European mortality monitoring shows increase in excess mortality

https://www.euromomo.eu/graphs-and-maps/

by gstipi • 6 years ago

74 124 6 years ago

euromomo.eu

justinclift• 6 years ago

The Swedish "Very high excess" is a bit surprising.

Extremely high excess: UK, France, Spain, Italy

Very high excess: Sweden

High excess: Switzerland

Moderate excess: Portugal

Low excess: Denmark

fiblye2• 6 years ago

Sweden's current policy is basically to do nothing and let nature take its course.

Bombthecat• 6 years ago

They have limitations.

They also started with way lower numbers. So keeping open with limitations is way easier.

Also, its probably the right approach. You cant lock down for one or two years at least..

styren• 6 years ago

I'm Swedish and this is not true at all, please stop spreading misinformation

henrikschroder12• 6 years ago

You are 100% wrong, Sweden is doing a lot of things to minimize the coronavirus impact.

But I'm curious why you are claiming something that is so completely wrong. Where did you get the idea from? Are you living in Sweden and strongly disagreeing with the official response, or are you living somewhere else, but "read it on the internet" ?

jobigoud• 6 years ago

Just yesterday there was a topic about how Sweden had low excess mortality and even in the negative, with all sorts of comments explaining the reason.

henrikschroder12• 6 years ago

No, that was someone here on HN who couldn't read the data from the NYT article correctly. Sweden is absolutely experiencing excess deaths right now.

What the article said is that a bunch of countries have fewer covid-19 deaths than excess deaths for a certain time period, while Sweden has more covid-19 deaths than excess deaths for the same time period.

The conclusion of that article is that most countries are probably severely undercounting covid-19 deaths, while Sweden probably isn't.

(This isn't necessarily true, another plausible explanation is that a lot more people than normal are dying of heart problems because they're too afraid to seek medical care since hospitals are pretty busy with all the covid-19 patients.)

gpderetta6• 6 years ago

Actually no, the excess mortality that was reported yesterday was not negative. What was (very slightly) negative was the excess mortality minus known covid deads, which means that sweden excess mortality is accounted for. Some places like Italy are seen excess mortality twice as big as the number of accounted covid deads, which means that they are way under reporting.

osobo• 6 years ago

These analysis are based on data provided by the countries. While they all may provide mortality as a number at some point, some may not already do so in the interval needed to draw accurate comparisons over the last two months. For example: I find the difference between England/Scotland, Northern Ireland and Ireland to be strange. I'd expect these to be closer together.

wiredfool• 6 years ago

There are different governments and health services in Northern Ireland and Ireland, and they've taken different paths in their actions and messaging.

derriz• 6 years ago

Why would you expect them to be closer? The respective government responses were very different - the UK initially decided to to pursue a "herd immunity" strategy - i.e. do nothing to stop the spread of the virus. In contrast, the Irish government locked down early and hard - the UK lock down occurred later (once the significance of the threat became more obvious) and even then introduced lock-down measures more gradually.

thu21115• 6 years ago

England has London, Scottish cities aren't on the same scale.

There also appears to be a pretty huge immigrant/racial skew in ICU and death cases for unclear reasons, conjectured to be vitamin D related. This is likely also skewing Sweden vs their neighbours, given Sweden's enormous migrant intake.

Arrrlex• 6 years ago

Could someone help me understand the all ages excess mortality graph? Naively it looks like

a) there were way more deaths in 2018 and 2019 than were expected, and

b) despite Covid-19, the excess mortality still hasn't caught up with 2018 levels

What am I missing here?

soco• 6 years ago

So much interesting data and no idea what it means... why the visible bumps for SOME countries in December 2016 or January 2018? Why Spain has that higher mortality in winter? Why why why...

rjtavares2• 6 years ago

Every European country has more deaths in the winter, mainly due to the flu season (which can happen from December to February).

mrow843• 6 years ago

IANAE but ... all three years show the pattern of what I presume is the flu season (broadly speaking), causing deaths (mostly in the elderly), and differing in severity/rate between the three years.

2020 is different because there is a clear step-change in death rate after week 11, and the implication/fact is that that is COVID-19.

The conclusion is that those deaths are excess deaths on top of those that we would have had as a consequence of the 2020 flu season.

mkagenius• 6 years ago

clear step-change in death rate after week 11, and the implication/fact is that that is COVID-19

How does that compare with 2018 excess death graph, clearly some other novel virus would have caused that, it broke all previous highs quickly.

henrikschroder12• 6 years ago

It's a cumulated graph. Earlier flu seasons were responsible for excess deaths in 2018 and 2019, and since flu season starts earlier in the year than pandemic deaths, by April, more people had time to die than has died in total from covid-19 in 2020 so far.

