Content uploaded by Nasrin Jaberghaderi
Author content
All content in this area was uploaded by Nasrin Jaberghaderi on Jun 30, 2020
Content may be subject to copyright.
Clinical Psychology and Psychotherapy
Clin. Psychol. Psychother. 11, 358–368 (2004)
Copyright © 2004 John Wiley & Sons, Ltd.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/cpp.395
A Comparison of CBT
and EMDR for
Sexually-abused
Iranian Girls
Nasrin Jaberghaderi,1Ricky Greenwald,*,2
Allen Rubin,3Shahin Oliaee Zand1 and
Shiva Dolatabadi1
1Allame Tabatabaee University, Tehran, Iran
2Child Trauma Institute, Greenfield, MA, USA
3University of Texas at Austin, USA
Fourteen randomly assigned Iranian girls ages 12–13 years who had
been sexually abused received up to 12 sessions of CBT or EMDR
treatment. Assessment of post-traumatic stress symptoms and
problem behaviours was completed at pre-treatment and 2 weeks
post-treatment. Both treatments showed large effect sizes on the post-
traumatic symptom outcomes, and a medium effect size on the behav-
iour outcome, all statistically significant. A non-significant trend on
self-reported post-traumatic stress symptoms favoured EMDR over
CBT. Treatment efficiency was calculated by dividing change scores
by number of sessions; EMDR was significantly more efficient, with
large effect sizes on each outcome. Limitations include small N, single
therapist for each treatment condition, no independent verification of
treatment fidelity, and no long-term follow-up. These findings
suggest that both CBT and EMDR can help girls to recover from the
effects of sexual abuse, and that structured trauma treatments can be
applied to children in Iran. Copyright © 2004 John Wiley & Sons, Ltd.
*Correspondence to: Ricky Greenwald PsyD, P.O. Box 544,
Child Trauma Institute, Greenfield, MA, 01302, USA.
E-mail: rg@childtrauma.com
Practitioner
Report
INTRODUCTION
Sexual abuse victimization can lead to a range of
consequences in children, potentially including
post-traumatic stress disorder (PTSD), develop-
mental delays, increased anxiety and depressive
symptoms, and disruptive, sexually inappropriate,
and regressive behaviours (Kendall-Tackett,
Williams & Finkelhor, 1993). There are relatively
few controlled studies of the efficacy of treatments
for the sequelae of sexual abuse (Saywitz,
Mannarino, Berliner, & Cohen, 2000) and other
childhood trauma (Cohen, Berliner, & March,
2000). There is currently greatest support for the
efficacy of various forms of cognitive-behavioural
therapy (CBT), typically including some combina-
tion of psycho-education, coping skills training
and exposure. Intervention with parents has also
been shown to be helpful, but is beyond the scope
of this review.
For example, Deblinger, Lippman and Steer
(1996) compared a 12-week programme of abuse-
focused CBT provided to children only, to parents
only, or to both children and parents, with
CBT versus EMDR 359
Copyright © 2004 John Wiley & Sons, Ltd. Clin. Psychol. Psychother. 11, 358–368 (2004)
standard community care for a total of 100 sexu-
ally-abused children ages 7 to 13 years. Results
indicated that all groups improved on PTSD symp-
toms; however, CBT provided directly to the chil-
dren resulted in significantly greater improvement.
In another study of 49 children ages 7 to 14
years, those who were provided with abuse-
focused CBT experienced significantly greater
improvement in depression and social competence
than those receiving nondirective supportive
therapy (Cohen & Mannarino, 1998). However,
the groups did not differ in PTSD symptoms,
possibly because the trauma was a focus in the con-
trol condition as well.
A newer treatment, eye movement desensitiza-
tion and reprocessing (EMDR), has been shown to
be efficacious in a number of adult trauma studies,
and preliminary findings with traumatized
children are also promising. EMDR is a treatment
for traumatic memories and their sequelae that
requires the client to perform bilateral eye move-
ments while concentrating on the trauma memory.
The procedure consists of a structured sequence of
treatment components (Shapiro, 1995) that have
been identified as being effective across trauma
treatment modalities (Hyer & Brandsma, 1997;
Sweet, 1995), including psycho-education, cop-
ing skills training, and exposure (including
emotional processing and cognitive restructuring
components).
A recent review of randomized treatment
studies, published as part of the International
Society for Traumatic Stress Studies’ Treatment
Guidelines, found EMDR to be an empirically sup-
ported efficacious trauma treatment (Chemtob,
Tolin, van der Kolk, & Pitman, 2000), with the
reservation that EMDR had not yet been directly
compared to other validated focused PTSD treat-
ments such as prolonged exposure. Since the pub-
lication of that review, several studies meeting
most of Foa and Meadows’ (1997) gold standard
criteria have directly compared EMDR to validated
CBT treatments (including an exposure compo-
nent) for PTSD. Findings across studies indicate
that both treatments were, in general, comparably
efficacious (Ironson, Freund, Strauss, & Williams,
2002; Lee, Gavriel, Drummond, Richards, &
Greewald, 2002; Power, McGoldrick, & Brown,
2002; Taylor et al., 2003; A. McFarlane, presentation
at the Annual Meeting of the International Society
of Traumatic Stress Studies, San Antonio, Novem-
ber 2000; B. O. Rothbaum, presentation at the
Annual Meeting of the Association for the
Advancement of Behavior Therapy, Philadelphia,
November 2001). Three studies (Ironson et al.,
2002; Power et al., 2002; A. McFarlane, presentation
at the Annual Meeting of the International Society
of Traumatic Stress Studies, San Antonio, Novem-
ber 2000) suggested that the therapeutic effect may
have occurred more quickly for EDMR than for
CBT. One study (Ironson et al., 2002) found that the
level of distress both during and between sessions
was lower for EMDR; this study also a found lower
drop-out rate in the EMDR group.
