Is Nutrition A Soft Science?: Problems with Nutrition Research

Someone once asked me, "Is nutrition a soft science, like psychology?"  

There may not be a straightforward "yes" or "no" answer. So, I will warn you right now, this might be a long post...

Some areas of nutrition overlap with "harder" sciences, like chemistry and biology. We know what a calcium ion looks like (chemistry!). We know that there are different forms of some minerals, like magnesium oxide vs. magnesium glycinate (again, chemistry). We know that sailors who didn't get enough vitamin C developed bleeding gums or scurvy (biology). 

But, other topics within nutrition are more "softer." They are more hotly debated and the research is conflicting.  

Let's explore this further...

Finding a Diet for Diabetes

Take, for example, the ketogenic diet or a high-fat, low carb diet. Some scientists and healthcare workers propose that it may be good for diabetes because carbs raise blood sugar. Repeatedly high blood sugar over time can result in complications, like neuropathy and kidney failure. In extreme cases, patients can enter a diabetic coma and even die. 

Photo Credit Sharon McCutcheon

Sounds like a good argument to avoid carbs then, right?

Well, no...Others say that a moderate-carb diet is better. Doctors and dietitians recommend "carb counting" or eating a certain number of carbs (often 45-60 grams per meal). This doesn't mean eating refined sugar, but rather carbs from grains, legumes, tubers, vegetables, fruit, and dairy. They argue that your body needs carbs and that people with diabetes may develop low blood sugar without them. There are people who are admitted to the hospital because their blood sugar drops too low and they can develop complications and even die. 

Photo Credit Monstruo Estudio

Wait, what? So, who's right?

Good question. The answer is still under debate because there's research to support both!

If you're interested in reading the research yourself, L. Gupta, et al. (2017) put together this review article that looked at studies of the ketogenic diet and diabetes. In general, measurements of diabetes like Hemoglobin A1c (HgbA1c) and fasting blood sugar improved. This suggests a ketogenic diet might be helpful. 

On the other hand, Drs. Ley, Hamdy, Mohan, and Hu (2014) put together a review supporting a carbohydrate-rich diet. They recommended eating whole grains, fruits, legumes, and vegetables--most of which would be avoided or limited on a ketogenic diet!

Photo Credit Brooke Lark

This is why it's important for people with diabetes or other health conditions to work with a doctor and a Registered Dietitian. They can help you determine what is right for you. 

Why do we have conflicting results?

One reason is that human studies can be hard to control. Unlike animals, we’re not locked in cages being fed what the researchers give us and only doing things they want us to do.

Usually researchers just ask participants to log food.  If participants didn’t accurately report what they ate to the researchers, this may affect results.

Photo Credit Ava Sol

Studies on nutrition often have confounding factors, too, which I will discuss more in a minute.

But, first, let’s look at some other issues with studies on particular foods.

Studying Individual Foods

Often, studies aim to determine whether or not a food is good for us. Is wine good for the heart? Does saccharin cause cancer?

In my last post, I wrote about wolfberries a.k.a. goji berries, but I didn’t share much research with you for several reasons.

First off, studies aren't always done in humans. Long and colleagues studied chickens who had wolfberry extracts mixed into their corn & soybean feed. Wanping Aw studied mice after his team first made them sick by giving them a chemical called dextran sodium sulfate. The results of these studies may not apply to humans.

Photo Credit Oxana Kuznetsova

Second, the doses used in studies aren't what you'd get from eating a real wolfberry (or whatever food is being studied). The recommended serving size for dried wolfberries is 40 grams or 1/4 cup. The above chicken study fed them 4 grams per kilogram body weight each day for 6 weeks. So, a 150 pound person would have to eat 272 grams of wolfberries or almost 7 times the recommended serving!

Even when studies are done in humans, the results don’t have clear implications for other humans, because of correlation and confounding.

Correlation and Confounding

I’ve mentioned these terms twice now. What are "correlation" and "confounding"? 

Correlation means that two things seem to be related but one may not necessarily cause the other. For example, maybe we look at a group of people living in China and say, "They eat wolfberries and they have fewer chronic diseases than Americans." This doesn't mean that eating wolfberries prevents chronic disease. 

Confounding means there might be other factors. One factor might be their genetics. Maybe it's due to other foods they eat. Or, maybe many of the study participants lived on farms instead of cities. Maybe they have better stress management skills, etc. It's hard to know the reason they had fewer chronic diseases. 

We see correlation and confounding in other types of science, too. For example, if we run a clinical trial of an antidepressant, the results may be confounded by other factors, like counseling and physical activity.  That said, in medicine, we can run placebo-controlled studies, where someone is given either the active drug or an inactive version...which is harder to do in nutrition. We can't make a placebo-controlled study of whole foods, like berries, because the participants would know if they were eating berries or not.

Photo by Cecilia Par

So, then, how do dietitians know what to recommend?

Healthcare providers, such as doctors and dietitians, don't rely on the results of a single study. They often use guidelines by professional organizations, such as the Academy of Nutrition and Dietetics, the Society for Critical Care Medicine, or the European Society of Enteral and Parenteral Nutrition. These organizations complete reviews. That is, they sort through the available research to try to make recommendations based on the results.

For example, the Academy of Nutrition and Dietetics Evidence Analysis Library grades the available studies as good/strong evidence, fair, limited/weak,  expert opinion only, or not assignable. See screenshot below.

What does this mean for the blog?

I am planning to write more superfood posts. But, I will try to avoid claims, like that a particular food is an immune booster. My posts will focus on what science says, like that wolfberries contain vitamin C.