Nutrition research is slower moving than pharmaceutical research and widely misreported but researchers have enough data to provide meaningful advice on major food categories.

One thing more or less all nutritional researchers would likely agree on is we don't know everything about how healthy foods are. Most evidence is circumstantial and should be read with significant skepticism, particularly if an effect is being reported for the first time.

'Stephen Pauker, a professor of medicine at Tufts University and a pioneer in the field of clinical decision making, says, “Epidemiologic studies, like diagnostic tests, are probabilistic statements.” They don’t tell us what the truth is, he says, but they allow both physicians and patients to “estimate the truth” so they can make informed decisions.'

- Excerpt from Do We Really Know What Makes Us Healthy by Gary Taubes

All large trial techniques are expensive and time consuming. Pinpointing a specific effect on a specific food is even more difficult than seeing if a pharmaceutical drug has its intended effect for a number of reasons. Researchers can not perform studies using the golden standard double blind technique in randomized control trials because test subjects will always know if they are eating chicken or broccoli while they have a much harder time telling if they are taking a pharmaceutical or a placebo sugar pill. The time period in which quality nutritional studies can be conducted is also much longer than pharmaceutical studies. While test subjects in pharmaceutical trials can typically tell if a treatment is working within a few months, the effects of food on mortality can take decades to be evident. Further, for obvious ethical reasons, clinical trials can rarely be used on studies of potentially harmful effects (i.e. modifying a non-soda drinkers behavior to include 3 sodas per day for the next year). The resulting limitations mean nutritional research is often derived from observational data using surveys. Memory also creates problems in nutritional studies as its easier for study participants to recall about how many cigarettes they smoked daily compared to what they usually ate in the last month. Finally, even when effects from foods are shown the study will need to correct for many other factors such as income and education level need to be corrected.

As a result, research quality varies significantly. The chart on the left below illustrates the traditional hierarchy of scientific evidence - placing systematic reviews and randomized control trials as most reliable. The chart on the right below shows a revisited hierarchy proposed in 2016 by researchers from the Mayo Clinic, illustrating that while ranomized control trials are typically better than cohort studies, well-designed cohort studies can be better than poorly designed randomized contrials and so on. Separately, it makes the point that systematic reviews can also vary in quality and are different from and a lens in which to view observational studies.

Traditional Hierarchy of Scientific Evidence
2016 Revisited Hierarchy of Scientific Evidence

Providing accurate dietary advice is critical because eating plays a hugely important role in health. Yet, widely-reported news stories and changes in the most popular fad diets give the false impression that the best dietary advice changes on a whim. Take for instance, a study of dietary advice in the top 10 selling UK newspapers graded on evidence scales developed by the World Cancer Research Fund (WCRF) and the Scottish Intercollegiate Guidelines Network (SIGN). Of 111 dietary health claims identified, 72% and 68% (assessed by WCRF and SIGN criteria respectively) had levels of evidence lower than the convincing or probable categories that are recommended for dietary health claims [1].

"It can certainly be confusing for members of the public to know what they should and shouldn’t be eating, and in what amounts. From saturated fats to sugar – it can often feel like navigating a minefield of inconsistent nutritional advice. It’s important to remember that information referred to as ‘expert advice’ is sometimes based on evidence that has known limitations. For example, it may involve a small sample of participants or ignore other factors that could affect the results."

- Excerpt from Healthy eating advice must be based on the best evidence - Public Health England

If you found this article interesting, you might also enjoy our free receipt analysis service. Upload your food receipts and our web app can instantly tell you cumulative stats from your purchases with important metrics like how much sodium and added sugar you’re getting relative to calories. It’s like a calorie counter but much faster to use. You can sign up here.