Hacker News Reader

Comment by sgc | Hacker News Reader

It's obviously not well tested enough to not get pulled out of use despite the pandemic, so more testing as we move forward seems appropriate. OP has a point, although the form of testing will probably be one less likely to cause people to avoid the vaccine.

Also, what are the odds of a 1 in 10k concern not being picked up as such in a 40k sample size? It would look like background noise if it was only one case in that particular group.

To be clear, I am very much pro continuing use of the vaccine, but it's not the case that every possible 1 in 40k or less concern has been discovered since we had one such group go through a controlled study. Since the odds of COVID causing serious problems are far, far higher than that, it is still a win, but we should of course learn as we go.

View on HN

> It's obviously not well tested enough to not get pulled out of use despite the pandemic

Or the people making the decision are incredibly risk averse and don't mind an extra few hundred or thousand Covid deaths to save a single person with a blood clot.

>Also, what are the odds of a 1 in 10k concern not being picked up as such in a 40k sample size? It would look like background noise if it was only one case in that particular group.

It's a specific type of blood clot that is incredibly rare and occurs more often in young women. If it had a 1 in 10k rarity it would have probably been found

I never said anything like "the blood clot risk might be 1 in 10k". I said that there can be things that come up a lot more than 1 in 100k that were not found in the trials (not referring to anything specific, just unknown unknowns), and it is not appropriate to stop looking for side effects just because we have done some testing.

I’m not sure whether you’re arguing that this wouldn’t be noticed or if we cannot measure it statistically.

For the former, the side effect is blot clots with thrombocytopenia. That combo is very rare, is indicative of an autoimmune etiology, and would stand out to healthcare providers and be reported. Even if there were a few cases that were not identified or reported, it would not lead to a 2x order of magnitude difference in the reported incidence as you suggest.

If your argument is a statistical one about the study being underpowered, then show your math. Because this is definitely not a small n.

Otherwise, be careful about presuming that something in a domain unfamiliar to you is equally unclear to experts in that domain.

Its not 1 in 10k. You're overstating the risk. It's because its not 1 in 10k it was only picked up after mass vaccination began.

See? you're inflating risk and misunderstanding consequences. How many people die, if nobody is vaccinated compared to a 1 in 100,000 risk of clotting? A lot more than 1 in 100,000 die from covid.

The risk of other non-fatal clotting increased 50% during the trials. The risk of this one specific type is 1 in 200,000+ which can result in death.

As I said to someone else, I was referring to unknown unknowns and the need to stay vigilant, not to blood clot risks specifically. And I stated expressly that I think the vaccinations should continue. But let's not just stop learning. It's not a big ask.

Reply on news.ycombinator.com