“Hypocritical Oath”

Lilienfeld, 2007

Of this week’s five articles, I found the by far the most compelling one to be (the always controversial) Scott Lilienfeld’s manifesto on Potentially Harmful Therapies (PHTs). Taking on quackery, DARE, and experiential therapies in one fell swoop, Lilienfeld sets forth a powerful argument for the real and present urgency of addressing PHTs in clinical psychology ­now – before, even, taking another step towards establishing firm ESTs. This is a powerful idea of which, I admit, I was quite skeptical at the outset. However, having walked Lilienfeld’s path of logic (sic) with him, I am largely compelled by his claim of importance in addressing PHTs

To catalogue the interesting and/or contentious ideas discussed in this article would be both overlong and boring. So, I will touch on some of what I see to be the most intriguing ideas:

-I have not before heard a more damning portrayal of the impact of the “file-drawer effect” and its brother, “Type III error” (p.57). That all finding should be somehow reported is, of course, prescriptively wonderful, but awfully hard to make happen, especially given the (necessary) rigors of the peer-review process. That the lack of reporting of Type III error may very well be masking a body of research demonstrating iatrogenic (at least) effects of various treatments is just downright scary. Though the International Journal of Non-Significant Results represents a step in the right direction towards resolving this mishap, the notion of “enrolling all intervention studies in a centralized data bank that is publicly accessible” (p. 57) is probably a better idea.

-I was particularly heartened to see Facilitated Communication (FC) listed on the provisional list of PHTs. Having seen firsthand the amount of false hope that such interventions can provide to families that are desperately seeking real hope, it is worrisome to me that its prevalence has not diminished at all over time. I find it telling, too, that, in additional to FC, several of therapies listed in the article are directed towards children with autism (chelation, hyperbaric chambers, etc). Perhaps this population is a good starting point from which to address this pressing concern, as most information-providing agencies for families are proscribed from providing preferential treatment information – except in the case of conclusive demonstration of harmful effects. To my knowledge, no treatment has been sufficiently widely accepted to meet this criterion.

-Most notably, there is a logical flaw in Lilienfeld’s paper when he claims, “the literature here suggests that it would be erroneous to presume that these [untested] treatments are safe prior to subjecting them to adequate tests.” (p. 62; italics mine) In contrast to the Dodo Verdict, Lilienfeld is proposing an “inverse treatment specificity” effect. As a result, he proclaims that we should indeed be wary of untested treatments, as they may, in fact, prove more harm than good. However, he may be committing a sort of Type I error, here, by too hastily rejecting his null hypothesis of treatment equivalence. Given this formulation, though, it seems we are stuck with trying to impose a 3-factor logic on a binary set of choices: though in principle we would like to presume a priori that a treatment is either efficacious, negligible, or iatrogenic, in practice one must either offer a treatment option to a patient or not.

As such, from a practical standpoint, it may behoove us to no longer view the large body of untested treatments as either “probably ok” or “maybe harmful,” but to use the existing body of research to make solid evidence for use in clinical practice prior to subjecting them to rigorous research. To wit: we have a body of literature speaking to “common factors” (such as positive treatment alliance) for efficacious treatments. It seems, then, that Lilienfeld’s list of potential “underlying principles of negative change, such as premature termination of exposure, vicarious exposure to negative role models, and induction of false traumatic memories, that cut across numerous specific techniques,” (p. 65) could be seen as “common factors” of PHTs (indeed, Lilienfeld calls them “mediators”). Of course, research would be needed to ensure that these are appropriately-defined constructs, but it seems that, as Lilienfeld notes, undertaking this line of research would be “more informative and parsimonious than a catalogue of PHTs” (p. 65). This, then, brings me to my point: using both sets of positive and negative “common factors,” one could evaluate an untested treatment as being at risk for being a PHT or as having a protective factor against such risk. Treatments, for instance, that hinge on the generation of “suppressed memories,” could be seen as at risk treatments, while those that adhered centrally to principles of exposure for phobia treatment could be seen as somewhat protected against the potential likelihood of being iatrogenic. This rather easy and straightforward process of evaluating the prima facie characteristics of treatments could provide the useful direction needed for those researchers hoping to heed the urgent call of this paper to begin to determine a more comprehensive list of PHTs.