Needle To Door Time (part 3)
In this miniseries I use survival analysis techniques to investigate delays between vaccination and death for 5,039 cases over the period 2020/w53 – 2021/w36 using data from an unknown NHS Trust
I hope that readers enjoyed my introduction to Cox regression in part 2 of this miniseries. Some may ask what did it all mean and the simplest answer I can give is that it is riskier being a vaccinated asymptomatic COVID case than a vaccinated symptomatic COVID case when it comes to early death. This is bizarre in the extreme and tells us that we’re missing a big piece of the clinical picture. There are several possibilities starting with vaccine harm and ending with dubious coding, but I’d like readers to offer their thoughts rather than me witter on!
That age as a covariate doesn’t feature in the model offered in part 2 is rather puzzling. Everyone knows that risk of death generally increases with age and so why is it that risk of early death in the vaccinated cohort isn’t a feature of age? One explanation is that the cause of death in these cases isn’t a function of disease progression, which leaves us with vaccine-induced early death as a distinct possibility.
What bugs me the most at this stage is why asymptomatic COVID is associated with early death (28-day death) so I decided to run Cox regression again but using a comprehensive factor I’ve called ‘death class’. The best way to introduce this is to plonk down a frequency table that gives us an idea of where I’m coming from and the sample sizes in each category: