Hacker News new | ask | show | jobs
by thunkle 441 days ago
> Of those who never drank, 40% had vascular brain lesions. Of the moderate drinkers, 45% had vascular brain lesions. Of the heavy drinkers, 44% had vascular brain lesions. Of the former heavy drinkers, 50% had vascular brain lesions.

I mean when you put it that way it doesn't seem so bad.
2 comments
throwaway7783 441 days ago
Yeah, it almost reads like "if you have to drink, drink heavy and don't stop drinking" - that seems to be the least harmful category other than not drinking at all!
link
doubled112 441 days ago
> impaired cognitive abilities were observed only in former drinkers.

Have to keep drinking or you’ll experience a loss in cognitive abilities.

link
ergonaught 441 days ago
Yeah. The 40% seems like the interesting bit.
link
bbarn 441 days ago
Yeah, the fact that 40% get them regardless makes the slight increases seem.. statistically irrelevant?
link
mikeyouse 441 days ago
It turns out we in fact have tests to determine statistical significance - and the fact that this study was peer reviewed and published means the results were indeed statistically significant!

> We included 1,781 participants (mean age 74.9 ± 12.5 years, mean education 4.8 ± 4.0 years, 49.6% women, and 64.1% White). Compared with participants who never consumed alcohol, moderate (odds ratio [OR] 1.60, 95% CI 1.19–2.15, p = 0.001), heavy (OR 2.33, 95% CI 1.50–3.63, p < 0.001), and former heavy (OR 1.89, 95% CI 1.41–2.54, p < 0.001) alcohol consumptions were associated with hyaline arteriolosclerosis while only heavy (OR 1.41, 95% CI 1.10–2.30, p = 0.012) and former heavy (OR 1.31, 95% CI 1.02–1.68, p = 0.029) alcohol consumptions were associated with neurofibrillary tangles. Former heavy drinking was associated with a lower brain mass ratio (β −4.45, 95% CI −8.55 to −0.35, p = 0.033) and worse cognitive abilities (β 1.31, 95% CI 0.54–2.09, p < 0.001). The association between impaired cognitive abilities and alcohol consumption was fully mediated by hyaline arteriolosclerosis (β 0.13, 95% CI 0.02–0.22, p = 0.012).

link
tbrownaw 441 days ago
Being statistically significant is about how good your measurement is. (In one particular way.)

It's entirely unrelated to being practically significant.

link