I continue here with my thoughts about Randomized Control Trials (RCTs). RCTs are considered to be a tool that would help us determine the effectiveness of a treatment, by comparing the outcome of this treatment (called experiment) to a treatment that we already know about (called control).
One of the most fundamental tenets of RCTs is that members of the control group receive identical (control) treatment. Similarly members of the experimental group must receive identical (experimental) treatment, as well. It’s only when this happens that you can collect statistically useful data.
That's is not so difficult to ensure when considering RCTs for pharmacological or other similar “medical” treatments. Making sure that identical amounts of some substance are administered is rather easy.
But how do you do it when considering other types of treatment, which are not easily measurable?
One of the most fundamental tenets of RCTs is that members of the control group receive identical (control) treatment. Similarly members of the experimental group must receive identical (experimental) treatment, as well. It’s only when this happens that you can collect statistically useful data.
That's is not so difficult to ensure when considering RCTs for pharmacological or other similar “medical” treatments. Making sure that identical amounts of some substance are administered is rather easy.
But how do you do it when considering other types of treatment, which are not easily measurable?
That's just a methodological problem, the proponents of RCTs would say. You must design your experiment in such a way that there are no variations of treatment within the control and within the experimental group.
Is this the same, then, for psychotherapeutic or psychoanalytic treatments?
Hmmm. That would be a bit more tricky.
Eliminating treatment variation can only be possible when the treatment in question is structured and does not allow for any improvisation.
For example, one can design a "psychotherapeutic" treatment consisting of set questionnaires, exercises, tasks etc. Such treatment can "easily" be the subject of a Randomized Control Trial.
What about psychoanalysis?
Unfortunately, psychoanalysis is never structured. The only rule in a psychoanalytic session is that of free association. One patient might decide to speak about the events of the day, while another might prefer to speak about his or her worries for the future. A third might decide to recite a recent dream, or an event from his or her childhood. The variation is infinite.
Please, bear in mind that this is an intrinsic characteristic of psychoanalysis –it is not a problem.
How can we use RCTs to assess psychoanalysis then?
I am afraid we cannot. Apparently psychoanalysis is a treatment that cannot be “tested” via RCTs.
Does it mean that it is worthless?
Some people feel inclined to believe so. They are satisfied in their conclusion that whatever claims psychoanalysis makes, they cannot be tested.
This is their argument broken down:
I hope you can see that he are having a fallacious argument here.
Without our consent RCTs have been elevated from being just one of all possible means to collect evidence, to be the one and only means to collect evidence. They clearly are not.
What do we do then? How do we test psychoanalysis?
That's a very different question, and, admittedly, not an easy one. But we should not do like the drunken guy, in John Searle's metaphor, who, having lost his car's keys under the dark bushes, tries to find them under the street lamp because it was brighter there.
In short, as far as evidence is concerned, the question regarding the effectiveness of psychoanalysis, remains open.
I will return to this.
Is this the same, then, for psychotherapeutic or psychoanalytic treatments?
Hmmm. That would be a bit more tricky.
Eliminating treatment variation can only be possible when the treatment in question is structured and does not allow for any improvisation.
For example, one can design a "psychotherapeutic" treatment consisting of set questionnaires, exercises, tasks etc. Such treatment can "easily" be the subject of a Randomized Control Trial.
What about psychoanalysis?
Unfortunately, psychoanalysis is never structured. The only rule in a psychoanalytic session is that of free association. One patient might decide to speak about the events of the day, while another might prefer to speak about his or her worries for the future. A third might decide to recite a recent dream, or an event from his or her childhood. The variation is infinite.
Please, bear in mind that this is an intrinsic characteristic of psychoanalysis –it is not a problem.
How can we use RCTs to assess psychoanalysis then?
I am afraid we cannot. Apparently psychoanalysis is a treatment that cannot be “tested” via RCTs.
Does it mean that it is worthless?
Some people feel inclined to believe so. They are satisfied in their conclusion that whatever claims psychoanalysis makes, they cannot be tested.
This is their argument broken down:
- RCTs can be used to collect evidence that a treatment works
- Psychoanalysis cannot be tested with a RCT methodology
- Therefore Psychoanalysis is an unproven (and possibly worthless) treatment
I hope you can see that he are having a fallacious argument here.
Without our consent RCTs have been elevated from being just one of all possible means to collect evidence, to be the one and only means to collect evidence. They clearly are not.
What do we do then? How do we test psychoanalysis?
That's a very different question, and, admittedly, not an easy one. But we should not do like the drunken guy, in John Searle's metaphor, who, having lost his car's keys under the dark bushes, tries to find them under the street lamp because it was brighter there.
In short, as far as evidence is concerned, the question regarding the effectiveness of psychoanalysis, remains open.
I will return to this.
RSS Feed
