CAUSATION VS. CORRELATION: THAT IS THE QUESTION!

In the early 1990s large scale studies demonstrated that patient’s undergoing vasectomies had a higher risk of prostate cancer. The common explanation was that there was some statistical correlation between these two findings (perhaps stemming from a younger population exposed to a urologist, i.e. those undergoing vasectomies, were more likely to engage in prostate cancer screening later in life). Despite this statistical finding, the biological causation of prostate cancer from a vasectomy was labeled dubious at best by the larger urological and scientific community (there wasn’t any plausible scientific explanation why this would be the case).

Although this seems like an obvious case, the tension between correlation and causation permeates throughout medicine. The most glaring example of the confusion between these 2 principles is seen in lipid theory – which underscores much of our understanding of cardiovascular disease. While I am not a cardiologist or a primary care physician, a cursory evaluation of history and the unintended consequences which resulted will easily prove this point. A researcher for Minnesota in the 1940s, Ancel Keys, set out to understand the risk of heart attack in (mainly white) men. He used data across 7 countries (mainly urban population) and showed a statistical correlation between serum cholesterol and risk of stroke and heart attack. There was much to be criticized about his methodology and findings including the very select countries he used to set out to prove his theory. Alternate theories were also offered in the 1960’s, including one that suggested that refined sugars were the main culprits in the Western diet causing chronic diseases including cardiovascular disease. Just as many studies disproved lipid theory as proved but didn’t gain traction in the field.
If you are a child of the 1980’s, you know what a profound influence the influence of lipid theory by government, industry and scientists had on our food consumption. Fat was out, more carbs were in. Processed food (margarin, Wesson oil, skim milk) was in, whole food was out (butter, taro, lard and whole milk) . Processed egg whites were in, and regular eggs were out (so much cholesterol in the yolk, a heart attack waiting to happen!) .As a young doctor (as is the case with other formative minds), I was all in on these theories and pushed them as much as anyone.
The result has been the worst escalation of diabetes, obesity and other chronic illnesses that we have ever seen. In urology, we are seeing an epidemic of erectile dysfunction and low testosterone in men (also likely related to the statin push in medicine … for all you biochemists, the building block molecule of Testosterone production is…you guessed it, Cholesterol! Any wonder why we all have low T..But I will save for another blog).
In short, WE WERE WRONG! Whole foods are better than processed, butter is better than margarin, and the list goes on. We’ve taken out the fat and put in refined sugars in foods which has been devastating. We’ve confused a whole generation of Americans of what is good food, and a whole generation of doctors who view the 3 digit cholesterol number as the barometer of mortality. While I believe cholesterol has some validity in understanding health, we’ve taken one variable in a complex, interdependent physiological system amongst a complex and variable species (black, white, Asian, urban, rural, sedentary, active, male, female, etc, etc, etc) and made broad generalizations for all. The longer I practice medicine the less convinced I am of the “purity” of scientific theories (some agenda is always driven), and question the relevance of these theories to be applied across a broad population. I encourage every clinician to challenge their basic assumptions and to asses where their own bias of correlation/causation may influence patient care.

Related Posts