Your points about being careful about labeling are fair. I guess what I'm saying is this: Imagine that you have to set rules. The health and well-being of many

Your points about being careful about labeling are fair. I guess what I'm saying is this: Imagine that you have to set rules. The health and well-being of many people depend on you doing a good job setting those rules. You want to determine the most optimal rules but you have limited time and budget. You need to pick a number. You are lower-bounded by what is feasible and you don't know where your upper bound is. In this scenario, I think it is reasonable to pick a number which is near the lower bound and probably below the upper bound (aka what they did) rather than figuring out what the upper bound is. -Because honestly it may not be possible to reliably determine the upper bound (without literally poisoning people). Data is often scarce or bad when it comes to things which cannot (legally, ethically) be tested. Moreover consider the failure modes: If you pick an upper bound which is too low but still trivially attainable for most people (97.5%) then you inconvenience a very small minority of communities. If you pick an upper bound which is too high, you are literally personally responsible for causing brain damage in children. It's the sort of situation that the term 'better safe than sorry' was intended for.