Different methodological positions

I think that part of what's going on here are a clash of two different methodological positions on the anwer to the question "Does EMDR work?"

One position is what I will call the "weak position" (only to distinguish it from the "strong" position, not derogatory is meant by the term). Sandra, Ricky, and for most part myself apply adopt this position when we conclude that there adequate evidence showing that EMDR "works" (e.g., Rothbaum et al., 1997; S. Wilson, et al., 1995; Power et al., 2003).

From the weak position, all that is required is evidence from adequately conducted studies comparing EMDR with some nominal control group, such as waitlist, relaxation, suppportive counseling, or whatnot. In other words,the package known as EMDR (or exposure therapy, or cognitive processing therapy, or stress inoculation training) is better than minimal or no-treatment. This is the situation we are currently in with regard to treatment of PTSD: We have several packages that "work" from this weak methodological position. Nominally, we have multiple treatments that work.

The second position is what I will call the "strong position." This position would seem to be endorsed by Landry and is explicitly endorsed in the methodological paper by Lohr et al. published in the 1999 special issue of Journal of Anxiety Disorders. From this perspective, you have to show that something is unique to a treatment before you can claim that the treatment works.

Consider the following hypothetical situation. Assume that we have two treatment packeages, TX1 and TX2. TX1 is comprised of a whole bunch of component A and and some of B. We can represent this as AAB.

Package TX2 is compristed of some A and some C, represented as AC.

Each package is shown superior to waitlist, but TX1 is better than TX2. Importantly, we identify these treatments or distinguish one from the other based on their unique features. Now, the question here is whether we have one treatment, two treatments, or three treatments.

If it turns out that the only effective ingredient is component A, then even though we have two treatment packages and three ingredients, we really only have one treatment, which is A. Because there is no evidence for an incremental effect of B or C, we cannot properly say, from this perspective, that TX1 or TX2 work, because what is really working is A.

Let me get silly for a bit. Let's say that after being throughly trained in CBT for PTSD, I then develop my own treatment that consists of CBT but I wear a purple hat when I do CBT. I then conduct an RCT and show that "Purple Hat Therapy" is better than waitlist or even relaxation. Can I claim that "Purple Hat Therapy" works?

From the weak methodological perspective, the answer is yes. I have demonstarted that my package is bettter than control.

From the strong methodological perspective, the anwer is no, I can't claim that Purple Hat Therapy works. Before I can make this claim, I have to show that there is something about the addition of the purple hat that makes treatment better than plain old CBT. If the purple hat makes no difference, then I've simply added another study showing the CBT works.

From the weak methodological perspective, I think we can safely say that EMDR, prolonged exposure, stress inoculation training, and cognitive therapy all "work."

From the strong methdological perspective, we have to recognize that these treatments have overlapping features, although they may vary in the specific amounts of those features. For example, PE, EMDR, and cognitive therapy all involve the patient thinking about the trauma to some extent, but the extent to which the patient focuses on the trauma may vary across treatments, with PE probably spending the most time focused on the "index" trauma. Because of this overlap, we cannot claim that the specific treatment works until we can show that there is some unique feature that works above and beyond the common features. So, from the strong position, we would have to start with the assumption that PE, EMDR and CT really represent a single therapy: thinking about the trauma therapy. And we can only claim that EMDR works if we can show that some aspect of EMDR adds to outcome above and beyond "thinking about the trauma." If we can't show a unique effect, then EMDR really boils down to "thinking about the trauma" therapy.

If I'm correct in my analysis, then perhaps we can save a lot of unproductive exchanges by recognizing in which sense the term "works" is being used. If Landry or somoene else says "there is no evidence that EMDR works" and is coming from the strong methodological position, then listing the studies that show EMDR is superior to WL, or Relaxation, or even PE do nothing to rebutt the position.

Similarly, if someone makes the claim the EMDR doesn't work, they should make it clear that they either are comming from the strong position, or else recognize that they have the buden of showing why such studies as those by Rothbaum, S. Wilson, Powers et al. are invalid.

Replies:

Re:Different methodological positions, by Realist, 04/24/03 Re:Different methodological positions, by Sandra Paulsen Inobe PhD, 04/25/03

| Behavior OnLine Home Page | Disclaimer |

Copyright © 1996-2004 Behavior OnLine, Inc. All rights reserved.