Exploring Vaccine Harm (part 3)

A possible method for the assessment of relative vaccine harm in younger people

In part 2 of this series I revealed the results of some exploratory generalised linear modelling (GLM) that relied on the weekly count of all cause deaths for the 80+y age group along with an indicator variable marking vaccine rollout as independent variables in the prediction of weekly all cause deaths for younger age groups. The 15 – 19y and 20 - 24y categories stood out like a sore thumb with elevated rates in the post-vaccine era that amounted to an 11% - 19% hike in weekly deaths, and we discovered elevated rates reaching statistical significance for all other age groups save for persons 35 - 39y and persons 45 – 49y.

Given that everything appears to be kicking-off in the lump of time that is the post-vaccine era I decided to break the data down into four distinct phases: pre-vaccine; main doses (jabs 1 & 2); primary booster (jab 3) and secondary boosters (jab 4 onward). To get the ball rolling quickly I pressed a couple of buttons that gave me the output of a Kruskal-Wallis non-parametric test for independent samples for all 18 of the vaccine-eligible subpopulations. Here’s the formidable-looking table that spewed out, being surprisingly easy to decipher:

The issue here is whether mean weekly deaths for a particular population subgroup have differed significantly between the four vaccine era phases. If the column headed ‘Sig.’ contains a probability value >0.05 then the answer is NO! If that column contains a probability value <=0.05 then the answer is YES! (if we go for the 95% level of confidence).

No less than eight age bands hit or exceed the 99% level of confidence of p<=0.01, with two (15 – 19y; 70 – 74y) hitting the 99.9% level of confidence (p<=0.001). Scorchio! Let’s have a look at an error bar plot of these two groups without even reaching for a biscuit…

There’s a great deal to consider here, so I shall carefully pick put what strikes me, starting with the younger age group.

Younger Death

What we have is an utterly bizarre situation whereby unvaccinated 15 – 19y persons in the pre-vaccination era of 2020/w1 – 2020/w49 exhibit a much lower mean weekly death rate than at any other time since. We’re talking about a novel virus at full strain strength for which pre-existing immunity didn’t exist, and for which early clinical knowledge and patient management was lacking to the point of being dangerous, yet this was the safest period of all for teenagers!

Teenagers started getting their initial doses when older folk were getting their primary booster, this era being characterised by the highest teenage death rate of all. If this is not an indication of vaccine harm I shall eat my Panama hat. Even if not an accurate indication of harm (other things can kick-off) it certainly offers a signal that should trigger serious lines of enquiry, yet we’ve come to realise that these are never going to happen: the apple cart is too laden to tip over.

In terms of quantifying the hike we’ve gone from a pre-vaccine mean of 13.00 deaths per week to a peak vaccine mean of 17.23 deaths per week, this yielding a factor of x1.325, being an increase of 32.5% in old money. With susceptible teenagers now a sad statistic we observe a decline in the death rate that may be attributed to survivorship bias plus a few other things.

No doubt the Chris Whittys of this world have been ordered to put a favourable spin on this, so we must expect to see something damning about alcoholism, drug abuse and risk-taking behaviour among this age group if not climate change, eating too many crisps during lockdown, and a backlog of undiagnosed heart problems that have only decided to plague the nation in recent months after decades of notable absence. Then there’s mauled broccoli and raspberries to content with.

Older Death

The situation for the 70 – 74y age group is different in that we see an elevated risk of death in the pre-vaccine era, as expected. We may note that this is not significantly different to the mean risk for main dosing (p=1.000; Kruskal-Wallis, n=90). Neither is it significantly different to the mean risk for the primary booster campaign (p=0.354; Kruskal-Wallis, n=75). Taken together these results would suggest that the vaccine up to jab 3 has done nothing to reduce the risk of death for this age group (all things being equal).

The situation dramatically changes for the secondary boosters and if I were in a position of authority I’d no doubt be forwarding this as confirmatory evidence of the effectiveness of jabs 4 and 5 and running out a slick nudge campaign aimed at folk who can have babies. We can verify this by considering the equivalent slide for non-COVID death for, as far as I am aware, the jabs don’t confer benefits for falling down the stairs, being hit by a bus, hypothermia, malnutrition or kidney failure:

Well there you go! It turns out the secondary booster period aligns with a period of naturally low death rates amongst the older generation, though I suspect survivorship bias comes into play here also (only the vigorous and healthy remain alive).

Coffee & Cogitation

A few simple slides and some pretty basic analysis and we are once more facing a situation that requires coffee, cake and cogitation. There’s a great deal I can still squeeze out, especially if I resort to multivariate or ARIMA modelling to account for various biases such as disease prevalence, service provision and seasonality. And so it is time to ponder on where I could take this approach next. Until then…

Kettle On!