You can click the "weekly" button and you'll see that covid-19 is deadlier than earlier flus, but has been killing people for a shorter amount of time.

ekianjo3• 6 years ago

a) there were way more deaths in 2018 and 2019 than were expected

Flu season every winter brings excess mortality. Also very hot summers, but there's less of a pattern.

b) b) despite Covid-19, the excess mortality still hasn't caught up with 2018 levels

You are missing a big factor: COVID is not behind us, and this excess mortality is happening DESPITE a lockdown, which never happens for regular flu seasons.

egman_ekki3• 6 years ago

The numbers have been revised, apparently, now it's higher.

itronitron• 6 years ago

Does anyone know what was going on in 2017 (winter 2016/2017) that caused the spike in excess mortality?

mshook• 6 years ago

Flu most likely?

https://www.ecdc.europa.eu/en/publications-data/summary-infl...

tilolebo3• 6 years ago

It's misleading because the 2017-2018 flu was actually much more lethal.

The 2016-2017 goes a bit higher during 02/2017, but its span is also smaller.

martopix• 6 years ago

Sad post: a lot of work, a lot of nice data, zero explanation of what it means.

rezeroed2• 6 years ago

Agreed. Newsflash: put people in cotton wool boxes they'll experience fewer accidents, but might suffer some other issues. Imagine that, more variables outside than inside.

gridlockd• 6 years ago

Numbers do not look that bad yet, but the trajectory is steeper. Is that sharp drop real or just lag?

gpderetta6• 6 years ago

I have been following these reports for the last few weeks. There is often a sharp drop in the end, but it goes away the next week. I believe it is just an artefact of the modelling

john_miller• 6 years ago

A general question: Can someone explain to me, why the integral of the z-score chart is not zero but positive?

Is the sum of the standard deviations not zero?

peter_retief2• 6 years ago

Am I correct to assume from this data that the current flu season (Early onset) is not more severe than previous years?

thepangolino• 6 years ago

According to data, the last previous two flu seasons in Europe have been rather mild.

peter_retief2• 6 years ago

I wonder how the current data compares to severe flu seasons and what the deviation is?

flr03• 6 years ago

When comparing figures between previous years and 2020 one has to keep in mind that the current excess mortality is despite people adopting unprecedented behaviours like strict social distancing, regular hands cleaning etc.

That put things into perspective.

bayesian_horse5• 6 years ago

Some excess mortality might even be caused by such measures. I feel like I would have a much higher threshold before looking for (Covid-unrelated) medical help. Suicides should be on the uptick. Other things like accidents should go down.

I'm not saying the measures and restrictions aren't necessary, just that there are some costs associated. Those deaths aren't really "Covid-19" fatalities even though they are indirectly caused by the pandemic.

onetimemanytime• 6 years ago

Many aren't going to doctor visits, aren't exercising, are eating more etc etc. It all adds up.

I guess Covid is like a really, really bad flu season showing up, in addition to the existing one. So when all is done. we'll really see. Flu does kill a lot of old and infirm people

icebraining2• 6 years ago

I feel like I would have a much higher threshold before looking for (Covid-unrelated) medical help.

In fact, our health authorities have insisted for people to come to the ER, because apparently the number of non-covid emergencies (like strokes) is unreasonably low.

gpderetta6• 6 years ago

That could be. On the other hands some countries that are in lockdown like Germany (although maybe not as strict as some other european states) are seeing no excess mortality at all, which makes the hypothesis unlikely.

DangerousPie2• 6 years ago

On the other hand there are much fewer traffic accidents, much less pollution and people aren't going to bars and getting into fights. So I'm not sure if the net effect of the lockdown on mortality is actually expected to be positive or negative.

ekianjo3• 6 years ago

Suicides should be on the uptick.

could go both ways really. If you are stuck at home with your old folks (probably the case for young people at risk of suicide right now), much less chance to commit suicide with total privacy in the current circumstances.

Also, there's usually less suicides when things go really bad in the society at large, because it puts things in perspective more than any other time in terms of human suffering.

aquadrop2• 6 years ago

Would also be nice if there was cause of death data, so you could filter out stuff like car crashes, murders, accidental falls etc.

danmaz74• 6 years ago

Even more impactful, most people are staying at home, meaning fewer accidents in the streets and workplaces.

kryptiskt• 6 years ago

It could cause extra death because people are reluctant to seek help for other ailments, for example the number of people coming to ER with heart attacks have decreased drastically in Sweden (-25% between February 24 and April 5).

tasuki• 6 years ago

Also keep in mind the graph isn't zero-based which makes it seem worse than it is. It's 73k compared to 52k last year, so about 40% worse than last year, not 100% as the graph might make one think.

egman_ekki3• 6 years ago

I think there was a correction since then. The peak is at 86k now.

aarroyoc2• 6 years ago

These kind of graphs, while useful, still miss a point. My experience in Spain is that there are a some hotspots like Madrid where hospitals and morgues are collapsed but other cities/villages, which have the same lockdown didn't have so many infected people per capita, so they will actually have probably the same or less mortality rates. The country on itself looks bad, but in the hotspots is terrible

cjfd• 6 years ago

In my country in places where there are fewer corona patients the hospitals used to be taking over IC patients from places where there are more. And now the total number of corona patients admitted to the hospital has been declining for about a month so that looks pretty good.

pjcv• 6 years ago

The data here seems to corroborate this: https://www.isciii.es/QueHacemos/Servicios/VigilanciaSaludPu...

