I was planning to write about Evidence Based Practice, but stumbled upon the question of Randomized Control Trials (RCTs).

I think that complex questions can be rendered a bit more manageable when broken down to their constitutive parts. RCTs form a constitutive part of any Evidence Based Practice, so I will start from them.

Let’s see.

In the literature we read: “Randomized control trials are the most rigorous way of determining whether a cause-effect relation exists between treatment and outcome and for assessing the cost effectiveness of a treatment”.

RCTs utilize powerful statistics to separate signal from noise and have the advantage that you don’t really need to have a comprehensive theoretical model to account for what you are researching and why. The only thing you need to be careful about is how you conduct the trial.

I am calling the absence of a model advantage, but it is questionable whether it actually is an advantage. To be sure, in much the same way to a computer that boots itself up from "nothing", RCTs help you to get something from nothing. For example, you can have a framework of mutually coherent hypotheses made up in little time, starting from one or more completely random research questions and building on the evidence you gather (or fail to gather).

One very characteristic example of this is the Dopamine Hypothesis of Schizophrenia.

The completely random research question was: “Is there any correlation between symptoms of schizophrenia and a disturbed dopaminergic signal  transaction?” I call it random because there was no deeper reason whatsoever to ask this question instead of any other.

One could ask, for example: “Is there any correlation between symptoms of schizophrenia and blood cholesterol levels?” Apparently the second hypothesis, if at all examined, led to no results, while the first did –much to the satisfaction of the flourishing pharma-industry, but that's beyond the point I am trying to make.

What I am trying to say is that we don’t have any conceptual understanding why disturbed activity of dopamine receptors should correlate to symptoms of schizophrenia. Don’t take me wrong, I am not arguing that they do not correlate –they may or they may not; I just point out that we don’t have a conceptual understanding why, and we don’t have a compass that would allow us to formulate other, similar questions and test them.

Now from some aspect this is certainly an advantage. Schizophrenia, to stay within our example, is a extremely complicated phenomenon, especially when we try to approach it in terms of neuronal functioning, and it is not easy to see how we could construct a theoretical model in absence of any knowledge about its internal workings.

So, RCTs can and sometimes do contribute to the study of complicated phenomena by identifying possible or potential causal relations (or hypotheses of relations), but they do not really worry about understanding those phenomena. As this happens by design, it is completely acceptable.

But RCTs “know” that at some point their contribution will reach its logical limits; at that point a human observer will need to take over and try to formulate a coherent conceptual theoretical model that would be able to account for all confirmed (via RCTs) hypotheses.

To give a measure of proportions, the Dopamine Hypothesis of Schizophrenia is still very very far from that point.

I stop here, but I will return to the question of RCTs.