These findings suggest that EMDR may be at
least equal to other CBT approaches in efficacy and
acceptability while being more efficient, in that
with EMDR much less homework is required, and
the treatment effect may be achieved in fewer ses-
sions. However, it would be premature to draw
this conclusion. There are many CBT approaches,
not all of which were compared; furthermore, such
findings need to be replicated and elaborated. Also,
one study found that EMDR fared much worse
than CBT in terms of drop-out rates as well as out-
comes for treatment completers (Devilly & Spence,
1999). Unfortunately, significant methodological
problems make this study difficult to interpret (see
Chemtob et al., 2000; Lee et al., 2002).
There is no published controlled research on
EMDR for sexually-abused children. EMDR has
been used successfully with women who were sex-
ually abused as children (Edmond, Rubin, &
Wambach, 1999). There are also several wait-list
studies that suggest EMDR’s efficacy with trauma-
tized children and adolescents (variety of types of
trauma events; Puffer, Greenwald, & Elrod, 1998;
Soberman, Greenwald, & Rule, 2002). One wait-list
study following a disastrous hurricane only treated
children with EMDR who had failed to respond to
a generally effective prior cognitive-behavioural
group treatment (Chemtob, Nakashima, Hamada,
& Carlson, 2002). The only direct treatment com-
parison studies of EMDR for children or adoles-
cents did yield results favourable for EMDR. One
study compared EMDR to Active Listening for
adolescent and young adult females (Scheck,
Schaeffer, & Gillette, 1998) and another compared
EMDR to Systematic Desensitization for Iranian
college students (Dolatabadi, 2000). EMDR has not
yet been tested with children or adolescents
against a credible alternative trauma treatment.
Our primary goal in the present study was to
directly compare the efficacy of CBT and EMDR in
the treatment of sexually-abused girls. We were
also interested in the relative acceptability and
efficiency of these treatments. Finally, we wanted
to see how these treatments would perform with
360 N. Jaberghaderi et al.
Copyright © 2004 John Wiley & Sons, Ltd. Clin. Psychol. Psychother. 11, 358–368 (2004)
Iranian children as compared with the mostly first
world English-speaking samples reported in the
literature.
METHOD
Design
Iranian girls ages 12–13 years with a reported
history of sexual abuse victimization were
randomized to two active trauma treatment
conditions, CBT and EMDR. Assessment of
post-traumatic symptoms and problem behaviours
were completed pre- and post-treatment. Because
of the wide range of symptoms experienced by
sexually-abused children (Kendall-Tackett et al.,
1993), the focus was not PTSD but a wider range
of symptoms and behaviours.
Measures
Child Report of Post-traumatic Symptoms
(CROPS)
The CROPS is a 26-item self-report questionnaire
covering the broad spectrum of post-traumatic
symptoms actually found in traumatized children
(Fletcher, 1993) including but not limited to those
symptoms of child/adolescent PTSD listed in the
DSM-IV (American Psychiatric Association, 1994).
The respondent is asked to endorse current
symptoms on a 0–2 scale of intensity (none, some,
or lots), total possible range of 0–52, with a score
of 19 or over indicating clinical concern. The CROPS
has shown good validity and reliability in several
settings and languages (Greenwald & Rubin,
1999; Greenwald, Rubin, Jurkovic et al., 2002;
Wiedemann & Greenwald, 2000) as well as sensitiv-
ity to change in post-traumatic status (R. Greenwald
et al., poster presented at the Annual Meeting of the
International Society for Traumatic Stress Studies,
Baltimore, November, 2002). The CROPS was trans-
lated into Persian, independently back-translated
to English, then reviewed by the original author
(RG) with item revisions as needed until the origi-
nal author was satisfied with the translation.
Parent Report of Post-traumatic Symptoms
(PROPS)
The PROPS is a 32-item companion measure to
the CROPS, also broad-spectrum and with similar
validity and reliability (Greenwald & Rubin, 1999;
R. Greenwald et al., poster presented at the Annual
Meeting of the International Society for Traumatic
Stress Studies, Baltimore, November, 2002; J.
Wiedemann & R. Greenwald poster presented at
the Annual Meeting of the International Society for
Traumatic Stress Studies, San Antonio, November,
2000). It must be filled out by an adult who is famil-
iar with the child’s behaviour; in this case, by the
child’s parent. The respondent is asked to endorse
current symptoms on a 0–2 scale of intensity, total
possible range of 0–64, with a score of 16 or over
indicating clinical concern. The PROPS was also
translated according to the procedures described
above. There is some item overlap with the CROPS
but substantial differences as well. The em-
phasis with the CROPS is thoughts and feelings,
whereas the emphasis with the PROPS is
observable behaviours.