It would be interesting to see if previous years exhibited similar hotspots.

thu21115• 6 years ago

But these graphs don't tell a story of collapse. Or perhaps they do but no worse than Spain normally experiences? Spanish excess mortality has peaked already at only slightly above what it peaked at in 2017.

Remember that everything we think we know about collapsed hotspots comes via the media. For every anecdote there are other anecdotes telling the opposite.

The new EuroMOMO site is really much, much better than before (kudos to whoever did that upgrade if you're reading this!). Usefully it now shows cumulative deaths. Compared to 2018 we still have had fewer deaths so far this year even with COVID included. That's staggering. It'll probably change soon given the slope of the graph, but that's really worth thinking about. So far fewer people have died than normal in 2020 than by this time in 2018.

aarroyoc2• 6 years ago

That's what I'm saying. These graphs are too broad for a country because the numbers change a lot between cities. I'm expecting that some cities even report less deaths because of the lockdown (less risky activities, less traffic accidents, less labour accidents,...). So even if the numbers look similar to the ones of previous years, it is not distributed as other years.

Endlessly• 6 years ago

Anyone know why data only goes back to 2016. This organization was started in 2008, did it really take them 8 years to be able to start collecting data?

SiempreViernes• 6 years ago

A common reason for this is some bureaucratic change in the collection procedure that breaks comparability with older data.

Another reason could be that it indeed took 8 years to get 26 countries to agree on, and implement correctly, common enough data reporting standards before the final analysis could be performed.

DanBC• 6 years ago

I think they have a 5 year limit on data. https://www.euromomo.eu/how-it-works/methods

Mortality baseline is modelled using a glm poisson corrected for over dispersion. The model is fitted on valid historical period as defined by the user, with a maximum of 5 years,

I don't know why they impose this limit. I'm guessing there's something about population drift over time in there.

They also changed the way they work in 2016: https://www.euromomo.eu/about-us/history

From 2016 onwards, the EuroMOMO network has been supported by and worked closely with the European Centre for Disease Prevention and Control (ECDC) and the World Health Organization (WHO) Regional Office for Europe.

rezeroed2• 6 years ago

I, Robot: stay in your homes, it is for your own protection.

Faaak• 6 years ago

Unrelated, but this new website is way better than the old one which was online just weeks ago (https://web.archive.org/web/20200323070430/https://www.eurom...).

Graphs are client-side rendered now ! And we can have the raw data too ! Great :-)

Aissen2• 6 years ago

For those interested, there's a breakdown by sub-region by French National Institute of Statistics:

https://www.insee.fr/fr/information/4470857

Needless to say, it matches closely the COVID-19 clusters. Kudos to this institution used to release yearly studies for the quick turnaround. (and all the data is available to download).

Aissen2• 6 years ago

Also of note for HN readers: in "Bouches-du-Rhone", a high-density population area, a ransomware attack has rendered the local cities unable to send the statistics for the second part of March, which they sent manually. I'm not sure how resolved this is, but it shows (yet again) how cyberattacks can have real-life impact (and this isn't a hospital…)

yummybear• 6 years ago

15-64 is a pretty wide age group covering everything from kids to senior citizens. I wonder why they havent divided it further into at least adolescent/adult and possibly +50

hrgiger• 6 years ago

Czechia is missing

tgflynn5• 6 years ago

I think it's important to note that while the cumulative excess mortality is indeed increasing (it could hardly do otherwise since it's defined as a monotonically non-decreasing function) the weekly excess mortality count sharply decreased in the most recent week (week 16).

That said the decline is so large from the previous week (9289 vs 20975) that I suspect there may be some error in the graph or the data.

EDIT: My claim that cumulative excess mortality is defined to be monotonically non-decreasing is not correct, as the below comment points out.

isoprophlex• 6 years ago

Is that true?

Excess mortality in the 0-4 yr cohort for example shows some negative number of excess deaths...

tgflynn5• 6 years ago

I was only looking at the graphs for all ages.

EDIT: Ah OK, I see what you're saying. Yes, my statement that the cumulative graph is defined to be monotonically non-decreasing was indeed incorrect because it assumes that excess mortality itself is non-negative, which is not the case (total mortality would be non-negative though). Still for the cumulative graph to decrease the weekly excess mortality would have to go into negative territory, which is far from the case for the "all ages" group.

gpderetta6• 6 years ago

as I reported else thread, I have been following these reports for a few weeks now, and there is often a sharp drop at the end which goes away on the next report. I think it is an artifact of their modelling for delayed registration that forces a regression to the mean.

tgflynn5• 6 years ago

Thanks, I suspected it was some sort of artifact.

European mortality monitoring shows increase in excess mortality

Related Stories