Yeah it's slightly concerning.
As I understand it the AZ vaccine was 62% efficacious after two doses.
If you postulate a 'risk' scenario of say a reduction of 20% in that figure due to one dose, a further 20% drop due to less protection against new strain and another 20% due to less protection in older age groups then the efficaciousness drops to about 32% ?i.e. instead of say 1000 people dying , 680 die.
A reduction of 10% in each of these categories would result in an efficaciousness of 45%?
I may be talking bollocks but the Government must have done these risk calculations.
Time to let people know the facts.
The efficacy figures are for "symptomatic covid confirmed by PCR"
ie the number of people who reported symptoms and tested positive.
The reduction in severe cases was 100% on the AZ trial. However, the number of severe cases on placebo was only ten (zero on active), so the confidence interval on that would be very large. Nobody died, so there's actually no direct evidence at all either way that these vaccines prevent deaths, though we fully expect they will, given the other evidence.
The "one dose" risk over just three months may well reduce by less than you suggest.
Nobody knows the effect of new variants; most experts seem to think it's small with the currently known ones, but that remains to be seen.
There are, unfortunately, no "facts" to define this, just judgements of risks. Different experts will make different judgements. I am not an expert. But I will speculate, against severe disease:
1. The one dose for three months strategy will make no measurable difference. Evidence: phase 3 data from the trials shows an excellent efficacy response before the second dose; studies on naturally occurring immunity from infection show persistence for more than 5 months.
2. Older people will have excellent protection. Evidence: phase 2 data on immune response in older subjects and emerging Israeli data.
3. New variants will have minimal impact. Evidence: what I understand of expert opinion.
I could be horribly wrong on all counts, but I don't think it matters, as there's no other option anyway. What I would argue is that the uncertainty on all of these makes it really important not to open up until case numbers are really low.