Re: differential mechanism

Ulrich: Thank you for your continued participation in this dialogue.

(from my earlier post) The reason we assert that the onus is on demonstrating the unique elements of EMDR contribute to treatment outcome is that the opposite position (that we skeptics have to prove the unique elements don't contribute to outcome) is logically impossible: you cannot prove the null hypothesis. The logic of scientific investigation is to assume that two things are the same, and then determine if there is adequate evidence to reject this null hypothesis and thereby indirectly advance the alternative position that things are not the same.

(from your reply) I totally concur with that statement. Accordingly, I have some reseravations about studies that appear to have set out to prove the null hypothesis, or at least report null findings without actually using appropriate sample sizes.

However, IMO the same holds true for studies setting out to show that EMDR and exposure rely on the same effect.

(my additional thoughts) I am glad that we agree on this logical and methodological point. I agree that studies should have adequate power to test the hypotheses of interest. I agree this holds true regardless of whether you are a believer, an agnostic, or an atheist. That said, the current dismantling studies are what they are. My complaint about some proponents of EMDR is that they feel it is acceptable to simply point out the issue of low statistical power and then ASSUME that the result would have been significant, had the study used enough subjects. The same goes when deviations in the procedure are noted: Deviations from the standard protocol are noted and then it is assumed the results would have different, had the researchers just done it right. Null results in a low power study, or null results following inadequate procedures cannot be reinterpreted as evidence for EMDR.

IMO, I think it is also important to view some of these issues in historical context. At the time several of the early dismantling studies were planned and conducted (e.g., Boudewyns et al., 1993; Renfrey & Spates, 1994; Foley & Spates, 1995), eye movements and other laterally alternating stimuli were still very much considered to be very important components of EMDR. The theory and clinical lore around EMDR would have readily predicted that removing these components should have a large detrimental effect. Large effects do not need as much power to detect as small effects. Thus, smaller n studies certainly seemed reasonable at the time.

To better illustrate the point, let me adopt and expand a bit on an analogy that Sandra has used previously, that of EMDR being a power tool. To be a bit more concrete, let's say that EMDR is a power saw. In the early years, eye movements and other bilateral stimuli were so integral to EMDR practice and theory that removing the eye movements seems to me a lot like unplugging the saw (thus rendering it a hand saw). The comparison between a the effectiveness of a power saw that is plugged in compared to one that is unplugged should be so obvious as not require too many observations to convince you that plugging it in is superior way to cut wood. The results of dismantling studies have failed to meet this expectation. Of course, we now know better and part of the reason we know better is the results from the earlier studies. We must turn elsewhere for hypotheses about which procedural elements are important in EMDR.

Shapiro (1999) provided some help by suggesting several things that may be important (e.g., repeatedly creating and discarding trauma-related imagery resulting in a sense of mastery; synchronizing thoughts, feelings, and sensations about the trauma; cognitive separation). However, she also made it clear she was not willing to stick out her neck and make any strong predictions: "...there are no assumptions of primacy or conjectures regarding the relative weighting of the various factors" (p. 45). This is unfortunate because it makes the task of dismantling EMDR all the more difficult, as there is no longer a clear suspect for the active ingredient. The exception, on which I hope we can all agree, is the focus on the trauma. If we included all elements of EMDR and simply replaced the content with non-trauma material, it would significantly lose its effectiveness in treating PTSD. However, if non-trauma focused EMDR reduced PTSD as much as trauma-focused EMDR, we could absolutely rule out exposure as an explanation.

Regarding the comment about "studies setting out to show that EMDR and exposure rely on the same effect," I try personally not to second-guess the motives of the investigators. If I were forced to guess, however, I certainly don't think that Leon Hyer and Richard Spates, two of the individuals who have been involved in dismantling studies of EMDR that failed to find evidence for the efficacy of eye movements, were out to prove that EMDR was just exposure. Hyer certainly doesn't seem to think so (Hyer & Brandsma, 1997: EMDR minus eye movements equals good psychotherapy, Journal of Traumatic Stress, 10, 515-522), and Spates appears to be coming to this conclusion as a result of his data, rather than as the result of an a priori commitment. Regardless of my guesses about peoples' motives, if some researchers have indeed set out to "prove" that EMDR and exposure are the same, they have chosen a loosing battle because one cannot prove that two things are the same.

A better strategy for us skeptics is to simply keep pointing out the fact that the burden of proof rests with those making the positive claim and insist that proponents of EMDR "show us the data." The other side of that coin is that proponents of EMDR should be willing to accept their burden of proof and show us the data. If they cannot show us the data, then they should be quiet. If they can show us the data, and the data survive critical scrutiny, then we critics should be quiet.

(from your reply) The suggestion to independantly vary exposure time and thus violating the standard EMDR protocol is a research approach that in some ways is caught within the behavioural paradigm. Bob Stickgold's posting may allow for a possible resolution of this problem.