Rutter Teacher Scale
The Rutter Teacher Scale was developed to assess
whether a child has a potential mental disturbance
(Kresanov, Tuominen, Piha, & Almqvist, 1998;
Rutter, 1967). It contains 26 statements in which
teachers rate the extent of problematic behaviours
exhibited by the child in school, such as
hyperkinetic behaviours, antisocial externalizing
behaviours, internalizing difficulties, relationship
problems, and dysfunctional habits. Scores can
range from 0 to 52, with 9 and above considered
the clinical range. The Rutter Teacher scale has
been found to have good reliability and validity in
a variety of studies across a variety of cultures,
including countries in Europe and Asia (Almqvist
et al., 1999; Kresanov et al., 1998; Kumpulainen,
Rasanen, & Henttonen, 1998; Morita, Suzuki, &
Kamoshita, 1990; Rutter, 1967). This Persian lan-
guage version omits two of the original items,
includes six additional items, has good psycho-
metric properties, and has a clinical range of 13
and above (Yousefi, 1998).
Subjective Units of Distress Scale (SUDS)
The SUDS (adapted from Wolpe as described in
Shapiro, 1995) measures intensity of subjective dis-
tress in response to a particular stimulus, such as a
traumatic memory. It is a widely used measure
which has been shown to correlate with several
physiological measures of stress (Thyer, Papsdorf,
Davis, & Vallecorsa, 1984; Wilson, Silver, Covi, &
Foster, 1996). Non-reactivity to a traumatic
memory is considered a primary indicator of
recovery (Horowitz, 1986). This 11-point scale uses
10 as the highest level of distress and 0 as the
CBT versus EMDR 361
Copyright © 2004 John Wiley & Sons, Ltd. Clin. Psychol. Psychother. 11, 358–368 (2004)
lowest level, or absence of distress. In this
study the SUDS was used as one of the termination
criteria, but not as an outcome measure.
Participants
A recruitment letter signed by the Principal Inves-
tigator (the PI; NJ) and the school principal was
sent to the homes of 125 sixth grade girls (ages
12–13 years) in an urban school in Iran. Volunteers
were asked, with their parents’ consent, to com-
plete the CROPS. Of the 123 who did, 62 received
a score of 19 or higher, indicating a clinically sig-
nificant level of post-traumatic symptoms. Then in
an office at the school, these 62 girls completed the
Lifetime Incidence of Traumatic Events (LITE;
Greenwald, 2000; R. Greenwald, A. Rubin, A. M.
Russell, & M. B. O’Connor, poster presented at the
Annual Meeting of the International Society for
Traumatic Stress Studies, Baltimore, November,
2002; also translated as above), a checklist allowing
the respondent to endorse exposure to a variety of
adverse events. Of these, 24 endorsed having been
sexually abused. These girls then participated in a
structured interview in school, conducted by the
PI, to determine the quantity and severity of the
sexual abuse. Of these, 19 girls were determined to
have experienced a qualifying abuse experience,
defined as unwanted oral, anal, genital, or breast
contact with another person, occurring 6 months
or more prior to the study. Consent/assent pro-
cedures were repeated for those eligible for
participation in the treatment study, and all
consented/assented.
All participants were in the same socio-economic
status, a de facto condition of attendance at the
school where the study took place. All participants
lived with their biological parents.
Participants were randomly assigned to treat-
ment condition, with some adjustments to promote
equivalence between groups. Four participants
had experienced penetration, and they were
blocked and then randomized, to ensure that two
went to each treatment condition. Three partici-
pants had only experienced a single abuse event;
they were also blocked; two went to the EMDR
group and one to CBT. Five of the girls’ sexual
abuse was due to incest. One of these was dis-
qualified because her brother was continuing to
molest her (despite the researcher’s efforts; chil-
dren’s protection is not mandated in Iran). The
other four were blocked to ensure that two went
to each treatment condition. The randomization
procedure for each subgroup of participants was
to pick their names out of the hat while alternat-
ing group assignments.
Ethical Considerations
It is not yet customary in Iran for universities
to have Institutional Review Boards or Ethics
Committees, so in lieu of such a structure, the
investigators discussed the design with respected
colleagues for feedback regarding scientific and
ethical concerns. The study was designed with the
usual precautions including the emphasis, at each
phase of recruitment, on the informed and volun-
tary nature of participation. The initial forms
included a way to contact the investigator to
discuss any distress engendered by participation;
and subsequent interviews were followed by
debriefing to ensure that participants were not left
in a state of distress. Consistent with other trauma
studies in which participants primarily reported
neutral or positive response (Brabin & Berah, 1995;
Martin, Perrott, Morris, & Romans, 1999; Newman,
Walker, & Gefland, 1999; Ruzek & Zatzick, 2000),
these participants indicated no special study-
related distress that required follow-up or referral,
except as already noted elsewhere in this paper.
Perhaps the most challenging ethical issue we
faced was how to balance the value of informed
consent with the sometimes conflicting value of
protection of participants. In Iran there is enor-
mous stigma associated with being a victim of
sexual abuse, and we were concerned that by
making the focus of the study explicit, we would
risk subjecting participants to ostracism by their
teachers and even by their own parents. On the
other hand, we wanted to provide as much infor-
mation as possible, so that participants and their
parents could make informed choices. Therefore
the study was presented as a trauma treatment
study, but with the nature of the trauma unspeci-
fied. Those parents judged by the PI to be suffi-
ciently open-minded and supportive were further
informed that the trauma being treated was sexual
abuse. Only five parents were so informed, three in
the EMDR group and two in the CBT group. Need-
less to say, this approach caused controversy
among the cross-cultural research team! However,
we ultimately recognized that the core value was
participant protection, and that specific procedures
designed to protect participants have to be
culturally sensitive and take local conditions into
consideration.
