Another Bad Vax Effectiveness Study

But not as bad as some people say.  The study was about the new bivalent vaccine and was conducted among Cleveland Clinic employees, many of whom likely had frequent exposure to CV-19. Among these over 50,000 employees, 41% had a prior documented CV-19 infection with 24% having had an Omicron infection.  (In reality, the infection rate was likely much higher) and 83% were vaxed (interesting that even a pre-eminent medical institution couldn’t get a higher vax rate among employees).  The headline finding is that including a bivalent booster (one that addresses original and Omicron CV-19 strains) had at most a 30% effectiveness, and that in the first weeks following vaccination with the booster, since this study didn’t even start until September of this year, when the bivalent vax became available.   (Medrxiv Paper)

The introductory paragraphs of the paper were clearly written by someone who either stayed up for three days without sleep or was under the influence of mildly hallucinogenic drugs.  Raving about the supposed effectiveness of the vax is not a sign of rationality at this point, so it alerts one to the likelihood of bias in the methods and interpretation.  And here the role of prior infection gets significantly underplayed.  People who had never been infected (that is, never reported a positive CV-19 test) were at highest risk of an infection during the study period, with those with any prior infection having a lower risk, and the lowest risk being a prior recent Omicron variant infection.

Interestingly, the risk of an infection during the study period also appeared linked to number of vaccine doses received–the more doses, the more likely the person would become infected.  Not clear to me, however, that the potential confounding interactive effect of prior infections, time from prior infections, prior vax doses and time from prior vax doses was fully considered.  And one general observation would be that the more vax doses a person had, the less likely that they had any prior infection, meaning they are potentially more susceptible to one during the study period.

The authors state that there is little reason to think prior infections were undetected at different rates among those who did and didn’t receive a bivalent vax, but don’t assess that by antibody testing, and they appear to ignore the possibility, even likelihood that those rates did differ among those vaxed and unvaxed before the bivalent boosters. There is a lot of hypothesizing about potential behavioral differences between vaccine seekers and non-seekers, which I always view as dangerous to a rigorous analysis.  But one important potential behavioral impact is ignored–if you know you had a prior infection(s), you may be less likely to get vaxed.

There is reason, as I have repeatedly said, most recently in a post warning against continued boosting, to believe that too much poking with the same or similar antigens may actually impair an immune response.  But I am dubious that there actually is an increased risk of infection among people who are vaccinated that is due to the vaccines.  At this point it is very hard to find pure cohorts in which to make comparisons, because so many people have been infected, and so many vaxed.  The residual cohorts of either uninfected, unvaxed or both, are very small and likely have some confounding characteristics.

What the study pretty clearly shows is the limited value of additional boosters.