https://www.ncbi.nlm.nih.gov/pubmed/31074816

Post menopausal women should be careful about over supplementing.

buffy said:

https://www.ncbi.nlm.nih.gov/pubmed/31074816

Post menopausal women should be careful about over supplementing.

Texas sharpshooter effect? ie. They looked at dozens of associations and selected out the strongest.

95% confidence interval is especially prone to texas sharpshooter, need 99% confidence interval.

> During follow-up, 2,304 of 75,864 women had a hip fracture. Both vitamin B6 and vitamin B12 were associated with increased fracture risk. Women with a combined high intake of both vitamins, exhibiting an almost 50% increased risk of hip fracture.

Let’s see if i can make sense of their definition of high and low. Of their P value for linear trend. Etc.

> 96,467 women with 2,812 hip fractures

Make up your mind.

> Compared with the reference category of total vitamin B6 less than 2 mg/d, an intake at least 35 mg/d was …

Come on, give me a sample size in both groups.

> there was no clear association between vitamin B6 from diet only and hip fracture or between vitamin B12 from diet only and hip fracture.

Hmm.

Need access to their tables, and can’t do it from android. The first thing i need to check is how many fractures in the high both vitamins category. Eg. If there are only 1,000 women in that category then we’re talking only about 30 hip fractures which would be statistically almost meaningless.

mollwollfumble said:

buffy said:

https://www.ncbi.nlm.nih.gov/pubmed/31074816

Post menopausal women should be careful about over supplementing.

Texas sharpshooter effect? ie. They looked at dozens of associations and selected out the strongest.

95% confidence interval is especially prone to texas sharpshooter, need 99% confidence interval.

> During follow-up, 2,304 of 75,864 women had a hip fracture. Both vitamin B6 and vitamin B12 were associated with increased fracture risk. Women with a combined high intake of both vitamins, exhibiting an almost 50% increased risk of hip fracture.

Let’s see if i can make sense of their definition of high and low. Of their P value for linear trend. Etc.

> 96,467 women with 2,812 hip fractures

Make up your mind.

> Compared with the reference category of total vitamin B6 less than 2 mg/d, an intake at least 35 mg/d was …

Come on, give me a sample size in both groups.

> there was no clear association between vitamin B6 from diet only and hip fracture or between vitamin B12 from diet only and hip fracture.

Hmm.

Need access to their tables, and can’t do it from android. The first thing i need to check is how many fractures in the high both vitamins category. Eg. If there are only 1,000 women in that category then we’re talking only about 30 hip fractures which would be statistically almost meaningless.

OK, have access to tables from Windows.

138 hip fractures in the category of High B6 and high B12. That’s sort of OK statistically.

If we lump in the higher rate of hip fractures with 419 fractures in the category of Medium B6 and high B12 it does become statistically significant.

But because Low B6 and High B12 is a very rare combination and can be ignored, what we’re really seeing is a correlation with B12 alone.

As we know, correlation doesn’t imply causation. Perhaps those prone to hip fracture are more likely to take vitamins as a preventative measure.