I am trying to reproduce the curves of Euromomo with data from Eurostat, but the curves of excess mortality for the age range 0-14 look very different from the ones of Euromomo. I guess that this is due to the fact that I did not normalized the raw mortality data from Eurostat as you did.
Unfortunately I can't attach a picture here. Is there a way to send it to the email of your substack? I guess that's not publich for spam reason....you DM me on twitter
Wonderful work! Please can you look at the euromomo mortality stats for 15-44 age range? It looks quite high, but it is not covid. And it keeps rising.
Late to the party and I don't know if it matters, but there are potentially uncertainties in the numbers from Sweden. Wether these would skew in any specific direction I'm not in a position to know.
The reason is how deaths (not just Covid but generally) are registred/reported. Covid deaths stands acc. to national health care system at 15 000 rounded, 80%+ elderly a/o with known comorbidities. However, the the Ministry for Social Security which also tracks mortality data for certain causes of death gives the number of Covid deaths as approx.10 000 rounded.
Furthermore, a person dying with the virus present may be registred as a Covid death, even if actual cause of death is complex or something else entirely. Most of the dead are elderly in nursing homes, and since many nursing homes regularly mistreat, malnourish and dehydrate the people in their care (this is an ongoing national scandal kept very quiet) and give them morphine instead of care, cause of death is often given as "general systems failure".
Autopsies and actual investigations in cause of death are only performed where there are uncertainties of concern (some diseases may make autopsy mandatory, I'm not up to date on the exact rules - you'd have to ask someone working in healthcare) or suspicion of crime.
As I said, I have no knowledge of if or how this skews the numbers. Bad data and muddled reporting of data has unfortunately gone from an aberration in the eighties to standard practice nowadays, not only in healthcare but in every state system as they started to fail around 2005.
A possible example of how our data-gathering services's bosses have weird priorities: crime figures and related data are kept at as abstracted a level as is possible without it looking as a coverup - meanwhile the Department of Agriculture is trying to create a national registry for people having chickens (even non commercial, as few as two or three, cats and rabbits).
So any numbers coming out of Sweden nowadays when we are rapidly becoming a failed state must be scrutinised as if we were an old Soviet block state.
I only use all cause deaths (not death categories). I don't think that this is by any means possible to get wrong. Counting deaths should work well. It would be easy to see if they are missing. Yes, the current last weeks do change, but it settles at some point.
Just a couple of notes. Would be nice to see a full "winter season" say June-May analysis. Strange, but in my analysis and over at "systems perestroika" calendar year and winter season numbers vary significantly.
Another level to analysis would be "place of death" and on addition some measure of health system capacity
Not absolutely sure which though. Maybe spare some index of accessible spare capacity, at peaks.
Also can you give an indicator of how much , if at all, absolute excesses vary between methods?
Nice job!
Are you able with your tool to track excess mortality per week?
That could be interesting to understand for example what happens from week 22 of 2021 in Europe, for the age range 0-14, or for the age range 15-44:
https://joejoejoe.substack.com/p/strange-data-from-europe
I am trying to reproduce the curves of Euromomo with data from Eurostat, but the curves of excess mortality for the age range 0-14 look very different from the ones of Euromomo. I guess that this is due to the fact that I did not normalized the raw mortality data from Eurostat as you did.
Yes we are looking into that. So far no clear signal on that issue. Even going male only for 20-25.
Unfortunately I can't attach a picture here. Is there a way to send it to the email of your substack? I guess that's not publich for spam reason....you DM me on twitter
I wrote you a DM you on Twitter!
USA 25-44yrs seems to be elevated starting april
Wonderful work! Please can you look at the euromomo mortality stats for 15-44 age range? It looks quite high, but it is not covid. And it keeps rising.
Same USA 25-44 yrs. Jumps up in april. Insurance professionals saying its unprecedented.
Late to the party and I don't know if it matters, but there are potentially uncertainties in the numbers from Sweden. Wether these would skew in any specific direction I'm not in a position to know.
The reason is how deaths (not just Covid but generally) are registred/reported. Covid deaths stands acc. to national health care system at 15 000 rounded, 80%+ elderly a/o with known comorbidities. However, the the Ministry for Social Security which also tracks mortality data for certain causes of death gives the number of Covid deaths as approx.10 000 rounded.
