2020/11/13
And just in time for Christmas? Who would have thought it? And Matt Hancock has millions of doses lined up for mass vaccinations starting from 1st December? Wow! What foresight!
Just a minute though - has this vaccine completed all its trials for efficacy and safety yet? (no).
Has our government pushed through measures that enable it to proceed with vaccinations before the usual regulatory approval has been given? (yes)
And has the vaccine manufacturer been granted legal indemnity in the event that this vaccine causes serious side-effects or even death? (yes)
Who is shouldering all the risk here? The same mob who never got the opportunity to decide whether it would be a good idea to order all these vaccine doses at (our) huge expense? Or to decide if we wanted to bypass the usual checks and balances? (yes)
The Trial:
"A PHASE 1/2/3, PLACEBO-CONTROLLED, RANDOMIZED, OBSERVER-BLIND,
DOSE-FINDING STUDY TO EVALUATE THE SAFETY, TOLERABILITY,
IMMUNOGENICITY, AND EFFICACY OF SARS-COV-2 RNA VACCINE
CANDIDATES AGAINST COVID-19 IN HEALTHY INDIVIDUALS" (download)
In this study, 80% of participants were given the vaccine, 20% the placebo (sodium chloride).
Participants were "healthy . . . determined by medical history, physical examination (if required), and clinical judgment of the investigator" " between the ages of 18 and 55 years, inclusive, and 65 and
85 years, inclusive (Phase 1), or ≥12 years (Phase 2/3), at randomization. Note that
participants <18 years of age cannot be enrolled in the EU"
There were various exclusions in phase 1 for those likely to have a poor result from catching Covid-19.
The above provides a flavour - read the download for the full detail.
From my reading of the trial documentation it isn't clear what part (if any) the groupings of 15 participants play in the analysis, nor what exactly the "90% successful" claim means in practice.
Based on a single sample of 94 instances of sickness, a mean value (the only value!) of around 10% of those falling ill despite being vaccinated (I'm interpreting here) has been reported.
We can't estimate the variability around this figure from a single sample. We would probably need somewhere between 10 and 100 such samples before we could predict the variability with any confidence.The trial however adopts the following assumption in order to control for variability:
"Bayesian approaches require specification of a prior distribution for the possible values of the
unknown vaccine effect, thereby accounting for uncertainty in its value. A minimally
informative beta prior, beta (0.700102, 1), is proposed for θ = (1-VE)/(2-VE). The prior is
centered at θ = 0.4118 (VE=30%) which can be considered pessimistic. The prior allows
considerable uncertainty; the 95% interval for θ is (0.005, 0.964) and the corresponding 95%
interval for VE is (-26.2, 0.995)." (my highlight)
This approach may or may not be valid but it's certainly an assumption and definitely outside my remnant of statistical expertise.
It would be good if instead of quoting a single result of 90%, the 95% confidence range could be made available (eg: between 80% and 95%). That would give us more information about the opportunity for disappointment.
So beyond noting that assumptions are undoubtedly a fertile source of error, I'm not going to comment further, but I would be happy to see more competent scrutiny from elsewhere.
If we are set on vaccinating the whole population in advance of completion of the normal approval procedures, then some hard scrutiny would seem to be an urgent requirement and my confidence in our Secretary of State for Health and Social Care to arrange this is best left unremarked.
