This week in 150Seconds, I take on a recent study in the journal of Neurology that examined the effects of Arts, Crafts, and Computer use on the development of mild cognitive impairment in the elderly. It's a nice study, but an error in the interpretation of subgroup analyses is one I see frequently, and decided to discuss in the following video:
In this piece, I look at a recent BMJ article that says that all the prior research regarding the health benefits of alcohol is due to inadequate controls. Cool idea, but is there enough power in the study to prove it?
Check out the video, below:
This study appearing in Cancer Prevention Research used an excellent dataset and really high quality statistical methods to show that metformin does not cause, or prevent, lung cancer. So why does the article and the press release focus on a small subpopulation?
Click below to view the video: