As before, the report itself states this is “not the most appropriate method” to assess vaccine effectiveness as it is not adjusted for various confounders (and they do not provide the data that would allow such adjustments to be made). ‘Fact-checking’ website Full Fact (funded by Big Tech) are currently trying to censor the Daily Sceptic because, they claim, this means it is ‘incorrect’ to use the data in the report to calculate vaccine effectiveness. This is not true, however: regardless of what PHE deems to be the “most appropriate method”, vaccine effectiveness is defined as the reduction in the proportion of infections in the vaccinated group compared to the unvaccinated group, and it is perfectly acceptable to estimate it from population data, as long as any limitations in the data are acknowledged.
It is certainly not ‘incorrect’ to use the latest population-based data to get an up-to-date estimate of unadjusted vaccine effectiveness as part of tracking how the vaccines are performing on the ground.
Perhaps the most important limitations in this data are that the high-risk were originally prioritised for vaccination and that those who have been previously infected may be more likely to decline vaccination. Both of these would artificially lower the estimate of vaccine effectiveness. However, a recent population study in the Lancet adjusted its vaccine effectiveness estimates to take account of no fewer than 22 different confounding factors, including these, and in almost all cases this resulted in very little change. For instance, here are the adjusted and unadjusted estimates against infection by age. (Note that the high values here are for the whole study period; what the study showed overall is that in more recent months vaccine effectiveness has been dropping fast.)
Two stay the same, two change by one point, one changes by two points and one changes by three points. This is typical of the vaccine effectiveness estimates in the study, with very few exceptions. This suggests that the unadjusted estimates from large population studies like this are already very close to the mark in most cases, with any adjustments being small. This gives us reason for confidence that the unadjusted estimates from the PHE data, even if, according to them, not “the most appropriate method”, will be sufficiently close to be useful.
So here, without further ado, is the table with the latest unadjusted vaccine effectiveness estimates, for the period September 6th to October 3rd. (For the previous three tables see my previous post.)
Note that unvaccinated here means actually unvaccinated, not partially vaccinated or post-jab. Hospitalisation means “cases presenting to emergency care (within 28 days of a positive specimen) resulting in an overnight inpatient admission”.
Strikingly, the (unadjusted) vaccine effectiveness (VE) in over-18s continues to drop. For those in their 40s it hits nearly minus-86% this week, down from minus-66% in last week’s report. This means the double-vaccinated in their 40s are now getting on for being almost twice as likely to be infected as the unvaccinated of the same age. Those in their 50s, 60s and 70s have similarly super-low VE estimates, while the unadjusted VE for those in their 30s goes negative for the first time, having been dropping for some weeks. For the under-18s, on the other hand – which is the group currently being vaccinated – it actually went up, from 84% to 88%.
Public health officials should be making a priority of investigating the reasons for this alarming inversion of vaccine effect in the over-30s. The fact that instead we have an effort from Government-approved ‘fact-checkers’ to suppress the reporting of it is disturbing, to say the least.
Vaccine effectiveness against serious disease and death continues to hold up well, save in the over-80s, where VE against hospitalisation has dropped from 59% to 51% since last week’s report, which is worrying as most of the deaths are in the over-80s. Effectiveness against death in the over-80s has been sliding more gradually from 70% in weeks 32-35 down to 64% in weeks 36-39, a month later.
Oddly, the text of the report contains an error. It states: “The rate of a positive COVID-19 test is substantially lower in vaccinated individuals compared to unvaccinated individuals up to the age of 39.” This is the same statement (word-for-word) the surveillance reports have made since they started reporting this data in week 36. However, it clearly is no longer true for those in their 30s, where the infection rate in the vaccinated is now slightly higher than in the unvaccinated, and needs updating.
Cases, hospitalisation, attendance and deaths by vaccination status are now presented in the COVID-19 vaccine surveillance report and therefore this data will not be produced in future editions of the variant technical briefing. These tables will be reinstated in the technical briefing if new variants of concern arise.
This is a pity as the Technical Briefing data, while limited to sequenced Delta positive tests, was useful because it went back to February and was published with a fortnight added at a time, allowing data for each two-week period to be analysed. The Vaccine Surveillance report data, on the other hand, only appears in four-week chunks a week at a time, preventing finer analysis, and only goes back to August.
What is really needed, of course, is for the full anonymised data to be released so that it can be analysed independently of Government and its favoured scientists. This is what those in Government and Parliament who care about transparency and truth should be pushing hard for, as without such full transparency the scope for real accountability is limited.
In the meantime, this real-world data from PHE, with infection rates in the double-vaccinated hugely outpacing those in the unvaccinated across many age groups, continues to make a mockery of the vaccine passports and mandates that have become oddly popular even as the data mounts-up that they are pointless.