A new study claims to link consumption of red and processed meat with increased risk of type 2 diabetes (T2D). Here’s why you should ignore it.
First, the study is more weak association epidemiology that is entirely inappropriate for studying chronic diseases:
- T2D is too common a disease and the reported “increase in risk” is too small. Epidemiology is only appropriate for high “risks” of rare diseases. Keep in mind that epidemiology is not technically capable of determining “risk” in the first place.
- No one knows what causes T2D. At the very least, it is multifactorial in nature — and competing causes (known and unknown) can’t be eliminated by waving a statistical wand. No case of diabetes in this study was medically attributed to meat consumption.
- The time period between exposure and disease onset (more than 30 years in this study) is too long for determining cause-and-effect;
- The reported statistical correlations between meat intake (on the order of RR=1.5) are well below the traditional epidemiologic threshold for reliability (on the order RR>2.0). They are most likely to be statistical noise.
- The study’s meat consumption and other confounding risk factor data were self-reported. No one really knows how much meat (or anything else) any study subject actually consumed, or what their T2D risk factors actually were.
Although you won’t read the following in media accounts of the study, the authors were forced to acknowledge at the end of their write-up that:
… because of the observational nature of our cohorts, the observed associations do not necessarily mean causation; although we adjusted for established and potential risk factors for T2D, unmeasured and residual confounding is still possible.
In our estimation, they seem to be more interested in grabbing headlines than doing good research — all the better to keep the grants coming from the incurably anti-meat National Institutes of Health.