In the quest for the ‘best diet’ for weight management, a google search of clinical trials will land you with all manner of conflicting results. Some studies conclude that a particular diet works for weight loss whereas the next one says it does not. So what gives?


Here is what I view as the Top 3 Problems with dietary research: (see this discussed in more detail in a recent JAMA Viewpoint article)


  1. For any particular type of diet, many variations exist. For example, does a ‘high protein’ diet include animal proteins, plant proteins, or both, and in what proportion? In a ‘low carb diet’, how low are the carbs exactly? And how much of the fat is saturated vs unsaturated?
  2. A diet may work in the short term, but not in the long term. This is because in many cases, we can only stick to a diet for so long before biological factors take over – the hunger hormone ghrelin is raging, our fullness hormones are in our boots, and our metabolism slows down, all of which favor dropping the diet and gaining weight back. Many studies concluding that a particular diet results in weight loss are way too short, with longer term studies of the same diet type usually showing no benefit over other types of diets. I would say at least 2 years is needed to really know whether a particular diet has any benefit.
  3. Adherence to the diet in a clinical trial is difficult to assess. We know that when people are assigned to a diet and then asked to report how much they actually follow the diet, they often don’t follow it very closely. On top of that, we also know that many people will report following the diet better than they actually do (meaning their overall adherence is even worse). This is called ‘social desirability bias’ – meaning the participants want to make the researchers happy, so they tell them that they followed the diet even though they didn’t. Some very excellent studies get around this by having participants live at their research facility for weeks at a time such that food intake can actually be controlled – but, as noted above, long term studies are really needed to know if a diet works long term, and I don’t foresee anyone being admitted to a research institution for 2 years!


For interested researchers, the JAMA article makes suggestions as to how dietary research can be improved. I like the idea of incorporating metabolomics, which can measure substances in a person’s body which reflect how well they have adhered to a particular diet (though I suspect they will confirm poor adherence, leaving us really no further ahead).


In the meantime, what works best? As I have said many times before – the approach to lifestyle changes has to be two things: individualized, and permanent. Finding the right pattern of eating that works for you, taking into account the foods you do and do not like to eat, cultural preferences, social circumstances and so forth. Eating in a way you do not like is unlikely going to be something that you can stick to in the long term, and we know that any lifestyle change has to be permanent if it is going to help manage weight long term.


Dr Sue Pedersen © 2019 

Follow me on Twitter! @drsuepedersen