I came across this lecture he gave called "Science For Smart People" and it is without a doubt the best (and funniest) crash course in epidemiology that I've seen.
If you watch this you will know more about health and nutrition science than 99% of people out there.
Correlation Does Not Equal Causation
To interpret research, it is critical to recognize the difference between the two types of studies: observational studies and intervention studies.
In an observational study, there is no intervention or treatment. The researchers observe the subjects over a period of time and gather data about them.
Observational studies use mathematical methods to crunch the data and find out whether a certain trait or behavior is associated with a particular outcome.
These studies can show, for example, that A (drinking) and B (depression) are associated, but they can not prove that A actually caused B.
Clinical Trials or Intervention Studies
In clinical trials, there are two or more groups that receive a different type of treatment. Often there is a group that receives no treatment at all (called the control group).
These sorts of studies are capable of demonstrating causation, e.g. that A caused B.
It is common that something that has been "proven" in an observational study turns out to be completely wrong when tested in a clinical trial.
Clinical trial and observational study. Remember that.
This Has Caused a Lot of Unnecessary Harm
Many of the high-impact headlines you see in the media are based on observational studies and the reporters make it seem like these studies prove something.
For example, the Nurses Health Study, one of the largest epidemiological studies ever conducted, showed that women who took hormones after menopause had a lower risk of heart disease.
However, when randomized controlled trials were conducted, it turned out that the hormone drugs actually increased their risk of heart disease.
I wonder how many women got heart attacks due to people buying into this observational study that turned out to be wrong?
I also wonder about the old observational studies that showed that saturated fat was associated with cardiovascular disease.
These studies made big headlines and gave rise to the damaging low-fat, high-carb dogma that probably played a major part in the obesity and diabetes epidemics.
I wonder how many deaths have been caused by the failed nutritional policy of the last few decades, based on observational studies that turned out to be wrong.