In general, we felt that we were successful in our
efforts to protect participants. Despite the discon-
certingly high response rate to the first wave of the
362 N. Jaberghaderi et al.
Copyright © 2004 John Wiley & Sons, Ltd. Clin. Psychol. Psychother. 11, 358–368 (2004)
study (123 out of a possible 125) – perhaps a reflec-
tion of the authoritarian nature of Iranian culture –
potential participants did seem comfortable in
declining participation as the study progressed
towards the treatment phase. Those children who
were eligible for treatment but did not participate
did agree to a follow-up evaluation which in fact
indicated no further concerns; these children were
already doing well in school and socially.
The parents of those children who dropped out of
treatment were provided an additional psycho-
educational intervention and referral options. The
parents of the child who was ineligible due to
ongoing abuse were also provided with psycho-
education and referral options. This did lead, even-
tually, to the girl moving to live with relatives,
where she was not subject to further molestation.
Procedure
The PI introduced the study and procedures to the
teachers in a group. Following the structured inter-
view in the screening procedures described above,
the pre-treatment assessment was conducted by
two psychologists who did not know the children
and were blind to assignment. For each participant,
the parent completed the PROPS, the girl com-
pleted the CROPS, and the participant’s several
teachers completed the Rutter as a group. The mea-
sures are all paper-and-pencil, and respondents
generally completed the forms with no assistance.
When assistance was needed, the psychologists
had been trained to help the respondent(s) to
understand the meaning of the item while avoid-
ing helping the respondent(s) to select their
response. Treatment was conducted after school
hours, at the nearby University clinic. Two weeks
after each participant’s final session, the post-
treatment assessment was conducted in the
same manner as described above, with the same
respondents for a given participant. Participants
requiring further treatment (those who failed to
meet termination criteria but had used up all their
sessions) were referred following the post-
treatment assessment.
Treatment Conditions
Each treatment was manualized, with reference to
published texts for more detail. In this study the
CBT procedure was based on Deblinger and Heflin
(1996), with additional activities from Camino
(2000). Although the activities were standardized,
they were tailored to the individual participants.
The CBT therapist was the fourth author (SOZ), a
professor and clinical psychologist who had been
trained in CBT by M. R. Mohamadi of Roozbeh
Hospital. The EMDR procedure was based on
Shapiro (1995), with age-appropriate modifications
suggested by Greenwald (1999). The EMDR thera-
pist was the fifth author (SD), a professor and
clinical psychologist who had been trained in
EMDR (levels 1 and 2) by G. Puk of the EMDR
Institute. Both had prior clinical experience with
children. The PI was present for all sessions, CBT
and EMDR, to observe and take notes.
Our challenge was to make the treatment condi-
tions as equivalent as possible, while also allowing
each of the treatments to be conducted (within
limits) according to its own standards. Implement-
ing a treatment approach on its own terms – the
way a practicing clinician would do it – supports
ecological validity. The trade-off is that the condi-
tions were not exactly equivalent across treat-
ments. The design reflects an attempt to balance
the interest in both internal and external validity.
In the CBT condition, the focus was on skill
development (e.g. symptom management) and on
exposure to the identified sex abuse memory. In the
EMDR condition, the skill development focus
was much more limited, and the focus on the iden-
tified trauma memory was not as strict. For
example, a minor upsetting memory was targeted
with EMDR as preparation for EMDR with the
identified sex abuse memory. The EMDR therapist
was allowed to treat other trauma memories with
EMDR if they arose spontaneously while working
with the index trauma, but was not allowed to sys-
tematically work through all trauma memories (as
would normally be done in clinical practice).
Duration of sessions was limited to 45min. In the
CBT condition, all sessions took 45min. In the
EMDR condition, most sessions were closer to
30min with some taking the full 45 min. Differ-
ences in number of minutes per session was not
systematically tracked.
In the CBT condition there was homework for
every session, such as checklists, drawings, and
listening to tapes of the exposure narrative. We
estimate that participants in the CBT group com-
pleted about 10–15h of homework in total, but
homework time was not systematically tracked.
Homework in the EMDR condition was minimal,
limited to drawing a ‘safe place’ on one occasion
between sessions early in the treatment.
Termination criteria was treatment specific, but
with a maximum of 12 sessions. CBT treatment
CBT versus EMDR 363
Copyright © 2004 John Wiley & Sons, Ltd. Clin. Psychol. Psychother. 11, 358–368 (2004)
also had a 10-session minimum in order to com-
plete certain activities; EMDR treatment had no
designated minimum. CBT treatment was termi-
nated prior to 12 sessions if the SUDS was 0–2 and
the primary abuse-related anxiety symptoms were
at a severity rating of 25 per cent or lower. EMDR
treatment was terminated prior to 12 sessions if the
SUDS was 0–2 and positive self-statements related
to the abuse were whole-heartedly endorsed, as
indicated by a 6 or 7 rating on a 7-point scale.
