Norman Fenton, Professor of Risk Information Management – Probability and Risk July 13, 2022

You don’t need to look very hard into the latest ONS data on all-cause mortality by vaccination status to confirm what we have been saying for a long time (see for example here and here): the ONS data is so obviously flawed (due to miscategorisation, missing vaccine deaths, and underestimates of the proportion of unvaccinated) to be worthless. 

This simple extract from Table 3 on the age standardized mortality rate*** for non-covid deaths over the full period (Jan 2021 – May 2022) is all you need to know to realise everything else in the report is problematic.

Even the most fervent vaccine supporters know that it is not a miracle cure for non-covid disease. Yet, if you believe the above ONS data, it must be – since the unvaccinated are dying at a much higher rate of non-covid illnesses.  We stopped doing detailed analyses of the ONS updates on this data in January when, having exposed the blatant flaws, we called on a) the ONS to publicly withdraw their dataset and b) the retraction of any claims made by others that are based upon it.

In case there has been an error either in our understanding of Table 3 or in the ONS presentation of the data we will be happy to set the record straight and update this article. But the bizarre data in the above table is consistent with the bias we found in the previous ONS reports. This bias (whereby mortality rates for the unvaccinated are overestimated and mortality rates for the vaccinated are underestimated) will inevitably occur when people dying (of any cause) shortly after vaccination are classified as unvaccinated and where the proportion of unvaccinated is underestimated.

The above table exposes that, far from fixing the many problems (that we discussed in detail with the authors of the ONS reports) these systemic flaws may have got even worse.

There was previously a laughable ‘explanation’ provided for the increase in non-covid mortality in the unvaccinated  – namely the ‘healthy vaccinee effect’. We already debunked that both theoretically and empirically. In any case, even if such a ludicrous hypothesis was correct, it would still mean that all the data – and conclusions from it – would have to be adjusted to take acount of the bias (which they certainly have not done).

Anybody attempting to draw any conclusions from the ONS data, let alone do any intensive data analysis, are now wasting their time. While it is possible there may be some interesting trends to discover from the data (such as possibly noticed here), these are more likely to simply reveal the extent to which the misclassification problems and other flaws have changed over time.

One of the most depressing things about the last two years has been how silent almost the entire academic community of statisticians and mathematicians has been in the face of so many obviously flawed statistics/studies that promote the ‘official narrative’ on Covid and the vaccine. A small number have contacted us privately to say they see through the nonsense but cannot speak out because they know how damaging it will be to their careers. In contrast, some prominent statisticians and mathematicians, who have refused to call out obviously flawed data supporting the ‘Covid narrative’, have been vocal in trying to downplay/ridicule more reliable data that supports an alternative narrative. Perhaps some of the prominent statisticians might at least finally publicly acknowledge a problem with the ONS data since it is so obvious?

*** we have also complained about the fact that the ONS started to use the age standardized mortality rate (unnecessary when you have the detailed breakdown by age) since it is essentially just a model that may not be suitable in this context.

Norman Fenton is Professor in Risk Information Management at Queen Mary University of London and also a Director of Agena, a company that specialises in risk management for critical systems.

Source