> There are other scenarios, e.g.
https://www.ox.ac.uk/news/2021-02-09-common-asthma-treatment... shows a cheap widely available steroid inhaler [...]
That is a good point, I should have defined the category more broadly as 'pharmacological interventions,' of which vaccines make up the biggest and best-studied subcategory (to date).
> That's ... not been shown. The Israeli study I linked to above (a report of it) says this assertion might be wrong for Pfizer for the sizable group of 16-30 ; The British think that's not true for AZ under age 30 ; many countries in the EU think that's not true for AZ under the age of 50 ; Norway thinks that's not true for frail elderly.
Yes and no.
Yes, E.g. for AZ there will be a threshold (say 30) at where Covid risk equals vaccine risk. At that point, both risks are very small, to the point that it's more dangerous to drive to the appointment by car.
No, when EMA says no AZ under 50 (in Germany it's 60) that's because for that age group we have other vaccines available. Still, even for a 40 year old female it is safer to get AZ than Covid.
For that reason current vaccines are safer for most people but not all. Even when they are not safer than Covid, data suggests they are still very safe.
> And basically, we only have short term safety data (pfizer: less than a year for 20K people, less than 6 months for the rest) [...]
I believe you do know that Pfizer/Biontech, Moderna, AZ all started in 04/2020, so we have data for 12m+. As a comparison safety data for Covid is only 6m older.
> I linked a BMJ+Science article that showed for a similarly EUAd vaccine, "Pandemrix" in 2009 that it took over a year to figure out that it was causing narcolepsy and worse than the flu it was supposed to stop.
Yes, Pandemrix is still remembered, especially in Northern Euope, and one reason why regulators are have been super careful, to the point where it didn't make sense to halt vaccination campaigns, see e.g. AZ blood clots. In general we expect side effects of vaccines to have an onset shortly after vaccination, and that has been also true for Pandemrix. However to detect two rare events, say narcolepsy and 5x narcolepsy, you need a lot of data. In the context of H1N1 and narcolepsy, this paper is super interesting [1] https://med.stanford.edu/content/dam/sm/narcolepsy/documents...
> It could be that for those 80%, that response would be enough to stop the virus before it can get systemic traction, [...]
That's not the conclusion of the paper and hopeful speculation. If you mean by 'systemic reaction' the replication of the virus in the host, it's demonstrably false.
I totally get the concern about 'unknown unknowns' but looking at the safety data you are currently engaging in lifestyle choices or are taking medication that are way riskier than current Covid vaccines.