Eye Movement Desensitization and Reprocessing (EMDR) Treatment for Psychologically Traumatized Individuals
Sandra A. Wilson The Union Institute
Lee A. Becker University of Colorado, Colorado Springs
Robert H. Tinker Colorado Springs, Colorado
(abstract) The effects of 3 90-minute eye movement desensitization and reprocessing (EMDR) treatment sessions on traumatic memories of 80 participants were studied. Participants were randomly assigned to treatment or delayed-treatment conditions and to 1 of 5 licensed therapists trained in EMDR. Participants receiving EMDR showed decreases in presenting complaints and in anxiety and increases in positive cognition. Participants in the delayed-treatment condition showed no improvement on any of these measures across the 30 days before treatment, but after treatment participants in the delayed-treatment condition showed similar effects on all measures. The effects were maintained at 90-day follow-up.
(discussion) The present results suggest that EMDR was effective in decreasing symptoms and anxiety associated with traumatic memory and in increasing positive cognition. EMDR effectiveness was demonstrated on different outcome measures after three 90-minute treatment sessions, with the effects being maintained at 90 days after treatment.
The SUDS and VOC ratings improved within each of the three treatment sessions. These findings are consistent with those of Shapiro (1989b) who found significant improvement on these measures in a single treatment session. The present results also buttress Shapiro’s findings of significant reductions in presenting complaints and anxiety.
The supplementary analyses suggest that EMDR was comparably effective for a range of traumas and individuals. Apparently the treatment was equally effective whether the trauma was related to sexual assault or molestation, physical or mental abuse, relationship trauma, or death of a significant other. Likewise EMDR worked equally well whether the participant had previous therapy or not and for those diagnosed as PTSD versus those who did not receive that diagnosis. Outcomes also did not vary for longstanding traumas or more recent ones, for severe or less severe traumas, or for gender of therapist or participant.
A number of factors limit conclusions that can be drawn from the present study. Behavioral measures of outcome were not used. Other than PTSD, no diagnoses were made, preventing investigation into the effectiveness of the treatment with other diagnoses or the effects of comorbidity. Individuals with strong secondary gain issues were screened from the present study. Treatment integrity was not evaluated by independent observers. Nonspecific treatment effects may be represented in the present results to an unknown extent. Treatment effectiveness with multiple traumatic memories was not investigated. Therapists were more rigorously monitored than what would be expected in usual clinical practice. The present sample was not representative of minority populations.
The effectiveness of various treatments for PTSD has been reviewed by Black, Abueg, Woodward, and Keane (1993). Although research is limited, they concluded that exposure therapies have somewhat better research support than the psychodynamic approaches and documented that the typical length of treatment for exposure therapies is 10 to 14 sessions. Although EMDR therapy contains a number of elements that are not typical of exposure therapies, to some degree imaginal exposure is involved in EMDR as the participant is exposed to images, cognition, emotions, and physical sensations of the traumatic event. However, length of treatment in the present research was considerably shorter than what was reported by Blake et al. for other therapies.
Despite the brevity of treatment, the mean effect sizes in the present study range from 0.44 to 2.07, with composite effect sizes of 1.82 for trauma-specific measures and 0.65 for general measures of functioning. These effect sizes compare favorably with well-controlled studies on psychotherapy as reported by Lipsey and Wilson (1993). The clinical significance analysis, using normative comparisons, indicated that EMDR brought the participants to within a normal range on the outcome measures.
Treatment effects were found to be stronger for measures that were specifically related to the trauma than for the more general measures. Trauma-specific measures (SUDS; IES Intrusion and Avoidance scales) contain questions about how the specific traumatic memory disrupts everyday functioning and how the participant avoids reminders of the specific trauma. The more general measures (State and Trait Anxiety; SCL-90-R scales) contain items that relate to general psychological functioning rather than to symptoms related to a specific traumatic event. It may be that the EMDR treatment was not of sufficient duration to effect larger changes on the more general measures. Furthermore, some participants may have had more than on traumatic memory. The more general measures of functioning might be resistant to change to the extent that additional traumatic memories were not treated. Blake et al. (1993) have noted that previous studies in PTSD typically only note improvement in the more obvious PTSD symptoms, such as intrusions and avoidance, and that evidence of improvement on more general measures of psychological functioning has been lacking. The finding that EMDR treatment had any effect on the more generalized measures raises the possibility that, as the traumatic memory was desensitized, the general functioning of the participant improved, with the participant becoming less anxious and depressed, experiencing fewer somatic complaints, and improving in self-esteem.
Demand characteristics were minimized in the present study by employing an independent assessor to collect pre- and post- treatment measurements. Therapist-collected SUDS ratings have been criticized as being overly sensitive to therapist demand (Acierno et al., 1994; Herbert & Meuser, 1992; Lohr et al., 1992). Because the termination of the assessment session was not contingent on the SUDS ratings taken by the independent assessor, and because the participant’s personal relationship with the assessor was minimal, it is likely that demand characteristics associated with those SUDS ratings were lessened. Even with those controls for demand characteristics in place, it might be argued that nonspecific, placebo treatment effects played a role in the present outcomes. However, PTSD has been noted to be resistant to placebo effects (Solomon, Gerrity, & Muff, 1992). In addition, when Lipsey and Wilson (1993) compared meta-analytic studies using placebo controls with those without placebo controls, they found that the average placebo effect size was .19. This suggests that, although placebo effects could account for some part of the effect sizes in the present study, it is not likely that they could fully account for the present results, where the composite effect size for the trauma-specific measures was over nine times greater. The present results suggest that EMDR shows promise in the treatment of traumatic memories, although the reasons for its effectiveness are yet to be understood.