is low delta CFR a function of vaccine efficacy?

a look at the UK data

many have claimed that the lower CFR of delta is caused by more vaccines in the overall population generating and adding immunity.

this seemed a plausible idea, so i pulled the UK variants of concern data on delta as far back is it goes and plotted it against the fully vaccinated rate (to give vaccines maximum advantage and steel man the case by ignoring the period of increased risk post dose one)

if vaccines are causing CFR to drop, we should see CFR drop as more people reach full vaccination. instead, we see the opposite: as fully vaccinated rose from 29.5% to 65.85%, CFR rose from 0.1% to 0.4%.

this is precisely the opposite of what one would predict if vaccines were having a protective effect for delta on a societal level and (unfortunately) looks to bolster the “vaccinated superspread hypothesis” with regard to this variant.

in general, one would expect CFR to be dropping in any event as more vulnerable cohorts are depleted and more natural immunity comes in to being from exposure and recovery.

as the shift to delta drives overall CFR down we do see the overall CFR of covid dropping in the UK, but this appears to be a mix issue in terms of dominant strain, not actual attenuation BY strain and is actually masking a simpson’s paradox.

the overall CFR is dropping, but the CFR of individual strains is flat or up. it’s just that shifting away from higher CFR’s (like aplha currently at 1.1%) to delta (0.4%) more than swamps the rise of delta from 0.1-0.2% to 0.4%.

and i think this is why a lot of people are reaching the wrong conclusion here. CFR on delta appears considerably lower than variants like alpha (64% lower CFR, basically 1/3 the lethality). but the shift in prevalence from others to delta is greatly reducing overall CFR.

the easiest way to think about this is imagine 10 bags of flour. some weigh 10 pounds, some weigh 100 pounds. if it’s 5 and 5 your average bag is 55 pounds. but if we up the weight in both bags 10% so they are 11 and 110 pounds, but shift the ratio from 5 and 5 to 6 and 4, avg. weight drops to 50.6 pounds despite every bag being more full.

that’s what’s happening here.

but the CFR of delta itself is rising and is up 100%+ since mid may.

and the vaccination rate has more than doubled in that period and is up 35 full percentage points.

this also seems contrary to the “delta has high CFR in the unvaccinated” claim as if this is so, shrinking that cohort should reduce overall CFR. instead, it rose.

further, claims that this is “all driven by the unvaxxed” do not align with the data.

in the UK over the last 2 weeks, we saw 596 VoC deaths. 450 (76%) were among the vaccinated.

this is, interestingly enough, a sign that vaccines ARE showing efficacy against death. but it implies something nasty going on elsewhere, like intensified superspread from the vaccinated as they carry and spread covid, perhaps as a newly potent symptomatic vector as laid out above in the vaccinated superspread hypothesis link.

so, while it seemed a useful idea to investigate and possibly a valid criticism, having now looked at the data, i’m not finding it compelling.

delta’s CFR does not look like an artifact of vaccines, at least not on a societal level.