[1] https://www.cdc.gov/vaccines/schedules/hcp/imz/child-adolesc...
Or maybe they said "we measured secondary immune response with a lab test, and group D had the highest number"?
If these weren't done for the COVID vaccine specifically, but some other vaccine, couldn't you still put some rough ballpark quantitative guesses on the COVID vaccine to help guide public policy better than "shrug, we have no clue what will happen if you do anything other than 21 days"?
Something like "We think 40% will be protected with one shot, 90% will be protected with two separated by 21 days, and we can fit a curve to these two data points which has the same shape as curves for other decades-old vaccines that have tons of data and is supported by biological theory [paper1] [paper2] [paper3] [paper4] [paper5], and that curve continues to monotonically increase, so waiting 5 weeks instead of 3 should be no problem, in fact we think you'd be 91% immune instead of only 90% immune (provided you don't actually get the disease in the extra intervening 2 weeks), r^2 = whatever, but waiting 1 week instead of 3 is probably super bad, you only get up to 50% based on the shape of the curve and what's known from other vaccines, [paper6] [paper7] r^2 = whatever.
Disclaimer: This assumes COVID-19 vaccine acts like the vaccines in [paper8] [paper9] [paper10] which collectively cover 80% of known vaccines, but it's always possible since it's MRNA the curve will be something super weird and there's absolutely no reason to think it'll be like that, but it'll take us 10 years to definitively rule it out, but if you need to make public policy and you don't have 10 years to wait, a policy based on a tentative analysis with a 20% chance of being wrong, which 20% we got by polling experts' Bayesian priors of the probability that a new vaccine will have a totally different curve, is probably better than a policy based on a completely random guess [paper11] [paper12] [paper13]."
Based on that you could do an expected-value calculation of how many less people will get sick under policies of holding back various percentages of vaccine for second doses (the factor you control), vs some distribution of possible random future production hiccups (the factor you don't). You might find it's better to get shots in arms now to have as many people as possible 50% protected as long as possible if you know it's implausible that it's going to be a disaster if those people wait 5 weeks instead of 3 for the second shot due to a manufacturing hiccup.
You definitely could (and this is exactly what many people are trying to do!) But this approach relies on many assumptions, and the FDA is hesitant to let assumptions (no matter how well-grounded) enter into its evaluations. Thus its near exclusive reliance on RCTs. This is mostly reasonable, but it causes confusion amongst people who don't understand evidentiary standards. Thus all the braying about "no data" or "no evidence."
It's very important to note that Phase 3 was about COVID19 disease and less about Sars-Cov2 infection. They do have some infection information, but the vaccine is licensed to prevent disease, not infection.
If there was more time, there would be more answers, but you can only run so many trials when people are dying from a pandemic.
0. https://www.fda.gov/media/144434/download page 53
* Take all of the above with a huge grain of salt, I am not a doctor, immunologist, public health professional, epidemiologist, etc
Presumably these choices weren't picked out of thin air, or made up by some businessperson based on business considerations -- "We think we can get paid twice if we put it in two doses, but if we try to go to three doses it'll be impossible to market" -- hopefully that's not what happened.
Hopefully instead this decision was made based on actual data -- for example in animals or a small group in Phase I or for some similar vaccine, they did actually try N-day dose separations for all 1 < N < 50 and then made some actual measurements of how fast white blood cells attacked the COVID virus in a lab dish or whatever. The lab test maybe wouldn't be as ironclad as the Phase 2 which actually sends the subjects back into the wild and sees how many got COVID in the next 3 months, but it's some concrete data.
Where is that data? Couldn't you use whatever data was used to decide on the dosing schedule in the first place to quantifiably answer how much worse were the alternative dosing schedules that weren't chosen?