Hacker News Reader

Comment by NonContro | Hacker News Reader

Not really, because if there are long-term side effects, we need to have a control group.

For example, the vaccine used for the 2009 Swine Flu:

https://en.wikipedia.org/wiki/Pandemrix

"Pandemrix was found to be associated with narcolepsy from observational studies, increasing the risk of narcolepsy by 5-14 times in children and 2-7 times in adults. The increased risk of narcolepsy due to vaccination in children and adolescents was around 1 incident per 18,400 doses."

If you don't have a large section of the population unvaccinated, you can't conclusively prove that the vaccine was the cause.

View on HN

“Problems you want to have“ are still problems, after all.

It’s just, which would you rather have: a perfectly run trial because your vaccine efficacy is so low that your control group doesn’t feel incentivized to get vaccinated? Or a vaccine that works well enough to quell a pandemic, causing your control group to jump ship and potentially lessen your statistical power?

There was a comment here that got deleted. It had something cool I haven't seen before: a risk calculator/estimator for Covid (both death and hospital admission), for the UK:

https://qcovid.org/Calculation

Interesting. Especially that it uses a number of different "properties" (e.g., post code) to "fine tune" the calculation.

Either way, the higher-order effects will continue and cascade in perpetuity. The world is not a well-controlled experiment, and if you think there is an obvious “better” choice here which can be separated from all the context, then you have no idea how the world works.

and if you think all choices are obviously bad (because there's no discernable 'better'), then i don't want you running a public health office.

(disclaimer: i've run experiments with patients, i'm well aware of how messy the world is)

> i'm well aware of how messy the world is

No, you are not. Not even close.

In this case, there is a suspected autoimmune link between the similarity of the inactivated viral protein and a brain protein that is associated with sleep regulation. The real question to ask isn’t whether the placebo group was better off, but whether people infected with Swine Flu had a lower risk of narcolepsy than vaccine recipients.

Not necessarily true. The incidence rates of those are well-known. Rarer side effects are impossible to find by comparing vaccinated/unvaccinated in a phase 3 and need larger groups to be found. That’s routine, though. Look at the VITT for the Adenoviral vector vaccines, no strict control group was necessary to identify that.

Reply on news.ycombinator.com