It seems like daily we are exposed to headlines implying that two variables are somehow connected. From there it is a short step to concluding that one variable is causing or strongly influencing the other variable. On the surface this is the “correlation suggests causation” camp of researchers. I would first like to ask these researchers how they decide which of two variables is influencing the other. In other words who is driving the bus? An example might be useful. This study suggests an association between sitting and anxiety.
http://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-015-1843-x
Correlation, Smorrelation
When the authors examined these variables they had to make some guesses about how one variable might be influencing the other variable. They had to come up with a plausible theoretical model to explain what they were observing. Their logic likely went something like this: Many studies have suggested that in general exercise is good for the brain and mental health. If this is true and you look at a group of sitters who aren’t exercising much, they should have more mental health problems like anxiety or depression because they aren’t getting the brain enhancing benefits of exercise. If your study shows a correlation between sitting and anxiety, your underlying thesis is supported. Of course scientists know that you need to take this one step further and do a randomized, controlled study. One group would exercise and the other group would not and then measures of anxiety would be assessed in both groups. Ideally these measurements would be nested in other measurements unrelated to anxiety so the study participants wouldn’t know exactly what was being measured. If the exercising group had less anxiety, your thesis would be supported but not proven.
Unfortunately controlled studies are much more time consuming and expensive than observational studies so they often never get done. Gary Taubes pointed this out in his New York Times article. Over time most scientists tend to give up in frustration. Their underlying thesis becomes “accepted knowledge” and the scientists move on to something else. This is often what happens in the real world as apposed to what we would like to see in an ideal world.
Just Who is Driving the Bus?
Let’s back up a little bit and look at the other possible relationship between these two variables—that being anxious somehow causes people to sit more. Most people would reject this possibility out of hand because it doesn’t match their own experience with anxiety. Common sense dictates that an anxious person is more likely to be pacing the floor rather than sitting. But you really don’t need to assume that sitting somehow causes anxiety because they could both be influenced by another unknown variable that produces this correlation. For example let’s assume that tea drinking causes anxiety and also pushes people to sit. If you didn’t look at tea consumption in your study it’s possible that a majority of study participants just happen to be tea drinkers. You would get a very different result if your study participants tended not to drink tea.
A Newer (and Perhaps Better) Theoretical Model
That’s why we have to be very careful about jumping to conclusions based on observational studies and we have to consider alternate theoretical models that look at new variables not previously explored. As a matter of fact, I believe that I can come up with a thesis that does make sense when it comes to examining the relationship between anxiety and sitting. My web site is dedicated to the concept that our diet of highly processed food can trigger a form of food induced brain dysfunction called Carbohydrate Associated Reversible Brain syndrome or CARB syndrome. We have observed that there at least 22 brain dysfunction symptoms associated with CARB syndrome and patients with CARB syndrome will tend to have most or all of these symptoms at the same time.
My Get Up and Go Got up and Went and I’m Anxious About it!
One of the symptoms of CARB syndrome is extreme physical and mental fatigue. It’s like your get up and go got up and went. People with this fatigue are much more likely to find themselves sitting a lot more than their healthy peers. People can develop fatigue for a lot of different reasons but the fatigue from CARB syndrome is by definition associated with the other CARB syndrome symptoms. Another classic symptom of CARB syndrome is excessive anxiety. Again a person can be anxious for many different reasons and there are numerous Psychiatric disorders where anxiety is a central symptom, but in these cases the anxiety is not usually associated with fatigue. Of course a person could have several Psychiatric problems at once leading to various combinations of symptoms being expressed at the same time, but from a scientific standpoint it always makes sense to find the simplest explanation for what you are observing.
Your Diet Can Simultaneously Affect Your Energy and Mood
I believe that CARB syndrome is now so common that most people who sit and are anxious likely have the disorder. There may be a few folks with a true anxiety disorder and fatigue say from a serious medical condition but of course they would be much less common that those with CARB syndrome. If they have other symptoms of the disease it provides further support that you are dealing with a single condition.
