Michael Pollan on how difficult it can be to disentangle the actual health effects of a characteristic or behavior from all the things associated with it.
“Be the kind of person who takes supplements. We know that people who take supplements are generally healthier than the rest of us, and we also know that, in controlled studies, most of the supplements they take don’t appear to work. Probably the supplement takers are healthier for reasons that have nothing to do with the pills: They’re typically more health conscious, better educated, and more affluent. So to the extent that you can, be the kind of person who would take supplements, and then save your money.”
The types of studies that show that people who take supplements are healthier are observational studies. The same types of studies that, say, tell us that hormone replacement therapy is a good idea [CTRL + F for “selection bias” to see the part that talks about that study]. Or tell us that people who eat red meat regularly die sooner. Or that people with a BMI >35 are more likely to die sooner. (In order to do a controlled study of the effects of BMI or body fat percentage on health, you’d have to randomly assign the people you’re studying to be thin or fat, rather than studying people who are already thin or fat. Any people dropping out because they were unable to become thin or fat would weaken your data. I’m sure most readers are familiar with how hard it is to change someone’s weight. And then you still wouldn’t know if it was the weight change itself or the behaviors that caused the weight change that caused any differences between the populations. In fact, the closest thing to “no behavior change”, liposuction, has few or no health benefits.)
With all the Concerned People repeating other Pollan quotes in an attempt to shame and/or control others, it’s easy to forget that he has a lot of sensible things to say, as well.