you are viewing a single comment's thread
view the rest of the comments
[–] 3 points 5 months ago (3 children)

What's HR= 1.12 for those of us out of the loop?

  • source
  • hideshow 3 child comments
  • [–] [S] 5 points 5 months ago* (last edited 5 months ago) (2 children)

    HR is hazard ratio, a term often used in observational epidemiology to express event hazard of the variable. The core problems is these studies are not tightly controlled, have innumerable confounders, and 1.12 means a 12% increased relative risk (not absolute risk)... which is so low it might as well be noise. This is a useful tool in the scientific process to generate hypotheses to plan further interventional experiments with, but it cannot establish cause and effect. However, lots of people REALLY REALLY REALLY want to use such data to make claims of cause and effect.

    Most nutritional papers are exactly this. Dr Bradford himself, the main who came up with the smoking->cancer link and wrote the definitive criteria for observational data, even indicated a hazard ratio less then 2 (so 200%) isn't meaningful.

    MD Ede explains this so much better then I can https://www.diagnosisdiet.com/full-article/epidemiological-studies

    Update - Here is probably one of the best published illustrations of this weakness - https://doi.org/10.3945/ajcn.112.047142 - Is everything we eat associated with cancer? A systematic cookbook review

  • source
  • parent
  • hideshow 2 child comments
  • [–] 3 points 5 months ago (1 child)

    Thanks! I actually used to teach stats - at least for awhile 😆. I just came across the meme and that didn't click for an association.

    That is a great explanation by the way. More coherent than I think I could manage.

  • source
  • parent
  • hideshow 1 child comment