As far as I understand, it is well documented that smoking correlates with B12 deficiency, and so many smokers take this vitamin as a supplement to make themselves feel better about their smoking habits.<p>This study then showed, not just that people got cancer, but they got lung cancer. This study was not a randomized trial that can show causation: it was a cohort study, where they ask a bunch of questions about peoples' past or ongoing behavior and then look for correlations from which you can try to guess at potential causes.<p>In a cohort study, there are attempts to predict confounding variables and "control" for them, but this involves essentially assuming the result and backing out the math, and assumes you trust that the participant is giving you correct data anyway.<p>A really fun confounding variable you get in these kinds of studies is that people who even bother to do your study correctly are also people who probably listen to their doctor and brush their teeth and wear their seatbelts and generally lead a safer life. (I am forgetting the name of this, but it is something like "compliance effect"?)<p>Further, if you have participants who don't just answer poorly but actively lie in ways that are correlated with the variables you care about (maybe "another way I make myself feel better about my smoking habit is to pretend I smoke half as much as I do or also smoke cigars and not count them"), you are effectively going to have no hope of "controlling" for that in the way you can with a blinded randomized controlled trial.<p>Given this, I read this and only think "that's an interesting hypothesis that is now worth digging into with real science to figure out what is going on", not "OK, everyone should stop taking B12 supplements if they don't want to end up randomly getting lung cancer"... that is just way too specific and too predictable of a cancer to be associated with this particular supplement.