I spend hours every day reading studies, articles, and researching health-related matters. When I find a new publication or exploration of a topic, I get excited to dive in. That being said, some studies and articles are more useful than others.
One type of study that is used frequently to make health claims and guide public policy is an “epidemiological study”. Epidemiology is the study of a set population, or group of people, to develop correlations or inferences.
The problem is that these do not prove anything. When we find a strong correlation between factors, we should use that as a starting point to conduct further research. An epidemiological study, by itself, should never be the basis for making health policies.
Let me give some examples.
Epidemiology suggests that soy is a healthy incorporation in a diet. This is due to the fact that Asian countries consume high amounts of soy on a regular basis and don’t experience the same health problems as Western nations.
However, no other factors are taken into account.
The soy that Asians consume has not been genetically modified to the same extent as ours, nor has it been grown in soils depleted of minerals. Also, most Asian dishes use fermented soy or the bean in its natural state.
Asian cultures consume more wild-caught fish (high in anti-inflammatory omega-3s), sea vegetables (loaded with vitamins and minerals), and opt for white rice, with less anti-nutrients and gut-damaging proteins than typical “heart-healthy” whole grains such as wheat and oatmeal.
Historically, Asians don’t consume as much processed food as Americans. They don’t cook in corn or canola oil, they don’t have packaged foods at every meal, and they don’t go out to eat as often.
And finally, they are far more active – walking, biking, and taking the stairs as part of daily life.
Because of these factors, we cannot confidently say that the consumption of soy in Asian countries is the cause of their better health.
When we look at soy mechanistically, we find phytoestrogens that have the potential to skew hormone levels, leading to fat-storage and growth of cancer cells. It is extremely high in inflammatory omega-6s. Take into consideration our growing practices, extensive refinement process, and consumption of soy byproducts, and soy consumption in the US no longer seems as safe.
Another example of epidemiology lacking substance:
In March of this year, there was a headline stating: “Animal protein-rich diets could be as harmful to health as smoking”. These news reports were based upon two studies: one epidemiological study of over 6000 adults and one study of mice in a laboratory.
The results of these studies suggested that a high protein diet (over 20% of calories) was “positively associated with diabetes-related mortality”. When you look at the numbers, one person in the “high-protein” group (consisting of over 1000 individuals) died from diabetes.
The lead researcher running this study owns a plant-derived protein supplement company…explaining the claim that only animal-protein is dangerous.
Some other issues:
There was no way to control for protein quality. There has never been a study showing negative outcomes from consumption of wild-caught fish, grass-fed beef, or eggs from pasture-raised chickens.
The mice that experienced growth of cancer tumors were implanted with melanoma cells before the study began. Plus, the study found that high protein consumption was “not associated with all-cause, CVD, or cancer mortality”. Therefore, the protein-cancer correlation was in fact disproved.
Finally, diet was self-reported. The average participant reported consuming 1,800 calories a day…30% lower than the national average. This suggests major under-reporting.
So, even though the study was riddled with flaws, and actually found no increased risk from animal-protein consumption, the results were phrased to dissuade individuals from consuming meat.
To get back to my original point – epidemiology is used too often to prove a pre-existing belief, promote a political agenda, or increase profits.
By itself, epidemiology is no different than trying to claim that the number of birds flying over a particular region somehow determines cancer rates in that area.
Certainly we should use any research tactic available to ask questions and form a hypothesis…but ultimately, we need to examine issues in every way possible.
Once we’ve investigated mechanisms, done cohort studies and some “food-diary” studies with pictures, it’s time to form a hypothesis and conduct a blinded, crossover, metabolic ward trial to draw some real conclusions!