Did the Vaccines Really Prevent 7,000 Covid Hospitalisations and Deaths in Summer 2022?

Did the Covid vaccines prevent 7,100 U.K. Covid hospitalisations and deaths in the summer of 2022? That’s the claim being made in a new study published in the Lancet. Here’s the report in the Times (it was also splashed in the Mirror and Guardian and on the BBC).

Thousands of hospital admissions and deaths might have been averted in the summer of 2022 if everybody had been fully vaccinated against Covid, a landmark study has concluded.

NHS data from all 67 million people in the U.K. was brought together for the first time to analyse the benefits of vaccines, in a significant moment for medical research.

Scientists found that 7,100 hospital admissions and deaths might have been prevented if everybody had had all their vaccinations and boosters.

The study, published in the Lancet, was based on anonymised health records covering the entire national population, making it a world-first for scientific research. The team came up with ways of pooling sources of routinely held NHS data that are stored and gathered differently across the four home nations. All the data were securely held, anonymised and available only to approved researchers.

They looked at Covid vaccination history, hospital records and death records for everyone over five years old between June 1st and September 30th 2022, and found that people who had not been up to date with their Covid vaccinations and boosters were twice as likely to die or be hospitalised with the virus.

The first thing to be said is that the period chosen is strange. The summer is obviously not the usual season for coronaviruses and those weeks in 2022 were one of the low points for Covid deaths compared to earlier in the pandemic – though there was a small spike associated with one of the Omicron subvariants (the red box below shows the study period). The late period is picked presumably because by this point the take-up of the endless ‘boosters’ on offer was seriously waning and so this allows the authors to press their main point, which is that skipping your boosters can be deadly. However, it hardly counts as a typical period for Covid deaths.

There is no comparison with the unvaccinated in the study. Instead all those not ‘up-to-date’ with the vaccines – collectively labelled “undervaccinated” (surely not a word) – are lumped together for comparison with the ‘fully’ vaccinated and boosted. How many doses counted as ‘under-vaccinated’ depended on a person’s age and hence how many he or she had been offered.

Overall it was a period of worryingly high excess deaths, somewhere around 10-20% (see below). Few of these were Covid deaths, however – around a quarter of the excess deaths that summer could be attributed to Covid, leaving around three quarters (14,000 deaths) due to other causes. Unfortunately, though not surprisingly, the study does not look at deaths from other causes, so we have no idea how many of these 14,000 unexplained deaths were vaccinated and how many doses they had received. Another missed opportunity.

It also needs to be stressed that the results are heavily modelled, with numerous adjustments and counterfactual scenarios coming into play. For example, from the methods section:

In each nation, we separately fitted logistic regressions in the age groups five-11 years, 12-15 years, 16-74 years and 75 years or older with undervaccination [sic] as the dependent variable. We then fitted Cox models with time to severe COVID-19 outcome as the dependent variable in the age groups five-15 years, 16-74 years and 75 years or older. Individuals were censored at non-COVID-19 death, deregistration or end of the study period. Vaccine deficit was included as a time-dependent exposure, changing levels on the date an individual received a vaccine dose that put them into a different category. We carried out analyses with a common set of adjustments, and an extended analysis that included further adjustments using additional variables that varied by nation depending on availability.

Is all this modelling and adjustment sound, does it move us closer to a realistic picture of the vaccines’ performance? It seems impossible to know. But one thing we do know is that mathematical models had a dire pandemic. It’s hard to see why these would be any different.

Despite the authors’ best efforts, however, even the tortured data would not confess the entire pandemicist creed. As the ‘Naked Emperor’ noted in his write-up, the authors were obliged to include one paragraph in the discussion section stating that the unvaccinated fared better than the rest of the ‘under-vaccinated’, i.e., those who had been vaccinated but were not up-to-date with the latest booster.

Our estimates for the 16-74 years and 75 years and older age groups show that being unvaccinated (strictly maximum dose deficit) was associated with similar or lower hazard ratio for severe COVID-19 outcomes compared with being vaccinated but having a vaccine deficit of at least one dose. This association could be due to vaccine waning and the fact that the most recent dose for those with a vaccine deficit frequently occurred many months before the study start date. The association could also be due to an uncontrolled selection effect for healthier individuals being more likely to be unvaccinated.

Note the attempt to claim a healthy unvaccinated bias – the opposite of the well-known tendency for people from groups with overall better health (e.g. those with higher socio-economic status or ‘white’ ethnic groups) to receive more vaccines. This healthy vaccinee effect is recognised to make vaccines appear in observational studies more effective that they really are and is one of the reasons why observational studies are often unreliable for estimating the effectiveness of medical interventions. Yet here we have the opposite being claimed (based on no cited evidence) to explain the unexpectedly poor performance of the vaccine.

It’s also worth drawing attention to some of the raw data in the study that, by themselves, do no favours to the vaccines. Overall, 30,407,626 of 68,204,268 people were ‘under-vaccinated’, a proportion of 44.6%. However, just 14,156 of 40,393 severe COVID-19 outcomes (hospitalisation or death) were in ‘under-vaccinated’ participants, a proportion of 35%. This means the ‘under-vaccinated’ were significantly under-represented in Covid hospitalisations and deaths – there were 22% fewer than we would expect given their numbers in the population (35 ÷ 44.6). Of course, these are raw figures which may be confounded, particularly by age. The authors would claim that that’s what all their modelling and adjustments correct for. I guess you’ll have to take their word for that.

It does mean though that, as a matter of actual events, nearly two thirds of Covid hospitalisations and deaths that summer were in the fully vaccinated-and-boosted, despite them only being 55.4% of the population. Hardly a pandemic of the unvaccinated.

So, did the Covid vaccines really prevent 7,100 Covid hospitalisations and deaths during summer 2022? It’s impossible to know from this study, though I’m doubtful. But we can note that, even if they did, it would only mean that the ‘fully vaccinated’ group would have gone up to 33,417 hospitalisations and deaths. So, even under the model assumptions, being ‘fully vaccinated’ only reduced serious Covid disease by 21% (7,180 ÷ 33,417). And of course, the completely unvaccinated actually fared better than the ‘under-vaccinated’ in the model.

Weren’t the vaccines supposed to “prevent“ COVID-19?

Stop Press: A WHO statement appeared today claiming from the WHO’s own research that “at least 1.4 million lives in our [European] region were saved thanks to safe and effective COVID-19 vaccines”. “The evidence is irrefutable,” says Dr. Hans Henri P. Kluge, WHO Regional Director for Europe – citing, er, a pre-print of a modelling study. Irrefutable. But why the sudden push to reassure the public about the efficacy of Covid vaccines – do they know something we don’t?