Parents also attended a single psycho-
educational session, the same for both groups, pro-
vided by the child’s therapist, within the first 2
weeks of treatment.
RESULTS
Pre-treatment Comparison of Groups
The groups did not differ significantly at pre-treat-
ment on age, socio-economic status, type, severity,
amount of sexual abuse, or extent of other
trauma/loss history. They also did not differ
significantly at pre-treatment on scores on any of
the outcome measures. In fact, the groups were
very similar on these pre-treatment variables.
Retention of Participants
One participant was disqualified prior to random-
ization because of ongoing abuse. Two other par-
ticipants, one from each condition, did not ever
start treatment. Both of these girls said that they
were not interested in treatment, and indeed both
of them were functioning well in school and other
areas of their lives. Two additional participants,
one from each treatment condition, dropped out
after starting treatment, because their parents
stopped bringing them, reasons unknown. The
drop-outs were not re-assessed. The other 14 par-
ticipants, seven in each condition, completed the
treatment as well as the post-treatment assessment.
Treatment Duration
In the CBT group, one participant terminated after
10 sessions, one after 11 sessions, and the rest after
the maximum of 12 sessions. In the EMDR group,
two participants terminated after four sessions,
one after five sessions, one after six sessions, and
three after eight sessions.
Post-treatment Referral
Those participants who did not meet termination
criteria at the post-treatment assessment were
determined to be in need of further treatment, and
were referred accordingly. Three participants in the
CBT group were referred for further treatment;
none in the EMDR group were referred.
Treatment Outcome
Table 1 displays the mean pretest and posttest
scores for each group. On the CROPS, compared to
pre-treatment, the EMDR group improved signifi-
cantly (p<0.05), but the CBT group did not (p=
0.116). However, the difference in improvement
between groups was not significant. On both the
PROPS and the Rutter, both groups improved sig-
nificantly (p<0.05), but the difference in improve-
ment between groups was not significant. Analyses
of covariance were used to test the significance of
between group differences, with pretest scores
serving as the covariate. Paired samples t-tests
were used to test the significance of within group
pretest to posttest differences.
Clinically Significant Change
One step in considering the clinical significance
of the pretest to posttest changes is to examine
Cohen’s d, an effect size statistic that is calculated
by dividing the pretest to posttest change by the
pretest standard deviation. When dis at about 0.8
or above, the effect size is considered large. When
it is at about 0.5, it is considered medium (Cohen,
1988). As shown in Table 2, large pretest to posttest
effect sizes were found for both groups on the
Table 1. Mean pretest and posttest scores on three
outcome measures, by group
Test Group Between
EMDR CBT groups p
Mean SD Mean SD
CROPS pretest 34.86 5.8 30.00 6.4
CROPS posttest 18.86* 7.9 22.71 6.9 0.15
PROPS pretest 21.00 6.2 22.43 10.3
PROPS posttest 10.14* 5.4 11.29* 6.6 0.96
Rutter pretest 13.71 12.2 8.86 7.7
Rutter posttest 5.00* 5.3 3.00* 2.9 0.42
* Pretest to posttest within group change statistically significant
(p<0.05).
364 N. Jaberghaderi et al.
Copyright © 2004 John Wiley & Sons, Ltd. Clin. Psychol. Psychother. 11, 358–368 (2004)
CROPS and PROPS, and medium effect sizes were
found for both groups on the Rutter.
Another way to consider clinical significance is
to see how many participants in each group moved
from clinical to normal ranges on each of the
outcomes. On the Persian version of the Rutter
Teacher Scale, a total score of 13 or more is consid-
ered in the clinical range (Yousefi, 1998). As
shown in Table 2, both of the CBT participants, and
one of the two EMDR participants, who started
in the clinical range at pretest moved to
the normal range at posttest. Although the other
EMDR participant did not reach the normal range
at posttest, she showed more pre- to posttest
improvement than all but one other participant
(also in the EMDR group) in the study. She had the
highest (worst) score at pretest, and improved from
33 at pretest to 15 at posttest. On the CROPS a total
score of 19 or more is considered in the clinical
range (Greenwald & Rubin, 1999). At pretest, all
participants in both groups were in the clinical
range on the CROPS. At posttest, three of the
EMDR participants and two of the CBT partici-
pants had moved to the normal range. On the
PROPS a total score of 16 or more is considered in
the clinical range (Greenwald & Rubin, 1999). At
pretest, five of the seven participants in each group
were in the clinical range. In the EMDR group, four
of those five moved to the normal range at posttest.
In the CBT group, three of the five moved from
the clinical to the normal range.
An additional way to consider clinical signifi-
cance is to calculate the reliable change (RC)
index (Jacobson, Follette, & Revenstorf, 1984), by
dividing each participant’s pre- to post-treatment
change score by the standard error of measure-
ment. If RC exceeds 1.96, the change is considered
unlikely to be due to imprecise measurement. As
shown in Table 2, on the Rutter Teacher Scale, four
of the seven participants in each group had RCs in
excess of 1.96. On the CROPS, all seven of the
EMDR participants, as compared to four of the
seven CBT participants, had RCs in excess of 1.96.
On the PROPS, five of the seven EMDR partici-
pants and six of the seven CBT participants had
RCs in excess of 1.96.
