I'm pretty unconvinced this thought is well-baked. It's more like, well, whoever thought it is baked. According to the CDC [0], Omicron had 9 deaths per 1000 cases, Delta 13, and original COVID 16. While I admit the hospitalization numbers are slightly more friendly for Omicron, it's clear no work was done to separate vaccination status in this data, and it'd take someone more dedicated than I to argue Omicron really exhibited markedly different characteristic than original COVID to a degree necessary to even make this plausible
It says the risk of death with Omicron is 69% lower than with Delta. Among the unvaccinated, the risk is 80% lower than with Delta.
These numbers jibe with the statistics on case rates and mortality that we've seen over the last three months, when Omicron has been dominant.
I suspect the difference would be much more pronounced if we were to account for asymptomatic cases. Looking at wastewater measurements, which have been a good proxy for (and predictor of) case rates, we see far fewer confirmed cases, relative to virus levels, during the Omicron phase than we saw during prior variants. The ratio of deaths to wastewater virus levels is about 1/10th of what it was with Delta.
There is one major bias in the sort of studies you just linked to that can't really be smarted away. When looking at things like a mortality rate, you need to somehow determine how many people are infected. Imagine there were some weird disease where it was completely asymptomatic in 99% of cases and fatal in 1%. The vast majority of contemporary COVID related studies would claim this disease would have an extremely high mortality rate, far higher than 1%.
The reason is because diagnosis is almost never done randomly, but instead relies on different avenues like hospitalization data. So you tend to already be biasing yourself to severe outcomes because milder cases are not going to end up getting diagnosed. In omicron's case this effect has been extreme as things like sewage samples showed dramatically higher rates of of the disease than were being officially reported. In the study you mentioned it determined diagnosis using:
"CDC used data from three surveillance systems to assess U.S. disease related to COVID-19 during December 1, 2020–January 15, 2022. COVID-19 aggregate cases and deaths reported to CDC by state and territorial health departments¶ were tabulated by report date.* ED visits with COVID-19 diagnosis codes were obtained from the National Syndromic Surveillance Program (NSSP).†† Hospital admissions and inpatient and ICU bed use among patients with laboratory-confirmed COVID-19 were obtained from the Unified Hospital Data Surveillance System."
So your numbers are biased to heavily rely on things like hospitalization data, which is going to make the numbers borderline useless for trying to evaluate the overall mortality rate. To be fair I'm not really attacking the study either. Like I said this is a problem that really can't be solved in any way other than an involuntary lottery with mandatory testing + reporting + profiling/classification, which is something I'd expect to see in China, but not the US - for better and for worse.
I'm with you. If conspiracy theories required well-baked thoughts, then we wouldn't have things like flat earthers. Not sure how well a thought logically makes sense is in any way relevant.
What seems more important is how conveniently it can prove a point that the person already is trying to push forward.
[0]: https://www.cdc.gov/mmwr/volumes/71/wr/mm7104e4.htm?s_cid=mm...