Thread: US excess death
1/
Rough estimation of US excess deaths, Oct 2019-Sep 2020 (“flu year”)
Impact of recent winter belongs to “flu year” Oct 2020-Sep 2021
(“Flu year” avoids arbitrarily split of winter mortality between two calendar years.)
At the end: Sweden
1/
Rough estimation of US excess deaths, Oct 2019-Sep 2020 (“flu year”)
Impact of recent winter belongs to “flu year” Oct 2020-Sep 2021
(“Flu year” avoids arbitrarily split of winter mortality between two calendar years.)
At the end: Sweden
2/
Inference derived from just 7 “flu years” (shown in rectangle) + heuristics.
Counts of all-cause deaths obtained from a table in this tweet.
Inference derived from just 7 “flu years” (shown in rectangle) + heuristics.
Counts of all-cause deaths obtained from a table in this tweet.
https://twitter.com/ProfessorAkston/status/1352371828245757955
3/
Notice a pattern: a year of “high” mortality is followed by a year of lower/similar.
Pattern disrupted in 2017/2018, severe flu season. Should have been lower/similar to 2016/17.
Estimated flu deaths: 61,000
Notice a pattern: a year of “high” mortality is followed by a year of lower/similar.
Pattern disrupted in 2017/2018, severe flu season. Should have been lower/similar to 2016/17.
Estimated flu deaths: 61,000
4/
Fit separate regression lines to 4 “high flu years” and 3 “low flu years” (until pattern disrupted).
Excellent fit for both (r^2=0.99), though few observations.
Used the equations to predict for 2017-18, 2018-19, 2019-2020.
Fit separate regression lines to 4 “high flu years” and 3 “low flu years” (until pattern disrupted).
Excellent fit for both (r^2=0.99), though few observations.
Used the equations to predict for 2017-18, 2018-19, 2019-2020.
5/
Previous pattern suggests that both 2017-18 and 2019-20 should have been “low”. (2018-19 should have been “high”.)
Therefore, predictions from “low years” is preferred.
Previous pattern suggests that both 2017-18 and 2019-20 should have been “low”. (2018-19 should have been “high”.)
Therefore, predictions from “low years” is preferred.
6/
Model estimates:
2017-18: excess 42,000 (CDC estimated 61,000 flu deaths)
2019-20: excess 254,000 (reported COVID death: ~210,000)
Translating to weeks of “normal” deaths, as suggested by @MLevitt_NP2013 : ~ 4.7 weeks.
Unaccounted excess: ~44,000
Model estimates:
2017-18: excess 42,000 (CDC estimated 61,000 flu deaths)
2019-20: excess 254,000 (reported COVID death: ~210,000)
Translating to weeks of “normal” deaths, as suggested by @MLevitt_NP2013 : ~ 4.7 weeks.
Unaccounted excess: ~44,000
7/
Next is deliberation (rounded numbers):
What is the split between excess deaths due to COVID and lockdown/panic?
Is it 44,000? (=254,000-210,000)
Some COVID deaths have been missed, but overcounting (death “with COVID”) likely much more common, given liberal coding guidelines
Next is deliberation (rounded numbers):
What is the split between excess deaths due to COVID and lockdown/panic?
Is it 44,000? (=254,000-210,000)
Some COVID deaths have been missed, but overcounting (death “with COVID”) likely much more common, given liberal coding guidelines
8/
Therefore, true COVID deaths < 210,000 reported.
Research will tell how many. For now, assume ~85%:
True COVID deaths: ~180,000 (70% of excess)
Lockdown/panic deaths: ~75,000 (30% of excess)
Therefore, true COVID deaths < 210,000 reported.
Research will tell how many. For now, assume ~85%:
True COVID deaths: ~180,000 (70% of excess)
Lockdown/panic deaths: ~75,000 (30% of excess)
9/
Estimated COVID deaths 2019-20 vs. estimated flu deaths 2017-18:
180,000/60,000 = 3.
COVID: 3 x more severe in the US than recent severe flu year. Order of magnitude seems reasonable given IFR & R0
Lockdown/panic made it 4+ more severe.
The toll in flu year 2020-21? By Oct
Estimated COVID deaths 2019-20 vs. estimated flu deaths 2017-18:
180,000/60,000 = 3.
COVID: 3 x more severe in the US than recent severe flu year. Order of magnitude seems reasonable given IFR & R0
Lockdown/panic made it 4+ more severe.
The toll in flu year 2020-21? By Oct
10/
And what happened in Sweden, without lockdown & panic?
Plus graph: updated comparison of current winter mortality in Sweden vs. 2017/18.
And what happened in Sweden, without lockdown & panic?
https://twitter.com/prof_shahar/status/1346631983896543233
Plus graph: updated comparison of current winter mortality in Sweden vs. 2017/18.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
