overkill.

In any case the only reason I posted in the first place was OP had an intentionally misleading bar chart. Arguably +20% was also misleading (since it wasn't excess deaths in 2020, just excess deaths post outbreak) but you can look at the time series here (and posted above) and look for yourself. My personal belief is Sweden saw larger deaths than Denmark/Norway in part due to their COVID policies, but they did not experience the complete disaster many predicted in March. Whether the trade off was worth it will be debated by public policy experts for years, and ultimately no one will agree.

Nobody said you can't count those deaths as COVID deaths when they actually were, that's not the argument at all.

The argument is that (a) not all excess deaths were caused by COVID, only some and (b) excess deaths are very low compared to the apocalypse that was promised, in fact Swedish excess deaths are totally normal and if the modellers had never showed up and told people the world was going to end unless we all lock down, nobody would have thought anything was happening in Sweden. There might have been a few press stories about an unusual form of pneumonia or a new coronavirus but it wouldn't have got more attention than the bad flu year of 2017 did in the USA.

It is possible in 3 years we look back 4 years and find no excess deaths over that time period

Right. The point I'm making is you don't have to wait 4 years. The data to make that call is already in: Swedish excess deaths for 2020 are on track to be normal. Very unlikely there'll be a sudden massive surge in the last month of 2020 given Sweden's near-zero COVID-linked death rate. Outside of the first wave excess death in 2020 has been lower than average for Sweden so it's already balanced out.

My personal belief is Sweden saw larger deaths than Denmark/Norway in part due to their COVID policies, but they did not experience the complete disaster many predicted in March. Whether the trade off was worth it will be debated by public policy experts for years, and ultimately no one will agree.

I feel the Danish economist's analysis is convincing w.r.t. "dry tinder". I fully agree with the last sentence though!

No, the FT article is referring to excess deaths during the outbreaks, which of course is higher than the average for those months.

Mortality for the year in total is on average.

True, my mistake. I'd agree that would be a major concern if on those trend lines post outbreak the 2020 deaths clearly dropped below the historical average. Given the fact that the year isn't over I'll take FT's analysis over that bar chart OP linked to. If what you say is true it appears to be due to the fact that Sweden's 2020 deaths were below average in February, based on those charts that seems unlikely though. However +20% is an overestimate as a result.

Wait, 273,000 excess deaths in the United States?

That is the population of Fort Wayne or Toledo.

Reply on news.ycombinator.com