Furthermore, a person dying with the virus present may be registred as a Covid death, even if actual cause of death is complex or something else entirely. Most of the dead are elderly in nursing homes, and since many nursing homes regularly mistreat, malnourish and dehydrate the people in their care (this is an ongoing national scandal kept very quiet) and give them morphine instead of care, cause of death is often given as "general systems failure".
Autopsies and actual investigations in cause of death are only performed where there are uncertainties of concern (some diseases may make autopsy mandatory, I'm not up to date on the exact rules - you'd have to ask someone working in healthcare) or suspicion of crime.
As I said, I have no knowledge of if or how this skews the numbers. Bad data and muddled reporting of data has unfortunately gone from an aberration in the eighties to standard practice nowadays, not only in healthcare but in every state system as they started to fail around 2005.
A possible example of how our data-gathering services's bosses have weird priorities: crime figures and related data are kept at as abstracted a level as is possible without it looking as a coverup - meanwhile the Department of Agriculture is trying to create a national registry for people having chickens (even non commercial, as few as two or three, cats and rabbits).
So any numbers coming out of Sweden nowadays when we are rapidly becoming a failed state must be scrutinised as if we were an old Soviet block state.
I only use all cause deaths (not death categories). I don't think that this is by any means possible to get wrong. Counting deaths should work well. It would be easy to see if they are missing. Yes, the current last weeks do change, but it settles at some point.
Hi Henning/Orwell here some excess by place from UK if you are interested
Excess deaths vs 5 yr av
2020
Home Hospital Care Other
Home
J -172 -2,687 -1,228 -47
F 477 -1,932 -497 70
M 2,480 2,129 1,666 40
A 9,695 15,830 19,296 673
M 4,830 529 7,419 41
J 3,635. -2,847 236 -345
J 3,199 -3,444 -667 -190
A 3,340 -2,196 -235 -120
S 2,890 -1,908 -396 -119
O 3,661 -292 -168 -94
N 4,121 3,611 515 -135
D 4,167. 4,878 61 -101
42,321 11,671 26,552 -329
2021
J 6,252 13,107 3,180 93
F 4,808 5,045 624 -126
M 3,086 -3,403 -2,470 -233
A 2,844 -3,932 -1,969 -112
M 2,564 -3,365 -1,167 -193
J 2,639 -2,462 -985 -209
J 3,294 -469 46 -98
A 3,159 1,259 296 -107
S 3,478 2,115 665 -82
O 3,610 1,829 480 4
35,732 9,724 -1,300 -1,061
Non-covid excess (assume Covid=excess)
2020
Home Hospital Care. Other
Home
J -172. -2,687 -1,228 47
F 477 -1,932 -497 70
M 2,473 1,809 1,656 39
A 8,152 -6,399 11,008 -4,002
M 4,290 -6,408 1,791 -321
J 3,429 -5,172 -1,072 -446
J 3,082 -4,211 -1,039 -232
A 3,290 -2,542 -383 -140
S 2,845 -2,440. -534. -131
O 3,452 -3,321 -711 -147
N 3,554 -4,770 -1,421 -329
D 3,541 -5,116 -1,925 -365
38,413 -43,190 5,643 -6,051
2021
J 4,328 -10,335 -3,412 -646
F 3,479 -7,893 -3,167 -584
M 2,610 -6,889 -3,326 -388
A 2,641 -4,793 -2,187 -150
M 2,463 -3,698 -1,240 -210
J 2,580 -2,763 -1,032 -221
J 3,181 -1,457 -63 -126
A 2,949 -916 90 -153
S 3,185 -765 332 -125
O 3,330 -912 123 -45
30,746 -40,423 -13,879-2,648
Just a couple of notes. Would be nice to see a full "winter season" say June-May analysis. Strange, but in my analysis and over at "systems perestroika" calendar year and winter season numbers vary significantly.
Another level to analysis would be "place of death" and on addition some measure of health system capacity
Not absolutely sure which though. Maybe spare some index of accessible spare capacity, at peaks.
Also can you give an indicator of how much , if at all, absolute excesses vary between methods?
I'm working on that: July to July instead of caledar year yes. Better. But I need to adapt (interpolate) the populations which are by calendar.