How is it that more vaccinated Brits are dying of the Delta variant than unvaccinated?

You may have seen a report doing the rounds of the internet claiming that significantly more of the people dying of the Delta variant of COVID in the UK are vaccinated than unvaccinated. It's hard to pin down where the report came from - some articles say the low-life UK newspaper, the Mail on Sunday - but it has gone viral on Facebook and Instagram and apparently on some pro-Trump websites. It provides yet more ammunition for anti-vaxxers and the vaccine-hesitant.

But can it be true? It seems improbable, especially given the claims from the US that COVID deaths are almost all among the unvaccinated. Luckily, it is not hard to find refutations, or at least plausible explanations, from more reputable outlets like The Guardian and the BBC, as well as fact-checkers like Full Fact.

The original stats in which the claims are made come from the latest Public Health England (PHE) Technical Briefing (Table 4). It says that, of the 117 people who have died from the Delta variant in the UK, it was unknown whether or not three of them had been vaccinated (what? why?). But, of the rest, 70 (61%) had received at least one dose of the vaccine, and 50 (44%) had received both, while 44 (38%) had not been vaccinated.

This might seem worrying at first glance, and lend credence to viral claims that the vaccines are ineffective or even dangerous, and that they make people more likely to die, at least from the Delta variant (which is now by far the most common variant worldwide, and particularly so in the UK). But, as more sober analysis reveals, it's actually not that unexpected.

For one thing, as the BBC points out, only 8% of recent cases have been among the fully vaccinated (although that's actually higher than I would have expected), so the vaccines are indeed quite effective at preventing the virus. But most of the fully-vaccinated people are among the over-50s, the very cohort that is significantly more at risk of dying if they do catch the virus, which is why they appear over-represented among the death statistics. An Israeli study shows just how much chronic "old people" comorbidites like hypertension and diabetes skew the death rates from COVID (only 6% with no comorbidities died, according to that study).

The vaccines are not perfect and, if everyone were fully vaccinated, then 100% of the deaths would be among the vaccinated. Britain has a pretty good vaccination rate, particularly among older adults, but some will always slip through the net and get infected and some of those will become hospitalized, and some of those will ultimately die. If the vaccines prevent 95% of deaths, as it is estimated they do, one in twenty will still succumb, more in the more deprived ethnic areas.

I kind of understand the logic of this as I read it, but it is really not easy to understand, and it does still seem a bit counter-intuitive. If a double-vaccination is supposed to reduce the likelihood of catching COVID by over 90% (and the likelihood of hospitalization and death by much more than that), then surely we should be expecting a figure closer to 1% or less for deaths among vaccinated people.

Also, I don't understand how these British statistics square with the American ones that suggest that very few unvaccinated people are dying (which is more in line with what I would have expected logically). The only difference I can see is that the US figures are over the last six months, during most of which a much smaller proportion of the population were vaccinated, rather than the much more recent past in the case of the UK figures.

UPDATE

Different figures were given in a recent news conference by Sir Patrick Vallance, Chief Scientific Advisor to the UK government, who reported (after a Twitter correction for an error in his verbal report) that 60% of British COVID hospitalizations were unvaccinated, suggesting that 40% were partially or fully vaccinated. This is still a hugely larger number than would be expected, and at odds with the American stats. 

Vallance explains it rather cursorily by saying that the vaccines are not 100% effective, and that, as more and more people are vaccinated, this kind of death level is not surprising. Well, I'm surprised.

UPDATE UPDATE

This graphic, also apparently based on UK government data, should give us more confidence that things are going as they should.

It shows new cases and deaths from the first 50 days of the second and third waves of the pandemic in Britain. It shows how deaths have stayed very low during the third wave (right-hand graph) compared to the second wave (left-hand graph), even as cases sky-rocket once more.