(my further thoughts) I too thought Bobs comments were excellent and, if you haven't already done so, I hope you'll look over my reply to him. But to respond directly to you a bit further, I (again) am not as pessimistic as you are about the possibility of showing a dose-response curve in EMDR. I'm open to the possibility that altering the standard EMDR procedure could result in better results, worse results, or no difference. It seems you are committed to the idea that altering the standard procedure can only reduce the efficacy of EMDR. We have already seen this was not the case with eye movements.

(from your reply) IMO, part of the reason why EMDR has generated such controversy, apart from the admittedly hokey looking eye movements, is that it is a treatment that is not based on any existing psychological theory.

(my further thoughts) I agree this is part of the reason. But I think EMDR (and the promotion of EMDR) has been rightly criticized on other grounds. EMDR has been charged with being a moving target: changes in the practice of EMDR occur based on "clinical experience" more rapidly than research can be completed to determine how effective the previous "standard procedure" was. And if the old practice was found lacking, the result was dismissed as not relevant to the new procedures. The history of eye movements in EMDR is a perfect example of this. Shapiro (1989) was quite clear about the central role of eye movements in the procedure: "...the crucial component of the EMD procedure is the repeated eye-movements while the memory is maintained in awareness" (p. 220).

This is a strong position that is easily tested by removing the eye movements and replacing it with nothing or with some other form of alternating stimuli (to control for expectancy effects) and the prediction is quite clear: the eye movement condition should be superior to no eye movements, which should not differ from a waitlist control group. However, by the time enough such studies could be completed, the EMDR procedure had been changed to include these other forms of bilateral stimulation and even fixed eye variations. Thus failures to find differences between eyes fixed and other conditions were dismissed as understandable because the authors were simply comparing variations of EMDR (Lipke, 1993 cited in Greenwald, 1996).

Shapiro has also been criticized for her aggressive marketing practices prior to adequate data to support EMDR's effectiveness. If we can all agree that the first convincing data for EMDR's basic efficacy came from the S. Wilson et al. study (1995), which were then independently replicated by Rothbaum (1997), then perhaps there is reason that critics were concerned by claims that over 4000 therapists had been trained in EMDR by the summer of 1993 (Shapiro, 1993).

(from your reply) Shapiro (1995) essentially uses computer science and cognitive neuroscience metaphors to explain the effects of EMDR. EMDR has been caught in a recent paradigm shift within psychology (also see Journal of Clinical Psychology, Volume 57, Issue 9, 2001: Special Issue: In Search of a Paradigm for Clinical Psychology: Neuroscience versus Behaviorism).

EMDR, in some ways has been a treatment in search of a theory. Using some recent findings within the field of neuroscience may aid in generating useful research hypotheses about a possible underlying mechanism with regard to EMDR as well as other treatments, thus potentially leading to an improved understanding of human behaviour.

(my further thoughts) From your perspective, EMDR has been a treatment in search of a theory. I have a different perspective. I think EMDR has suffered from too much theory in the absence of a clear set of findings to specify what needs explaining. IMO, the facts that need explaining are:

1. Compared to waitlist control, EMDR works (S. Wilson et al., 1995; Rothbaum, 1997)

2. Compared to active listening, EMDR works (Scheck et al., 1998). This works toward ruling out the hypothesis that EMDR effects are the result of "common factors" or "non-specific factors" of therapy.

3. It is not clear what make of the findings that EMDR was slightly better (Edmond et al., 1999) or clearly better (Marcus, Marquis, & Sakai, 1997) than "treatment as usual" given the mixture of treatments involved in the comparison condition. I am especially concerned about the quality of care that patients may have received through Kaiser Permanente in the Marcus et al. study, given that Sakai has recently reported a study in which HMO therapists had been using Thought-Field-Therapy. (The current issue of Journal of Clinical Psychology is a special issue devoted to T-F-T, which is where Sakai's paper appears. The target papers were published without having to survive peer review. Instead, they were published alongside critical papers).

4. EMDR works as good (Vaughan et al., 199?) or slightly better than relaxation (Carlson et al., 1998). The finding by Carlson et al. that EMDR was superior to relaxation was muddied by the fact that EMDR sessions were 60-75 minutes each and relaxation sessions were only 40 minutes each. Thus EMDR participants had the benefit of somewhere between 240 - 420 additional minutes of therapy.

5. There is no convincing evidence that eye movements or any other bilateral stimuli contribute much to treatment outcome.

6. There is no convincing evidence that installation trials improve outcome.

7. There is no evidence that any other unique aspect of EMDR improves outcome.

8. The question about relative efficacy of EMDR and exposure therapy or other forms of CBT that have independent evidence of efficacy has not yet been answered. Several studies have been completed recently and their results should help to answer this question. However, I suspect there will be many arguments from both camps about treatment fidelity. Until the dust settles, claims about the unique benefits of EMDR (e.g., its supposed speed relative to traditional CBT) are mere speculation.

If these indeed are the facts currently to be explained, there is nothing that cannot be accounted for in terms of current theories. Before EMDR needs a theory, it needs some findings that exceed the boundaries of our current theories. EMDR is a procedure in need of a finding to justify its wild theorizing.

Replies:

Re: differential mechanism, by Göran Högberg Sweden, 11/23/01

| Behavior OnLine Home Page | Disclaimer |

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