Some extra points I learnt while reporting this story. Bharat Biotech has claimed that because it is a whole virus vaccine and attracts a wider range of antibody responses (to more than just the spike protein), it is more likely to protect against UK variant https://twitter.com/PriyankaPulla/status/1346392237282713601

SEC seems to have accepted this argument. But as Vineeta Bal told me, a wider range of antibody responses may also raise the risk of antibody dependent enhancement (not yet seen in any COVID vaccine, but impossible to rule out without wider administration) @abledoc

So, if one is going to ascribe a theoretical probability of higher efficacy against the UK variant, why not also consider the theoretical probability of greater risk?

Even if Bharat Biotech has given 11000 people the first dose (a fact that has given comfort to some experts), how much time has gone by since the first dose? If one can can investigate the SII adverse event, which occurred after 10 days of first dose, for possible relatedness...

...how are we not asking how many days these 11000 people in the Bharat Biotech trial got monitored for safety? Given that 22500 people were enrolled only by 3 Jan, even if all got the first dose, are we saying fifteen min of safety data is enough?

And as @gkang said, Bharat Biotech is using a new adjuvant. She pointed out that after an H1N1 vaccines were developed in 2009, instances of narcolepsy in Scandinavian countries, identified post vaccine introduction, were attributed to the adjuvant.

Important to remember that in some ways, Phase 3 itself is an arbitrary minimum line drawn for establishment of safety. It is a *reasonable* point at which to introduce a vaccine, but a Phase 3 trial also cannot confirm 100% safety.

That's because vaccine safety is a question of numbers, as Dr Kang said. Rarer adverse effects only show up when larger groups get the vaccine

But now that we have agreed upon Phase 3, why are we trying to skip that too?

Also, as our estimates of safety become more and more iffy (we have no estimate for the Bharat Biotech vaccine - we have zero data), even rarer adverse events should start figuring in our risk benefit calculus

Why give a healthy person a vaccine with low efficacy (below 50% for instance) even if the adverse effect is rare? What is the ethical justification for this? An effective vaccine - sure. But here, we know absolutely nothing about efficacy.

Plugging this story I had written about why Phase 3 trials are necessary when ICMR announced the August 15 deadline last year https://science.thewire.in/health/icmr-covid-19-vaccine-bharat-biotech-clinical-trials/

Vaccines are complicated. A mechanistic understanding of what contributes to safety/efficacy *CANNOT* predict safety and efficacy in real life. That's why we do Phase 3 trials, instead of assuming immune response will protect, or assuming an inactivated vaccine will lead to ADE.

Hypotheses of safety/efficacy based on mechanisms discovered in in-vitro and animal studies fail, disastrously, and often. It may not fail in Bharat Biotech's case, but we should know better than to assume it won't.