Cucumber versus lettuce and COVID-19

Home Forums General Discussion Forum Cucumber versus lettuce and COVID-19

Viewing 3 reply threads
  • Author
    Posts
    • #8711
      Aaron Hartstein
      Moderator

      Hi everyone,
      Since the board has been pretty quiet this month, I figured I would add some levity and a small reminder of the difference between pre-print and peer-review literature. Additionally, I thought this might potentially start a small discussion of causation versus correlation regarding study design. As a clinician, when reading a study, how much does the design influence your analysis, interpretation, and application? Are correlational studies valuable? If so, in what way? Should we only make decisions from cause and effect studies? Thoughts?

      I hope everyone is healthy and safe.

      Aaron

    • #8715

      I think correlational studies do have their place. In the world of nutrition literature correlational studies are commonplace because of the large amount of confounders and behavioral aspects to diet. It is difficult to control for exposure to foods, nutrients, etc. as opposed to our ability to control exposure to an intervention in medical research.

      This does not however excuse this study for the oversight of cultural differences, exposure rates, healthcare quality, and the large amount of uncertainty that is associated with COVID-19. A little concerning that the sections on limitations and the interpretation explaining why the results didn’t make sense took up most of the paper.

    • #8716
      awilson12
      Participant

      Looks like cucumber and cabbage are out. Maybe there is a correlation between foods that start with the same first letter as COVID and COVID mortality?

      Design definitely influences my thoughts on the credibility of the findings and my clinical reasoning within the limitations of the article.

      I do think that correlational studies have value but have to be careful not to in your mind make it a causation. An example that comes to mind for me is some of the studies on ACL-R return to sport that correlate decreased re-injury rate with waiting until 9 months for return to sport. This isn’t saying that returning before this is the cause of injury or giving a black and white cutoff, but still gives good information to guide return to sport based on data from large ACL databases.

      I think that nailing down cause and effect in PT studies is hard to do, so I feel like there are a lot of correlational studies that can still add value to helping with decision making.

    • #8718
      helenrshep
      Participant

      Good thing I haven’t been eating cucumbers and cabbage recently… think this means I don’t have to wear a mask?! Just kidding of course.

      My earliest memory of correlation vs causation is the example of ice cream consumption leading to murder because of the correlation between rate of murders and ice cream sales (since both are high in the summer). In glaring examples like this one it’s easy to clearly understand that correlation is not causation, however, things aren’t always that clear cut like in Anna’s ACL-R example.

      It’s incredibly difficult if not impossible to control for every variable. While correlational studies are still important and add value, I tend to weight them lower on the totem pole of evidence than causation studies.

Viewing 3 reply threads
  • You must be logged in to reply to this topic.