by Michael D. Anestis, M.S.
Today's article adopts a similar theme to Monday's article on dolphin-assisted therapy (DAT): just because a treatment is popular and widely used does not mean that it actually works as advertised. Today's focus is on eye movement desensitization and reprocessing, created by Frances Shapiro (1989) and typically known by its acronym, EMDR. The approach was originally introduced as a novel treatment for post-traumatic stress disorder (PTSD), although it has since been adopted as a treatment for a variety of other mental illnesses, particularly other anxiety disorders. The effectiveness of this treatment approach has been the subject of rather intense controversy, with proponents making sweeping claims about the robust impact of EMDR and skeptics pointing to data that simply offers no support for those claims. Before detailing that data, let me first provide a quick description of the therapy itself, which is described well in an article by Cahill, Carrigan, and Frueh (1999), one of a series of articles in an issue of the Journal of Anxiety Disorders that was devoted entirely to this particular topic.
What is EMDR?
EMDR, as used to treat PTSD, requires that a client engage in a series of lateral eye movements while describing and/or imagining details of a traumatic event. These eye movements are typically generated by following the therapist's finger as he or she moves it back and forth across the client's visual field. After a set of 10-20 eye movements, the therapist assesses the client's current level of distress by asking them to rate their subjective units of distress (SUDS; Wolpe, 1973) from 0 to 10. This procedure is repeated until the client's SUDS is reported to be 0 or 1.
Following this phase - desensitization - the client is told to come up with a positive thought to keep in mind in conjunction with the traumatic images and subsequent physiological sensations. The same procedure described for the desensitization phase is then repeated until the client reports a high level of belief in the new, positive interpretation of the images and physiological sensations.
In summary, then, EMDR thus exposes clients to thoughts and descriptions of a traumatic event until they no longer experience intense anxiety and then asks them to develop alternative interpretations of those images and descriptions. For those of you who have read our description of cognitive processing therapy and prolonged exposure therapy, this might seem familiar. I will return to this point later, but this is, in fact, one of the central points of criticism of EMDR: that it is simply an alternative version of already validated treatments with the potentially meaningless addition of eye movements.
Following this phase - desensitization - the client is told to come up with a positive thought to keep in mind in conjunction with the traumatic images and subsequent physiological sensations. The same procedure described for the desensitization phase is then repeated until the client reports a high level of belief in the new, positive interpretation of the images and physiological sensations.
In summary, then, EMDR thus exposes clients to thoughts and descriptions of a traumatic event until they no longer experience intense anxiety and then asks them to develop alternative interpretations of those images and descriptions. For those of you who have read our description of cognitive processing therapy and prolonged exposure therapy, this might seem familiar. I will return to this point later, but this is, in fact, one of the central points of criticism of EMDR: that it is simply an alternative version of already validated treatments with the potentially meaningless addition of eye movements.
Does EMDR Work?
Asking whether or not EMDR works is overly simplistic. A better approach, and one taken in several reviews of this topic (see Cahill et al., 1999 and Lohr, Lilienfeld, Tolin, & Herbert, 1999 for good examples), is to compare the impact of EMDR to a variety of other approaches and to examine whether the unique aspect of EMDR - namely, repeated lateral eye movements - adds anything to the impact of the treatment.
As noted in Cahill et al (1999), a number of studies have compared EMDR to no treatment waitlist conditions. Comparisons to a no treatment condition are valuable for a couple of reasons. First, they provide data indicating that a treatment's effects are not due simply to the passage of time. If it was just time that accounted for the improvement, the no treatment condition would equal the treatment condition at the end of the study. Second, they ensure that regression to the mean - the tendency for extreme scores to drift toward the mean across time - do not account for decreases in symptoms.
In studies comparing EMDR to no treatment conditions, EMDR has consistently been shown to be superior, indicating that, in the treatment of PTSD (as well as multiple other anxiety disorders), EMDR is better than nothing (Boudelwyns & Hyer, 1996; Carlson et al., 1998). This, obviously, does not mean a whole lot, but it does at least provide evidence that EMDR has positive effects and is definitively not iatrogenic.