Acceptability and Efficiency of Treatments
In a ‘horse race’ study in which no efficacy differ-
ence is detected between the two treatments, other
considerations become more prominent. One is
treatment acceptability. If the participant will not
mount the horse, or cannot stay on it, she will not
finish the race. Therefore, the drop-out rate is a
Table 2. Indicators of clinical significance on three outcome measures, by group
Measure Group EMDR CBT Group
EMDR CBT Normal Clinical Normal Clinical EMDR (n=7) CBT (n=7)
Cohen’s dindicator of effect size (mean pretest to posttest change divided by pretest standard deviation)
CROPS 2.8 1.1
PROPS 1.8 1.1
Rutter 0.71 0.72
Number of participants in clinical and normal range
CROPS
Pretest 0707
Posttest 3425
PROPS
Pretest 2525
Posttest 6152
Rutter
Pretest 5252
Posttest 6170
Number of participants whose reliable change index (RC) scores were significant (RC >1.96)
CROPS 74
PROPS 56
Rutter 44
CBT versus EMDR 365
Copyright © 2004 John Wiley & Sons, Ltd. Clin. Psychol. Psychother. 11, 358–368 (2004)
critical consideration when comparing treatments.
One participant was excluded from the study prior
to randomization, as noted above, due to ongoing
victimization. Following randomization, there
were four drop-outs, two from each condition. In
each condition, one drop-out occurred prior to
treatment onset, and one subsequent to treatment
onset. Thus, in this study there was no evidence
of differential acceptability of one treatment over
another.
Another concern relates to the efficiency of a
treatment, the amount of time and effort it takes to
achieve the desired effect. Other things being
equal, a more efficient treatment can relieve
suffering more quickly, and can make limited
treatment resources more widely available. To
determine efficiency, we first compared the mean
number of sessions it took each group to complete
treatment. (As we noted earlier, treatment lasted
up to 12 sessions, and was terminated sooner if the
SUDS associated with the abuse memory reached
0–2, along with other treatment-specific criteria.)
As shown in Figure 1, the mean number of sessions
per group was 6.1 for the EMDR group and 11.6 for
the CBT group. Using an independent samples
t-test, this difference was statistically significant
(t=7.1; df =12, p<0.000) and large (ES =4.2).
Here we abandon the horse metaphor in favour
of the more modern custom of measuring fuel effi-
ciency in miles per gallon. We then divided each
participant’s change score by the number of
sessions each had. This indicated the mean amount
of change per session on each outcome measure, as
shown in Figure 2. The mean per session change
on the Rutter Teacher scale was 1.4 for the EMDR
group and 0.50 for the CBT group (p=0.04). On the
CROPS, the mean per session change was 3.0 for
the EMDR group and 0.67 for the CBT group (p=
0.04). Thus, the difference between the two groups
was statistically significant on both the Rutter and
CROPS measures. The EMDR group also had a
higher mean per session change on the PROPS;
however, the difference was not significant (p=
0.18), with a mean of 2.4 for the EMDR group and
0.96 for the CBT group. The effect sizes for all three
comparisons were large: 2.0 on the Rutter; 2.3 on
the CROPS; and 2.0 on the PROPS.
DISCUSSION
Several features of this study limit confidence in
interpretation of findings. The limited power
inherent in such a small sample size makes it dif-
ficult to detect possible actual differences between
the treatments. The small sample size also limits
the study’s external validity. The use of one thera-
pist for one treatment condition and another ther-
apist for the other condition makes it impossible
to distinguish therapist effects from the specific
effects of the treatments. Also, although the PI,
who was knowledgeable regarding both treat-
ments, observed all sessions, there were no formal
procedures to monitor or evaluate treatment
adherence. These limitations were unavoidable
given the limited resources (including scarcity of
therapists in Tehran trained in structured trauma
treatment methods) available to the investigators.
Figure 1. Mean number of sessions for each group Figure 2. Mean amount of change per session on each
outcome measure
366 N. Jaberghaderi et al.
Copyright © 2004 John Wiley & Sons, Ltd. Clin. Psychol. Psychother. 11, 358–368 (2004)
An additional limitation is the absence of a no-
treatment condition. The inclusion criteria mitigate
this problem, in that a clinically significant level of
post-traumatic symptoms more than 6 months
after the trauma indicates a degree of chronicity.
Thus, although having a no-treatment control con-
dition would have been preferred, it is unlikely
that the apparent treatment effects found in this
study would have occurred within that 3-month
period in lieu of treatment.
EMDR was clearly far more efficient in terms of
number of sessions to termination criteria, as well
as amount of change achieved per session, at least
on two of the three outcome measures. An addi-
tional efficiency issue, from the participants’ point
of view, is the amount of homework required for
each treatment—much more for CBT. This adds to
the trend of similar findings favouring EMDR re
efficiency, across controlled comparison studies.
Despite the apparently superior efficiency of
EMDR, questions remain on this issue. First of all,
participants in the CBT condition had a required
minimum of 10 sessions whereas there was no
minimum in the EMDR condition. This discrep-
ancy carries the risk of differential expectancy
effects as well as extension of CBT treatment
beyond what would have otherwise been needed.
