idk where Sanjay Gupta got his info, but I got mine from Pfizer's own indications, which were obviously aimed at being very conservative given that typically people are only warned not to take it if allergic to any ingredients in the vax. Pfizer went the extra mile on this, warning against people with ANY allergies getting it and making sure to take it at a place with resuscitation facilities.
Those instructions somehow didn't trickle down to the vax givers and 2 NHS nurses who were vaxed had reactions and used their epi-pens. The fact that they already had epi-pens on them suggests a history of allergic reactions, and indicates Pfizer's info was not followed to the letter.
So yeah, snafu.
WornDown - they are still trying to figure out what to do about people who got the placebo. Typically, they would be followed for 1-2 years (still in placebo group) to have a group for comparison to the vaxed group. However, it could be considered unethical to withhold an effective treatment from people who could benefit from it.
There are a few options:
1. break the blind and give placebo patients the vax. Scientists don't want to do this because it will reduce the amount and the validity and therefore value of future data from the rest of the trial.
2. Dr. Fauci suggested a valid and ethical way of proceeding: turn it into a crossover trial, ie give the placebo to the original vax group and vice-versa. The trial could have originally been planned that way and still be valid and ethical. This issue is that the original design is expected to be more powerful for estimating the effect of the vaccine compared to placebo. So this would be a way of 'rescuing' the trial from the ethical and validity perspectives, with the 'cost' coming in terms of loss of power.
3. Try to force people to stay in their assigned groups. Again, this suffers from the ethical concern that a treatment is being withheld from people who could benefit from it (Tuskgegee anyone??). Plus I'm sure that several individuals have already figured out which group they are in, so the study is not completely blinded anymore and people who got the placebo may be inclined to drop out of the trial and get the vax. This will complicate any further data analysis.
4. Try to ameliorate dropout by doing a nested second study using just the placebo group. Give some proportion the vax and the rest the placebo. This doesn't need to be 50-50, the percent getting the vax could be higher by some amount, based either on expected power or on an ad hoc basis or on some other criterion (eg, 75-25, or 2-1, or whatever).
There could be other options as well. Of these options, the best way to preserve the validity and ethics of the trial are 2 and 4.
Trials do end up getting stopped early all the time. The usual criterion for this is that the difference between treatment and placebo/control reaches a sufficiently high value at some point in time that it would be unethical to continue the trial.
It is problematic to stop a trial early if the placebo group is doing much better than the treatment: this is because the effect estimate in the middle of the study will likely be biased unless there is sufficient (statistical) POWER at that point in the study. This is an unusual situation, since studies are designed to have good power at the END of the study, not the middle. However, it does happen.
An example is the Women's Health Initiative and other HRT studies provide notable examples - these were stopped due to specific types of HRT being associated with increased incidence of breast cancer, blood clots in the lungs, strokes, and heart attacks/coronary events). The trials were stopped.
On the other hand, if the treatment is much better than the control then some may consider it unethical to not provide the control individuals with the treatment.
When a clinical trial is stopped early like this, it is called 'stopped early for benefit' (or sometimes a truncated randomized controlled trial). It is scientifically advised to avoid this, becase the treatment effect tends to be overestimated compared to what it would be were the trial carried out to the end. Why is this bad? Because any NEGATIVE (side) effects may not have had a sufficient time to show up. So the overall performance of the treatment will be exaggerated.