For the sake of argument let’s assume for a moment that I am right. If that is the case then in many people it is highly processed food that is causing both the fatigue and anxiety. Why is it so important to choose the best underlying thesis, the one that best reflects the real world? It makes a big difference when it comes to doing a controlled study to support your thesis. If your theory is that exercise improves brain function and thus lessens anxiety then you will take two groups of similar folks and have one group exercise on a regular basis and the other will not. At the end of the study if the exercising group has less anxiety then these results will support your thesis. But be careful. If your study participants have a traditional type of anxiety disorder and don’t have CARB syndrome, then they likely will have less anxiety when they exercise. I agree that exercise is good for brain function in relatively healthy people. This outcome is far less predictable when dealing with CARB syndrome patients. I believe that exercise is an important component of an effective treatment program for CARB syndrome but its effectiveness will be strongly influenced by whether or not the other components of a treatment program—for example a healthy diet, are also present. Although they also often have less anxiety when they exercise, you also need to consider these points:
- If you don’t control for diet then participants with CARB syndrome who consume a lot of highly processed food will look much different than those who eat a healthy diet. Each scenario will result in a different level of anxiety. In some people the anxiety simply disappears.
- Those with CARB syndrome will have a much more difficult time exercising because of their underlying fatigue. When they do exercise the intensity of their exercise will likely be less than people without the illness and intensity of exercise is challenging to measure in this type of study. This will tend to further skew the results in unpredictable directions.
- Normal healthy people with fatigue will react much differently to exercise than those with CARB syndrome so the study results will vary tremendously depending on who is in the study. If CARB syndrome is a “real disease” and I believe that it is, you will need to control for this variable.
- Because CARB syndrome is a new disease model, virtually nobody is controlling for it in their controlled studies, calling into question many study results that are focused on the connection between diet and brain function.
If CARB syndrome is the disease you have chosen as your theoretical model, then your controlled study will likely look at how different diets affect the symptoms of the disease. A Paleo diet should result in improvement in core CARB syndrome symptoms whereas a diet loaded with highly processed food should have the opposite effect. Because CARB syndrome is associated with metabolic parameters like obesity, insulin resistance and type 2 diabetes, these markers should also improve on a Paleo diet.
At the present time we don’t have strong evidence based on controlled studies to show us which of these competing theoretical models is more likely to be correct, and who knows—there might be other models out there that we haven’t even considered. That’s where empirical science can come to the rescue. Empirical observations are not worthless. If just about every time you try to cross a busy highway you get hit by a car, it is likely safe to assume that attempting to do so is not a good idea. There likely is a cause and effect factor at play in this scenario. Empirical observations can also be a guide to what type of controlled study you should pursue. After all, you don’t want to waste a lot of time and money studying sideways falling apples!
When the Gold Standard Fails Us
Another common problem is poorly done controlled studies. Although double-blinded controlled studies are considered to be the gold standard in biological research, if these studies are done poorly they do more harm than good. A recent example might help to clarify this point. One of the core components of my CARB syndrome treatment protocol is a low carbohydrate Paleo style diet. I also am a big fan of high fat (healthy fats) ketogenic diets. Recently headlines around the world noted that a new study shows that a Paleo style diet is “unhealthy and fattening”. This was supposedly based on a controlled study in mice. The problem is, it was a very flawed study. Aaron Blaisdell did a nice job of slicing and dicing this study in his latest blog post. Of course his perspective won’t have the reach of the study he criticizes because it’s not as “newsworthy”.
I know that I haven’t answered a lot of important questions today. What I have tried to accomplish is to point out the complex relationship between theoretical models (theories), empirical observations and observational studies and controlled studies. Despite all this complexity, when it comes to your individual health you can grab the bull by the horns and decide for yourself whether the CARB syndrome model or any other model for that matter has merit and might apply to you. Because the treatments based on this model are very low risk, after a careful risk/benefit assessment you can decide whether or not to try them. If your health improves it may be a chance occurrence rather than the direct result of the treatments that I recommend. In either case if your health and ability to function improves, I can’t see much down side. Until all the controlled studies are done, the risks and the results of a given approach will always rule. As a practicing physician with decades of clinical experience, I think that most people recognize when their health and well-being are improving. I guess we could call it a study with just one participant and that person would be you.
Indeed science by nature tends to be very messy, yet without it we would be lost in complex and dangerous world. If we have a good understanding how science works in the real world it can provide us with the tools we need to gain a better understanding of the world around us. This in turn will allow us to manipulate that world in way that is beneficial to our health and well-being.