Medical “study” needs a course in logic.

The mantra “Studies show…” reveals a lot, although often not the conculsion being drawn! Here’s yet another example of irrationality that passes for science…

A study in Diabetes Care found that sleeping five hours or less, OR nine hours or more each night may increase your risk of developing diabetes. Sounds like if I just sleep the exact right amount, that I could lower my risk of diabetes, right?

Nope. Looking at the study, here’s what I found:

The study was done only on men, age 40-70. It claims to have controlled for confounding factors, but it also says that the effects of sleep on diabetes risk may be mediated by changes in testosterone. The men who slept 7 hours (which they conclude is ideal) were the younger ones with higher testosterone! What they’re seeing is a relationship between low testosterone and factors associated with diabetes such as insulin resistance and obesity. And there’s no evidence of any causal relationship, just a relationship.

Sometimes insulin resistance is a risk factor for diabetes, but there are actually times when we’re supposed to be in insulin resistance - in the summertime, to store fat for the winter. Insulin resistance itself, like blood pressure, doesn’t mean much outside of the whole living context of the person. But of course test result numbers are easier to work with!

This kind of analysis is interesting to me not only from the point of view of sorting out the medical issue at hand, but also for seeing the pattern in the way that these types of studies paint deceptive pictures. And that so much of what we accept as “common knowledge” about medicine and health is based on these deceptions.