However, since only one participant in the CBT
condition finished after 10 sessions, this minimum
number of sessions did not artificially extend the
duration of treatment for at least six of the seven
participants in the CBT condition. Even so, the ter-
mination criteria may have biased the findings
regarding the efficiency of EMDR. EMDR is struc-
tured to reduce the SUDS as quickly as possible,
whereas the CBT approach includes skills training
that may be less likely to lead to rapid symptom
reduction. Although quicker relief does seem
preferable, we do not know what the longer term
follow-up results would be. Although participants’
gains have maintained in most of the other com-
parison studies (for both treatments), we cannot be
sure how the participants in this study would
compare over time. Would the EMDR group con-
tinue to do well as a result of participants’ lower
level of post-traumatic symptoms post-treatment?
Or would the CBT group continue to improve over
time, as participants recovered from the exposure
and gained more experience in using the skills they
had learned?
On the other hand, study eligibility criteria, treat-
ment procedures, and termination criteria may
have biased the findings to favour CBT’s efficacy.
Participants were eligible for inclusion if they met
sexual abuse victimization criteria, whether or not
they had experienced additional trauma as well.
In fact, all participants in each group did report
additional trauma history, and furthermore all par-
ticipants in each group endorsed at least one
additional event as continuing to be highly dis-
tressing. Although in the EMDR condition treat-
ment could be terminated once the participant was
no longer distressed by the memory of the index
trauma, participants may have been suffering from
additional trauma that would remain untreated
and possibly continue to drive post-traumatic
symptoms and problem behaviours. On the other
hand, CBT focuses in part on coping skills that can
be broadly applied; and termination criteria
included reduction of anxiety symptoms that were
not specifically referenced to the index trauma
memory. Considering that the outcome measures
also did not target memory-referenced post-
traumatic stress, but global symptom measures, it
might have been more fair to EMDR to continue
with treatment until all trauma had been resolved
(up to the 12 session limit), not just the index
trauma.
Conclusion
Both CBT and EMDR appeared to be efficacious in
helping these sexually-abused Iranian girls to
reduce their post-traumatic stress symptoms, and
in helping the girls to improve their behaviour.
EMDR was considerably more efficient in achiev-
ing these results. These findings suggest that both
CBT and EMDR can help girls to recover from the
effects of sexual abuse, and that structured trauma
treatments can be applied to children in Iran.
However, replication with better power and
controls is warranted.
REFERENCES
Almqvist, F., Kumpulainen, K., Ikaheimo, K., Linna,
S.-L., Henttonen, I., Huikko, E., Tuompo-Johansson, E.,
Aronen, E., Puura, K., Piha, J., Tamminen, T., Rasanen,
E., & Moilanen, I. (1999). Behavioural and emotional
symptoms in 8–9-year-old children. European Child &
Adolescent Psychiatry,8(Suppl. 4), IV/7–IV/16.
American Psychiatric Association (1994). Diagnostic
and statistical manual of mental disorders (4th ed.).
Washington, DC.
Brabin, P.J., & Berah, E.F. (1995). Dredging up past
traumas: Harmful or helpful? Psychiatry, Psychology &
Law,2, 165–171.
CBT versus EMDR 367
Copyright © 2004 John Wiley & Sons, Ltd. Clin. Psychol. Psychother. 11, 358–368 (2004)
Horowitz, M.J. (1986). Stress response syndromes (2nd ed.).
Northvale, NJ: Jason Aronson.
Hyer, L., & Brandsma, J.M. (1997). EMDR minus eye
movements equals good psychotherapy. Journal of
Traumatic Stress,10, 515–522.
Ironson, G., Freund, B., Strauss, J.L., & Williams, J. (2002).
Comparison of two treatments for traumatic stress: a
community-based study of EMDR and prolonged
exposure. Journal of Clinical Psychology,58, 113–128.
Jacobson, N.S., Follette, W.C., & Revenstorf, D. (1984).
Psychotherapy outcome research: Methods for report-
ing variability and evaluating clinical significance.
Behavior Therapy, 15, 336–352.
Kendall-Tackett, K.A., Williams, L.M., & Finkelhor, D.
(1993). Impact of sexual abuse on children: a review
and synthesis of recent empirical studies. Psychological
Bulletin,113, 164–180.
Kresanov, K., Tuominen, J., Piha, J., & Almqvist, F. (1998).
Validity of child psychiatric screening methods. Euro-
pean Child & Adolescent Psychiatry,7, 85–95.
Kumpulainen, K., Rasanen, E., & Henttonen, I. (1998).
The persistence of teacher-reported behavioral prob-
lems among children aged 8 to 12. European Child &
Adolescent Psychiatry,7, 225–234.
Lee, C., Gavriel, H., Drummond, P., Richards, J., &
Greenwald, R. (2002). Treatment of PTSD: Stress Inoc-
ulation Training with Prolonged Exposure compared
to EMDR. Journal of Clinical Psychology,58, 1071–1089.
Martin, J.L., Perrott, K., Morris, E.M., & Romans, S.E.
(1999). Participation in retrospective child sexual abuse
research: Beneficial or harmful? What women think six
years later. In L.M. Williams, & V.L. Banyard (Eds),
Trauma and memory (pp. 149–159). Thousand Oaks, CA:
Sage.
Morita, H., Suzuki, M., & Kamoshita, S. (1990). Screen-
ing measures for detecting psychiatric disorders in
Japanese secondary school children. Journal of Child
Psychology and Psychiatry, 31, 603–617.
Newman, E., Walker, E.A., & Gefland, A. (1999). Assess-
ing the ethical costs and benefits of trauma-focused
research. General Hospital Psychiatry, 21, 187–196.
