How tactful, "we were unable to find that the vaccine has been effective". Though Gawd knows we tried. Heh, heh.
Had they found a 26% relative benefit, it would have been meaningless because of unmeasured confounders. So, to be unbiased, I must say that a relative detriment also is unreliable due to potential confounding by indication. These are Cleveland Clinic employees and 82% got the flu shot. A small minority (18%) may be assumed to be different somehow. Maybe they work in less-public roles and such people did not get the shot because they correctly estimated that they were at much less risk of exposure than their colleagues, or because unvaxxed patient-care staff would be punished by being made to mask all winter and they wouldn't. You can't correct for these things.
The upshot: ONLY randomized placebo-controlled trials of preventive interventions can provide evidence of benefit.
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
branchandroot
Re: 24-5 flu vaccine study
Date: 2025-04-09 06:31 pm (UTC)Had they found a 26% relative benefit, it would have been meaningless because of unmeasured confounders. So, to be unbiased, I must say that a relative detriment also is unreliable due to potential confounding by indication. These are Cleveland Clinic employees and 82% got the flu shot. A small minority (18%) may be assumed to be different somehow. Maybe they work in less-public roles and such people did not get the shot because they correctly estimated that they were at much less risk of exposure than their colleagues, or because unvaxxed patient-care staff would be punished by being made to mask all winter and they wouldn't. You can't correct for these things.
The upshot: ONLY randomized placebo-controlled trials of preventive interventions can provide evidence of benefit.