I recently read this paper. It says there was a major mathematical error in the meta study used to dictate the Vitamin D standards for many countries: https://www.ncbi.nlm.nih.gov/m/pubmed/28768407/
back
2 comments
If you're going to cite anything, cite the original letter, which was featured in a relatively high impact journal: http://www.mdpi.com/2072-6643/6/10/4472
I've not had time to review it yet.
Having spoken with a mathematician, I'm not sure this letter's assumptions are entirely sound.
The major finding is the original targets are based on the average response (so up to 50% still in deficiency), not getting 97% to a base level.
However only being able to use 10 papers and without their data (only standard deviations) hamstrung their ability to peg a better range. More, their approach requires a normal distribution and a solid link between input and output levels, neither of which really exist here. Random error scuppers their analysis.
Again, this is not to say there isn't something here, it just needs more work, especially when recommending people take 2-3 times the current max tolerable limit.