How much uncertainty we accept depends a lot on the potential downsides. For example there isn't much of a downside to wearing masks, so that is a measure you could defend on rather flimsy data if you had to.

The potential downside here is very large, it could mean that the vaccine loses a lot of efficacy if the delay to the second dose is too large. It also could create circumstances that favor mutations that escape the vaccine, because you have a lot of people with a weaker vaccination response while the virus is circulating widely.