Taking the example of the various RET outcome studies that were analyzed in the Engels paper: alright, so, it’s been quantitatively determined that, on the whole, patients who were treated with rational-emotive therapy did significantly better than patients who weren’t treated at all or were treated with “placebo,” but did no better on average than patients treated with other forms of therapy that’ve also been shown largely effective with many different types of patients. Ok, so now how exactly do we interpret this finding? For instance, how do we know it doesn’t actually reflect the fact that for half the patients in each of these studies, rational-emotive therapy was the most appropriate and effective treatment possible, given the particular nature of their problems (e.g., an extremely distressing irrational worldview combined with a well-developed reasoning capacity and a receptiveness to criticism), and for the other half of patients rational-emotive therapy was actually the dead-wrong therapeutic approach to take (e.g. because they were already convinced of the irrationality of their beliefs but couldn’t motivate themselves to quit drinking or washing their hands compulsively or what-have-you, or because they hadn’t reached the level of self-insight to be able to articulate their beliefs in the first place)? Or, alternately, how do we know that some 90% of the participants of each of those studies did no better through RET than they would’ve done through pure CBT or systematic desensitization or plain old-fashioned talk therapy, because only 10% of study participants actually possessed the characteristics (whatever they may be) that make someone an excellent candidate for RET – and since the studies in question lumped those patients in with all the others, the data washes away the dramatically superior effect of RET on that minority subgroup within the much larger, hopelessly heterogeneous sample of participants? And even setting aside patient heterogeneity, how do we know that only a fraction of the therapists administering the RET weren’t actually doing it in an effective, appropriate, client-tailored way, and that RET isn’t simply more sensitive to therapist competence than are the other treatments it was being compared to (thus introducing still further unintended and uncontrolled variability into the variables allegedly being manipulated)? Taking just this last example, if it were indeed the case that a particular treatment had potential to be dramatically efficacious – for certain kinds of patients, of course – but was also particularly sensitive to the mediating variable of therapist competence (or even a particular aspect of therapist competence, such as a therapist’s perceived genuineness and warmth or ability to motivate the patient), I would be extremely interested to know that—as would any therapist or therapist-in-training, I imagine. And yet we would never glimpse the interactive or mediating effects of such variables on patient outcomes from the kinds of analyses described here, I don’t think, because the studies don’t break down their variables anywhere near as finely as that. The finest discrimination they make with regard to the nature of the therapy, as far as I can tell, is “degree of emphasis on cognitive techniques” (“exclusively cognitive,” “primarily cognitive,” and “equally cognitive-behavioral”); with regard to therapist competence, the one relevant mediating variable they mention is “experience of therapist” (which we already know from prior research to be an incredibly shabby predictor of actual competence); in fairness, they did examine certain client characteristics—such as “socioeconomic status, sex, age, intelligence, and type of complaint” – which is a reasonable start, but: what about level of client insight? What about degree of ruminative thinking style or anhedonia? What about chronicity vs. situation-specificity of their mood disorder (supposing it’s a mood disorder they’ve got)? What about comorbidity? And as for “intelligence,” well, that’s a great starting place, but – what type? Are we talking verbal intelligence, mathematical prowess, abstract reasoning skills, or what? The WAIS fullscale IQ score that no doubt serves as the sole measure of “intelligence” in the majority of these studies would not allow for such fine differentiation. And yet I’m quite strongly convinced that the kinds of variables we would need to look at to actually uncover any of the basic causal relationships underlying therapeutic change would be at least this specific and narrowly defined, if not more so. Instead, we are looking at prepackaged conglomerations of largely unknown and uncontrolled variables, with no real, systematic reason for the packaging except that someone did it that way before or it makes for a convenient experimental design. What is it about the “primarily cognitive” package of treatment techniques that might work better or not work better than the “primarily behavioral” package? If we can’t answer this question—what about it (i.e., what’s the mechanism?)—then we can’t really interpret any result we might get. If it turns out to work, we still don’t know why it worked; and if it turns out not to work, well, we still don’t know why it failed.
I think it's revealing, and somewhat ironic, that, of all our readings for this week, the one I've so far found most compelling--and most illuminating into the nature of RET and the conditions under which it might be effective—-was Ellis’ own 1999 article, “Why Rational-Emotive Therapy to Rational-Emotive Behavior Therapy?" In that article Ellis makes a great case for regarding (and implementing) RET as fundamentally a behavioral as well as a cognitive treatment—and he makes this argument with reference to the basic nature of RET’s intended mechanism. No amount of armchair intellectual discourse and purely theoretical reasoning can truly change a person’s conviction until that person actually experiences, in a concrete, visceral way, that the conviction bears out in reality—and the only way to solicit such experiential proof is by behaving differently first. Did the “primarily cognitive” therapists in the studies analyzed by Engels et al not subscribe to this view? Or did they incorporate the essentially behavioral components of RET in much the same way that Ellis describes doing it, but because they dressed it in what’s traditionally packaged as “cognitive” rather than “behavioral” lingo they were rated as “primarily cognitive” according to the Engels scale? It’s well nigh impossible to tell. Ironically, Ellis warns in his article about precisely the kinds of client as well as therapist variables that I’ve noted are prominently absent from the Engels analysis (and most other analyses like it that I’ve read to date): regarding clients, he writes, after a detailed description of the specific client characteristics that do lend themselves well to treatment with RET (along with fairly compelling explanations of why they lend themselves well!), “Many clients have special kinds of disorders or may react idiosyncratically to therapy and therefore may not benefit from the usual procedures of REBT or cognitive behavior therapy (CBT).” Are you listening, Engels (and psychotherapy outcome researchers everywhere)? And he goes on to cite various particular reasons why this might be: e.g., “because they are temperamentally opposed to them, will not do required homework, are hostile to the therapist, get neurotic gains from their disturbances, are convinced that they are hopeless, do not want to risk getting better, or for a variety of other idiosyncratic reasons.” And as regards the variable effectiveness of therapists: “Similarly, therapists who try to use the methods of REBT and CBT and do so ineffectually may not truly understand them, may be temperamentally opposed to them, will not take sufficient time or energy to apply them... As many research studies have shown, the success of therapy depends on many relationship factors between clients and therapists” [ding ding ding! Just goes to show us that everything is connected…] “Negative aspects of the relationship may interfere with REBT and CBT techniques that are usually effective but do not work for a particular client [and] therapist… Some clients favor one or a few techniques that may not be the best for them and even favor irrational or inelegant methods that rarely work. What does this mean for therapists who wish to be effective for as many of their clients as possible? It means that they can specialize in a particular mode of treatment… [but] should be prepared for clients to resist their ‘best’ methods and require different and perhaps ‘inferior’ ones."
How come Ellis was able to articulate so eloquently all these limitations and contextual constraints of his own therapy — the one he himself designed and staked his whole career on (talk about a “therapist allegiance!”) - whereas some of the most rigorous and self-avowedly impartial of psychotherapy outcome researchers seem to overlook them fairly regularly? I daresay that at least some part of the explanation (and yes, my broken record I shall persistently replay) is that Ellis was more theoretically minded - in the best sense of the term, which also closely corresponds to the quality of being causal-mechanism-minded - than today's psychotherapy researchers tend to be. As someone hoping to embark, by slow and cautious degrees, on my own eventual career as a psychotherapy researcher (of one sort or another), I hope to shake things up by stirring a bit more of the Ellis-esque mentality back into the stagnant waters of today's psychotherapy research establishment.
No comments:
Post a Comment