A more stringent way to test the strength of a treatment approach is to compare it to other unvalidated treatments. Unvalidated treatments are treatments for which there is no data indicating that the particular approach is an effective intervention for a specific diagnostic presentation. Studies comparing EMDR to unvalidated treatments for PTSD have been conducted on numerous occasions. The results appear to indicate that EMDR is at least equally effective to several unvalidated treatments for PTSD, including relaxation (Carlson et al., 1998; Vaughn et al., 1994). These results indicate that, relative to other treatments lacking empirical support for their efficacy and effectiveness, EMDR is at least equally useful. Again, this type of comparison can be useful, but it essentially boils down to comparing a treatment approach to something less than the best available treatment. When a client is learning about his or her options, she likely does not concern herself with how a particular treatment stacks up against treatments not recommended for everyday practice, but rather, the relative utility of a treatment compared to other treatments already shown to be effective.
A third and significantly more important way to test the strength of a treatment approach is thus to compare it to empirically supported treatments (ESTs). There are multiple reasons for doing so, the most important of which is determining the extent to which the treatment exhibits incremental validity, value above and beyond that provided by already established approaches, relative to additional costs. Despite numerous claims to the contrary by proponents of EMDR during the 1990's, that particular era produced no studies comparing EMDR to ESTs and, as such, claims of EMDR's superiority were based upon a complete lack of evidence. In an effort to address this empirical void, Devilly and Spence (1999) published such a study in the same issue of the Journal of Anxiety Disorders mentioned earlier in this article. In this particular study, Devilly and Spence (1999) compared EMDR to a form of cognitive behavioral therapy (CBT) in the treatment of PTSD. They found that clients in the CBT condition exhibited both statistically and clinically significantly greater improvement in symptoms after treatment and that the superior impact of CBT was not only maintained at 3-month follow-up, but actually increased. This, quite obviously, presents a problem for proponents of EMDR. Although the treatment performs better than no treatment at all and likely better than other unvalidated treatments, when compared to an EST, it fails to produce the desired effect.
Studies examining the utility of EMDR have typically only found positive effects on self-reported measures of distress such as SUDS or on therapist rated measures of improvement. Such measures can be useful at times, but are problematic here for a variety of reasons. First, if the self-reported distress measure does not match up with results on validated measures of symptoms, the results can be called into question. Additionally, in the studies that used clinician ratings, the clinicians were not blind to the treatment condition. In other words, the researcher writing about EMDR, who was trained in EMDR, knew that the client received EMDR and was motivated to report that the client improved, as this would validate the researcher's work. This is not to say that the researchers were lying, but rather that all people are subject to their own biases and removing those biases from measurement is important in order to ensure that results are legitimate.
As noted in Cahill et al (1999), a number of studies have compared EMDR to no treatment waitlist conditions. Comparisons to a no treatment condition are valuable for a couple of reasons. First, they provide data indicating that a treatment's effects are not due simply to the passage of time. If it was just time that accounted for the improvement, the no treatment condition would equal the treatment condition at the end of the study. Second, they ensure that regression to the mean - the tendency for extreme scores to drift toward the mean across time - do not account for decreases in symptoms.
In studies comparing EMDR to no treatment conditions, EMDR has consistently been shown to be superior, indicating that, in the treatment of PTSD (as well as multiple other anxiety disorders), EMDR is better than nothing (Boudelwyns & Hyer, 1996; Carlson et al., 1998). This, obviously, does not mean a whole lot, but it does at least provide evidence that EMDR has positive effects and is definitively not iatrogenic.
