In an article this week highlighting side-effects resulting from the AstraZeneca Covid vaccine, BBC Health Correspondent Fergus Walsh states:
It is estimated that the Covid vaccine programme prevented over a quarter of a million hospital admissions and over 120,000 deaths in the U.K. up to September 2021.
These figures come from a UKHSA Vaccine Surveillance report published on September 30th 2021. As it happens, the UKHSA estimates were actually for England rather than the U.K. and the precise numbers were 261,500 hospitalisations averted in those aged 45 and over by September 19th 2021 and 127,500 deaths averted by September 24th 2021.
The estimates were calculated by applying vaccine effectiveness rates available at the time to known vaccination, hospitalisation and mortality data. With such models, the first question to ask is whether the results pass a basic plausibility test. Let’s see.
In the first nine months of 2021, the ONS reports about 60,000 Covid-related deaths in total. If we believe vaccination prevented 127,500 deaths, then without vaccination there would have been nearly 190,000 deaths in that period.
In the first nine months of the pandemic before vaccination (i.e., March to November 2020), the ONS also reports about 60,000 Covid-related deaths. In other words, if vaccination prevented 127,500 deaths, then without vaccination the claim is that we would have seen three times as many deaths in the nine months post-vaccination as in the nine months pre-vaccination. Given that vaccination was rolled out during a period when Covid deaths were already decreasing following the huge infection wave at the end of 2020, that is more than a stretch – it is simply implausible
Consider further that we know that a large proportion of deaths in 2021 were actually in the vaccinated (the Vaccine Surveillance Report mentioned above reveals that in September 2021, for instance, only 21% of Covid-deaths were amongst the unvaccinated) and many more occurred early in the year before people had access to vaccination. That makes the potential pool of people whose lives might have been saved by the vaccine even smaller and the estimate of 127,000 deaths averted even less credible.
The problem with the UKHSA estimates is fairly obvious. The vaccine effectiveness rates they use relate to hospitalisation or death rates for the vaccinated relative to rates for those unvaccinated and not previously infected. For this and other reasons they dramatically overestimate the effect in real world populations in which significant proportions of people have immunity from a previous infection and in which (as we now know) vaccine effectiveness wanes rapidly over time.
There is another way to test the plausibility of the estimates quoted by the BBC. In January 2023, the UKHSA produced estimates of the expected number needed to vaccination (NNV) to prevent a hospitalisation or death. These were also modelled estimates, but in contrast to the earlier approach, they used updated vaccination effectiveness rates and, crucially, took account of the fact that effectiveness waned over time.
For the primary vaccination rollout, UKHSA reports NNV for hospitalisations but not for deaths (it published estimated NNV for deaths only for the booster period). I’ve taken the NNV to prevent a single hospitalisation for the primary vaccination programme and applied these figures to the total numbers vaccinated with two doses.
Note UKHSA provides separate NNV estimates for those “in a risk group” and “no risk group”, but equivalent vaccination numbers are not easy to identify for these separate categories. For this reason, I have used the reported NNV for both groups combined. I’ve also estimated hospitalisations prevented for those aged 40-plus rather than over-45s on which the 262,000 estimate was based. This is because the NNV is only reported for the whole 40-49 age group. Finally, my estimates go to the end of September, slightly longer than the period used by the UKHSA.
The NNV estimates indicate that just under 32,000 hospitalisations might have been prevented by vaccination by the end of September 2021. The (outdated) estimate reported this week by the BBC’s Fergus Walsh was eight times higher.
We can do a similar exercise for deaths prevented. For this we only have the NNVs reported by UKHSA for the 2023 booster programme. For hospitalisation, the NNVs are nearly three times higher in the booster period than for the primary dose (reflecting lower effectiveness of boosters). If we similarly assume the NNV to prevent one death was three times higher in the 2023 booster period, then we arrive at a figure of about 12,500 deaths prevented by the end of September 2021, just one tenth of the estimate reported by Fergus Walsh.
It is worth noting that the UKHSA still derives its NNVs from modelling and they may not reflect real-world population effectiveness. But at least the estimated numbers of hospitalisations and deaths prevented by vaccination are plausible in the context of actual data.
Presumably the BBC Health Correspondent is aware of the later UKHSA estimates of numbers needed to vaccinate. Why then is Fergus Walsh still quoting estimates of hospitalisations and deaths prevented by vaccination that are not only implausible but also way out of date?
David Paton is Professor of Industrial Economics at Nottingham University Business School. He tweets as @CricketWyvern.
To join in with the discussion please make a donation to The Daily Sceptic.
Profanity and abuse will be removed and may lead to a permanent ban.