Enrico Fermi considered broad questions that no one knew the answer to, but that you could get good guesses on — perhaps being within just 1% to 2% of the truth of the matter.
Francis Galton recorded guesses on the weight of an ox at a 1906 fair, and the average of the guesses was just 1% from the truth of the matter (real weight of dressed ox came to 1,198; while guess averages, mean and median, were both under 10 pounds off).
Applying Fermi (“sniff test”) analysis to the accumulated excess deaths reveals that observations are more than double what is possible, if you use the already-questionable assumption that a respiratory infection caused all (100%) of the excess deaths.
Here is the accumulation (in total excess deaths per million) to the end of May 2022:
And here is the same graph with notes on it:
It appears for it to be impossible that COVID, in all of its variants, could explain even half (50%) of the excess death recorded. Even the 99% upper bound of accumulated excess death (the light green bar above the dark green bar) was unable to reach 1,500 total excess deaths per million population.
That means that “something else” is responsible for most of the excess deaths which have been recorded.
How the Fermi Analysis was performed
Using 117 deaths from 52,000 Alpha variant infections (Page 3 of UK Tech Briefing #5), I obtained a 99% credible interval on the infection fatality rate (IFR).
Using a rough guess that the 99% lower bound IFR (0.178%) would lead to an average of 1.5 daily excess deaths per million (DEDpM) — everything else got scaled to that.
The wild type COVID strain was assumed as deadly as Alpha, while Delta involved a “lethality step-down” so that it was only 42% as lethal as Alpha — and Omicron involved a second “lethality step-down” so that it was only 31% as lethal as Delta.
You can either assume full susceptibility to the new variant each time but without increased transmissibility (which is what I did) — i.e., assuming increased transmissibility is perfectly offset by the reduced base of susceptibles — or you combine the existing scope of naturally-aquired immunity (from infection with former variants) with the new basic reproduction number for each new variant.
Reference
[787 untrained people guess to within 1% of the true weight of ox] — GALTON, F. Vox Populi . Nature 75, 450–451 (1907). https://doi.org/10.1038/075450a0
[on Enrico Fermi’s unique way of questioning] — FERMI QUESTIONS AND ORDERS OF MAGNITUDE. Patrick Liscio. November 2019 Math CoOp. https://www.dam.brown.edu/math-coop/presentations/Fermi.pdf
[As of 19 Jan 2021, from 52,000 COVID infections with Alpha (SGTF) variant, there were 104 deaths by Day 28 (corrected to 117 total deaths, by using Linton et. al., leads to IFR=0.225%)] — Page 3. Epidemiological findings. UK Technical Briefing #5. https://www.gov.uk/government/publications/investigation-of-novel-sars-cov-2-variant-variant-of-concern-20201201
[From ~700,000 cases confirmed by sequencing, Delta was 42% as fatal as Alpha] — Table 3. Number of confirmed and probable cases by variant as of 11 October 2021. UK Technical Briefing #25. https://www.gov.uk/government/publications/investigation-of-sars-cov-2-variants-technical-briefings
[From 1.5 million confirmed cases, Omicron is 31% as lethal as Delta, making Omicron much safer than seasonal flu (much safer)] — Nyberg T, Ferguson NM, Nash SG, Webster HH, Flaxman S, Andrews N, Hinsley W, Bernal JL, Kall M, Bhatt S, Blomquist P, Zaidi A, Volz E, Aziz NA, Harman K, Funk S, Abbott S; COVID-19 Genomics UK (COG-UK) consortium, Hope R, Charlett A, Chand M, Ghani AC, Seaman SR, Dabrera G, De Angelis D, Presanis AM, Thelwall S. Comparative analysis of the risks of hospitalisation and death associated with SARS-CoV-2 omicron (B.1.1.529) and delta (B.1.617.2) variants in England: a cohort study. Lancet. 2022 Mar 16:S0140-6736(22)00462-7. doi: 10.1016/S0140-6736(22)00462-7. Epub ahead of print. PMID: 35305296; PMCID: PMC8926413. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8926413/
[“Time-to-death” model for COVID; using the 95% UB of both the mean and SD of the lognormal model that fit the actual deaths best] — Linton NM, Kobayashi T, Yang Y, et al. Incubation Period and Other Epidemiological Characteristics of 2019 Novel Coronavirus Infections with Right Truncation: A Statistical Analysis of Publicly Available Case Data. Journal of Clinical Medicine. 2020 Feb;9(2). DOI: 10.3390/jcm9020538. PMID: 32079150; PMCID: PMC7074197. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7074197/