What is a Randomized Controlled Trial (RCT)?

Positives and Negatives of RCT's

nutrition scientists
Hero Images/Getty Images

A randomized controlled trial is an experiment to judge and compare the outcomes of different treatments. This is different from an observational study, where people are simply observed or tracked, looking for what might be roughly called "things that go together". For example, you might observe a group of people and find that the ones that ate the most bananas got in the fewest car accidents. Did the bananas prevent the car accidents?

You don't know until you do an experiment where you tell one group to eat bananas and the other not to eat bananas, and then you check the accident rate. Otherwise you don't know if people who are more careful tend to like bananas more, or some other factor that might link up the two things.

A randomized controlled trial is an actual experiment, not just an observation. The "randomized" part is very important. For example, in diet research, if everyone in the study was allowed to choose what diet they wanted to eat, you don't know if there is something about the people who chose a certain diet that makes them different from people who chose a different one. Also, randomization is way of trying to make sure that the groups have roughly equal mixes of various characteristics (some that you may not even know about) that might influence the outcome.

The "controlled" part of an RCT means that you try to account for anything you think might make a difference in the outcome.

In a diet study, the ultimate control is a "metabolic ward study" where your subjects all live in an area where all their diet inputs and energy outputs (exercise and other movements) are carefully measured. As you've guessed, this is not often very practical. The next most effective diet control method is to give people all the food you want them to eat.

Otherwise, you have to resort to careful education and followup. Or, as is all to often the case, not-so-careful education and followup.

It's also important to control characteristics of the participants that you think might affect the outcome.For example, if you have reason to believe that diabetics will respond differently to a diet than non-diabetics, you may choose to either exclude diabetics, only use diabetics, or analyze their data separately. These types of "co-variates" are often accounted for by statistical methods as well.

There is a possible shortcoming to RCT's that is creeping into current diet research. It's important to realize the danger that RCT's will wash-out the effects of subgroups.  For example, in this diet study, when it was first analyzed the researchers reported that the different diets had about the same effect.  Later, though, they looked WITHIN the diet groups and realized that there was a huge variation in how people were reacting.  So even though the averages were similar, some people were responding very differently.

  In this case, it turned out that people who were insulin resistant were more apt to get positive benefits from the low-carb diet than the low-fat high-carb one.

In other words, if a diet is affecting different people in opposite ways, it can look like the groups were similar if you just look at averages.  At this point in diet research, it is still more common to see RCTs with little attention to who is benefiting and who is not. I think that in the future this will become more and more important in diet research, as we gain knowledge about, for example, genetic differences that influence how different people respond to different foods.

See Also: How to Read a Diet Study

Who Benefits Most From a Low-Carb Diet?

Continue Reading