Here are graphs from the CDC of deaths by vaccination status (black line unvaccinated, dashed blue line vaccinated but not boosted, and solid blue line boosted), with December being of particular interest due to the arrival of the omicron variant:
For younger and middle-aged people, the risk of dying from omicron was so low that the advantage of a booster shot isn’t visible on the graph, at least through December. (Note that deaths prior to omicron tended to have about 3 week lag after becoming a case count, so extremely few people died of omicron in December, unlike today, when deaths remain moderately high despite case counts falling rapidly this month.)
This doesn’t mean the risk of younger people going without a booster doesn’t exist in a relative sense, but it’s minor in an absolute sense.
Consider a toy model: Intervention X cuts the chance of Bad Thing Y happening in half.
If the chance of Bad Thing Y happening without doing X is 10%, then doing X cuts your chance of Y happening to you by 5 percentage points from 10 to 5.
But the importance of doing Intervention X depends not just on its relative efficacy against Y, but also on the absolute frequency of Bad Thing Y.
For instance, if the chance of Bad Thing Y is only 0.1%, then doing X cuts your chance of Y by only 0.05 percentage points. You might well choose to do Intervention X with 50% efficacy if the chance of bad thing Y happening to you without it is 10% but not if it is 0.1%.
It’s hard for humans to keep both relative and absolute efficacy in their heads at the same time.
It’s even harder when:
This stuff is complicated, so let me try to explain what is apparently going on.
Here’s another new CDC graph, this time using COVID case counts rather than deaths:
The efficacy of having two shots of the vaccine at preventing being counted as infected clearly dropped sharply in December when omicron arrived.
On the other hand, the absolute chance of becoming an omicron case count went up so much that the efficacy of the booster is clearly visible.
For young people 18-49, the efficacy against being counted as a case of being vaccinated dropped to about 50% and the the efficacy of the booster relative to two-shot vaccination was even lower.
What about for middling-bad outcomes like being hospitalized? Here’s another new CDC graph from a different document (thus with different colors).
For 50-64 year olds during the omicron-heavy week ending Christmas Day, the rate of hospitalizations for the unvaccinated was 73 per 100,000, for the two-shot crowd 9, and for the boosted 2.
So, the relative efficacy of the booster for the middle-aged against the unpleasant outcome of being hospitalized was pretty strong, but the absolute value wasn’t huge.
An even more complicating analytical factor is that while the efficacy of the original 2-shot vaccination against being a case tends to both be reduced under omicron and wear off over the months, it still seems to leave long-term capabilities to resist COVID doing its worst to you.
In other words, the short-term antibodies from the 2-shot vaccination that might keep you from becoming symptomatic seem to fade out over only a number of months, rather than the year or longer as was hoped in late 2020. On the other hand, the 2-shot vaccination seems to leave behind enough capability to mount a longer-term defensive reaction that continues to cut down notably on the chance of severe illness.
So it appears that in the omicron era, being vaccinated is pretty good and the Israeli obsession with multiple boosters might be over the top for the non-elderly.
But it’s a complicated thing to think about, so maybe I’ve got it all wrong.
