We found in the last post that if you compare unvaccinated hospitalizations to unvaccinated positive tests, you find a similar proportion in December 2021 to that of December 2022. As a statistic, it’s been consistent. The same can not be said for statistics on vaccinated hospitalizations.
This is expected, to some degree. If the hospitalization rate were not improved for the vaccinated, there would be no upside at all to these treatments. But efficacy against hospitalization should be a consistent effect, or one that declines steadily as the vaccine’s effects “wane”. That’s what we saw with the vaccine’s steadily declining efficacy against infection.
But for hospitalizations, historical efficacy is anything but steady.
It’s obvious from looking at the age groups that this statistic was unreliable until late October at least, when the efficacy values start to converge for the first time. Data this noisy and implausible shouldn’t be used for anything. Unsurprisingly, it was used with great enthusiasm to support the autumn vaccination campaign.
Just combine this garbage data over 28 days to smooth its rough edges and age-adjust it so that the impossible 80x efficacy seen for 40-49 year olds boosts the humble 10x efficacy of 70-79 year olds, and out comes more amazing garbage — 35x efficacy! — for public consumption.
You might guess that one factor in the volatility of the hospitalization efficacy stat is the sparsity of data, and you’d be right. For much of the past six months, there just haven’t been enough SARS-CoV-2 positive hospitalizations to produce a stable efficacy statistic. But the reason there were so few hospitalizations among the vaccinated is rather extraordinary. They weren’t testing them.
Hospitals were required to test every unvaccinated patient regardless of what they came in for, but not required to test doubly vaccinated patients. With a methodology as rigged as that, it’s no surprise that the unvaccinated group had wildly more positive results. You can’t be tested positive if you aren’t tested at all.
This has significant implications for the hospitalization number as a whole. When the vaccines became available, everyone was being tested for SARS-CoV-2 who went into the hospital. It was only later, during the autumn Delta wave, that doubly vaccinated patients were excluded. A similar situation applied to testing health care workers.
Fewer known SARS-CoV-2 infections among patients and hospital workers meant fewer alarm bells going off, less panic, and less of a crisis. We were allowed a relatively normal autumn in Quebec (if showing proof of vaccination to access a private business fits your idea of normal). But to some extent, we owe that period of calm to the fact that many infections were allowed to pass underneath the radar.
Uniformity in testing procedures was restored on December 23rd, long after it had become clear that vaccinated people were catching SARS-CoV-2 and were quite capable of bringing it into hospitals.
With this history in mind, we can make some sense of the crazy twists and turns of observed vaccination efficacy against hospitalization.
The efficacy stat briefly got its act together in late October, only to loose its marbles again in mid November. There’s nothing obvious to attribute that first correction to. In October, infections within the population were relatively flat and slightly down from their unimpressive September peak. One factor is that hospitals weren’t forbidden from testing vaccinated patients, and some hospitals rightfully set their own policies to test patients regardless of vaccination status. Perhaps infections were concentrated in those hospitals?
In any case, hospitals managed to produce a plausible looking efficacy statistic for a few weeks until infections picked up again in mid November, at which point SARS-CoV-2 positive unvaccinated 30-59 year olds were again counted widely out of proportion.
It’s only when universal testing was finally reinstituted in hospitals province-wide on December 23rd that the efficacy age groups re-converge and begin to move in plausibly straight lines. Everything before that point, unfortunately, is garbage.
Now that vaccine efficacy has come down to earth, we still have the problem of the Legault government using hospitalization numbers to badger and divide the public. In the next item we’ll look at what SARS-CoV-2 hospitalizations actually represent, and why they were never the right category for monitoring the impacts of this virus.