[lmm]

> I think due to its generalized nature there are exceptions where it doesn't even apply (like people who are weight training, pregnant women etc).

All models are false, some are useful. BMI is imperfect, but arguing we should ignore it because it has exceptions is like arguing we should ignore that the Earth is round because it isn't perfectly spherical. Looking at BMI alone is occasionally misleading, but it's a hell of a lot better than not looking at anything, which is the realistic alternative.

> So maybe it is the best that institutions can do to gauge the health of a population with the data given, but I think it could be improved drastically with extra metrics (I've heard of measurements around the body being used).

No-one has identified any such drastic improvements. Measuring body fat percentage gives you a small improvement over BMI alone, but it's hard to do at home and only really matters for people with very high muscle mass (who are most likely already using more advanced measurements than BMI alone).