Discussion about this post

User's avatar
The Scam Doctor's avatar

You’re correct that vitamin D deficiency is more likely associated with diseases than a cause. However, poor study design does not really contribute to this belief. Not all studies are equal, and some are better quality than others. Because we do now have high quality studies on certain human outcomes, we can push the lower quality studies to the side. When we look at higher quality studies, we get more confident answers.

For example, VITdAL-ICU and VIOLET studies show us that there is no improvement in infections when we supplement vitamin D in the sickest people with the lowest levels. Similarly, when we study Vitamin D supplements the same way we study pharmaceutical medications, in a Randomized Controlled Trial, we consistently see that the supplements have zero effect on blood sugar levels.

I went directly to the data in the meta-analysis you reference in women with PCOS showing “improvements in insulin [and androgen] markers.” It showed just one study in the analysis with statistically significant insulin level decreases. However, that study also had a wide confidence interval meaning highly variable results. More importantly, the decrease in insulin level was not *clinically significant*, meaning that it doesn’t translate to any meaningful effect in humans. To explain that better, I can statistically significantly increase someone's bank account by $0.01 (meaning it will reliably happen), but it will make no difference in their life.

No posts

Ready for more?