A third and significantly more important way to test the strength of a treatment approach is thus to compare it to empirically supported treatments (ESTs). There are multiple reasons for doing so, the most important of which is determining the extent to which the treatment exhibits incremental validity, value above and beyond that provided by already established approaches, relative to additional costs. Despite numerous claims to the contrary by proponents of EMDR during the 1990's, that particular era produced no studies comparing EMDR to ESTs and, as such, claims of EMDR's superiority were based upon a complete lack of evidence. In an effort to address this empirical void, Devilly and Spence (1999) published such a study in the same issue of the Journal of Anxiety Disorders mentioned earlier in this article. In this particular study, Devilly and Spence (1999) compared EMDR to a form of cognitive behavioral therapy (CBT) in the treatment of PTSD. They found that clients in the CBT condition exhibited both statistically and clinically significantly greater improvement in symptoms after treatment and that the superior impact of CBT was not only maintained at 3-month follow-up, but actually increased. This, quite obviously, presents a problem for proponents of EMDR. Although the treatment performs better than no treatment at all and likely better than other unvalidated treatments, when compared to an EST, it fails to produce the desired effect.
Studies examining the utility of EMDR have typically only found positive effects on self-reported measures of distress such as SUDS or on therapist rated measures of improvement. Such measures can be useful at times, but are problematic here for a variety of reasons. First, if the self-reported distress measure does not match up with results on validated measures of symptoms, the results can be called into question. Additionally, in the studies that used clinician ratings, the clinicians were not blind to the treatment condition. In other words, the researcher writing about EMDR, who was trained in EMDR, knew that the client received EMDR and was motivated to report that the client improved, as this would validate the researcher's work. This is not to say that the researchers were lying, but rather that all people are subject to their own biases and removing those biases from measurement is important in order to ensure that results are legitimate.
What about the eye movements? Are they useful?
Thus far, empirical evidence does not support the contention of EMDR proponents that lateral eye movements add value to treatment. In numerous studies comparing EMDR to treatments that did not include lateral eye movements in the treatment of PTSD, no differences were found between the groups on the degree to which their symptoms improved (Boudewyns & Hyer, 1996; Dunn, Schwartz, Hatfield, & Wiegele, 1996; Renfrey & Spates, 1994). Additionally, in a study that assessed whether adding lateral eye movements to exposure therapy in the treatment of a public speaking phobia, Carrigan and Levins (1999) found that eye movements did not positively impact treatment in any way. In fact, the only difference between individuals who received eye movement training and those who did not was that those who did were less likely to be willing to give a speech post-treatment. Not a positive result for EMDR.
So what does this mean? It means that EMDR can be useful, but not because it is EMDR. It can be useful because it utilizes several components of validated treatments, particularly CBT. The addition of the eye movements makes the treatment seem novel and exciting, which likely explains a large portion of its popularity, but they are not actually active ingredients in the treatment of mental illness. Here, again, we have an example of misguided priorities. The philosophical appeal of eye movement training for some clinicians appears to distort their understanding of the purpose of therapy. Therapy is designed to help struggling clients to reclaim their lives. In order to do this, clinicians must offer the best available treatment, not the one that seems the most interesting, particularly if the most interesting treatment is not supported by data.
The reviews cited in this article are 10 years old...what is the story today?
There have been many studies on EMDR since the 1999 special issue devoted to the topic in the Journal of Anxiety Disorders. A summary of these studies would be well beyond the scope of today's post. I will make an effort to wade through all of these studies in the future and to provide readers with an update on the empirical status of EMDR. In the meantime, however, I suggest reading the summary of the current empirical status of EMDR on the website of the Society of Clinical Psychology, American Psychological Association Division 12, the organization responsible for updating the list of currently empirically supported treatments for mental illness.
Division 12 states that the current empirical evidence remains consistent with the conclusions mentioned above and that, in the previous ten years, no new evidence has emerged to counter the data indicating that, while EMDR is better than nothing, eye movements add little or nothing to treatment and EMDR is not as effective as EST's in the treatment of mental illness.
If you would like to learn more about empirically supported treatments for PTSD, we recommend the following products, all of which are available through our online store:
Division 12 states that the current empirical evidence remains consistent with the conclusions mentioned above and that, in the previous ten years, no new evidence has emerged to counter the data indicating that, while EMDR is better than nothing, eye movements add little or nothing to treatment and EMDR is not as effective as EST's in the treatment of mental illness.
If you would like to learn more about empirically supported treatments for PTSD, we recommend the following products, all of which are available through our online store:
Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University
