Nobody seems to have pointed out that there is a logical fallacy at the heart of the NICE guidelines, which put randomly controlled trials (RCTs) at the top of the list of approved evidence. The basis of RCTs is that it is a question of cause and effect. By setting up properly controlled research studies we can isolate the causative factors involved. If CBT comes out well on such trials, it must be because the technique of CBT is responsible, and that therefore CBT has to be recommended.

The fallacy involved here is that is has been shown fairly conclusively now, as neatly summed up in books like those of Miller, Duncan and Hubble,¹ that it is not the therapeutic techniques that produce the good or bad outcomes. The research shows that there are four main factors involved in therapy: what the client brings; how good or bad the relationship is; what the therapist does; and expectancy, hope and placebo factors. What it also shows is that this represents the order of effectiveness in producing good outcomes. What the client brings, in terms of readiness to work, is the most effective factor, accounting for about 40 per cent of the results. The relationship comes next, and we now understand, with the help of relationship theory and research, why this should be. Smaller still is the contribution of the therapist and the techniques of therapy.

Yet the whole vast apparatus of the RCT system is on the techniques of the therapist. It is these techniques which can be most easily measured and controlled; it is these techniques which can be manualised in technical handbooks. And it is these techniques which are the least relevant factor in accounting for outcomes.

In other words, the RCT enthusiasts are measuring away, getting more and more accurate and effective, but attending to the wrong thing. These researchers, it seems, are doggedly measuring techniques and taking these to be the most important variables, as if they just cannot believe that it is not the techniques that do the work. It must be, they cry, it just must be – how could it be anything else? But this is irrational, just as irrational as the famous case of continually going into the same channel in the maze because that is where the cheese should be! But measuring the incidence of the techniques because that is where the cause should be is just as irrational.

This is such an obvious point that I find it hard to understand why it is not made more often. Anyway, here is it for now.

1. Miller S, Duncan BL, Hubble MA. Escape from Babel: toward a unifying language for psychotherapy practice. New York: WW Norton; 1997.