Power, K., McGoldrick, T., & Brown, K. (2002). A con-
trolled comparison of EMDR versus exposure plus
cognitive restructuring versus wait list in the treatment
of post traumatic stress disorder. Clinical Psychology &
Psychotherapy, 9, 299–318.
Puffer, M.K., Greenwald, R., & Elrod, D.E. (1998). A single
session EMDR study with twenty traumatized children
and adolescents. Traumatology, 3(2). Available Internet:
http://www.fsu.edu/~trauma/v3i2art6.html.
Rutter, M. (1967). A children’s behaviour questionnaire
for completion by teachers: Preliminary findings.
Journal of Child Psychology and Psychiatry, 8, 1–11.
Ruzek, J.I., & Zatzick, D.F. (2000). Ethical considerations
in research participation among acutely injured
trauma survivors: An empirical investigation. General
Hospital Psychiatry, 22, 27–36.
Saywitz, K.J., Mannarino, A.P., Berliner, L., & Cohen, J.A.
(2000). Treatment for sexually abused children and
adolescents. American Psychologist, 55, 1040–1049.
Scheck, M.M., Schaeffer, J.A., & Gillette, C.S. (1998). Brief
psychological intervention with traumatized young
Camino, L. (2000). Treating sexually abused boys: A practi-
cal guide for therapists and counselors. San Francisco:
Jossey-Bass.
Chemtob, C.M., Nakashima, J., Hamada, R., & Carlson,
J.G. (2002). Brief treatment for elementary school chil-
dren with disaster-related posttraumatic stress disor-
der: A field study. Journal of Clinical Psychology,58,
99–112.
Chemtob, C.M., Tolin, D.F., van der Kolk, B.A., & Pitman,
R.K. (2000). Eye movement desensitization and repro-
cessing. In E.B. Foa, T.M. Keane, & M.J. Friedman
(Eds), Effective treatments for PTSD: Practice guidelines
from the International Society for Traumatic Stress Studies
(pp. 139–154). New York: Guilford.
Cohen, J. (1988). Statistical power analysis for the social
sciences. Hillsdale, NJ: Erlbaum.
Cohen, J.A., Berliner, L., & March, J.S. (2000). Treatment
of children and adolescents. In E.B. Foa, T.M. Keane, &
M.J. Friedman (Eds), Effective treatments for PTSD: Prac-
tice guidelines from the International Society for Traumatic
Stress Studies (pp. 106–138). New York: Guilford.
Cohen, J.A., & Mannarino, A.P. (1998). Interventions for
sexually abused children: Initial treatment findings.
Child Maltreatment,3, 17–26.
Deblinger, E., & Heflin, A.H. (1996). Treating sexually
abused children and their nonoffending parents: A cognitive
behavioral approach. Thousand Oaks, CA: Sage.
Deblinger, E., Lippmann, J., & Steer, R. (1996). Sexually
abused children suffering posttraumatic stress
symptoms: Initial treatment outcome findings. Child
Maltreatment,1, 310–321.
Devilly, G.J., & Spence, S.H. (1999). The relative efficacy
and treatment distress of EMDR and a cognitive
behavior trauma treatment protocol in the ameliora-
tion of Post-traumatic Stress Disorder. Journal of
Anxiety Disorders,13 (1–2), 131–157.
Dolatabadi, S. (2000). The comparison of EMDR (eye move-
ment desensitization and reprocessing) and SD (systematic
desensitization) in Iranian female college students. Unpub-
lished thesis, Alame Tabatabaee University, Tehran,
Iran.
Edmonds, T., & Rubin, A. (1999). The effectiveness of
EMDR with adult female survivors of childhood
sexual abuse. Social Work Research, 103–116.
Edmonds, T., Rubin, A., & Wambach, K.G. (1999). The
effectiveness of EMDR with adult female survivors of
childhood sexual abuse. Social Work Research, 23,
103–116.
Fletcher, K.E. (1993). The spectrum of post-traumatic
responses in children. Presented at the Annual Meeting
of the International Society for Traumatic Stress
Studies, San Antonio.
Foa, E.B., & Meadows, E.A. (1997). Psychosocial treat-
ments for posttraumatic stress disorder: A critical
review. Annual Review of Psychology,48, 449–480.
Greenwald, R. (1999). Eye movement desensitization and
reprocessing (EMDR) in child and adolescent psychother-
apy. Northvale, NJ: Jason Aronson.
Greenwald, R., & Rubin, A. (1999). Brief assessment of
children’s post-traumatic symptoms: Development
and preliminary validation of parent and child scales.
Research on Social Work Practice,9, 61–75.
368 N. Jaberghaderi et al.
Copyright © 2004 John Wiley & Sons, Ltd. Clin. Psychol. Psychother. 11, 358–368 (2004)
women: The efficacy of eye movement desensitization
and reprocessing. Journal of Traumatic Stress, 11, 25–44.
Shapiro, F. (1995). Eye movement desensitization and repro-
cessing: Basic principles, protocols and procedures. New
York: Guilford Press.
Soberman, G.B., Greenwald, R., & Rule, D.L. (2002). A
controlled study of eye movement desensitization and
reprocessing (EMDR) for boys with conduct problems.
Journal of Aggression, Maltreatment, and Trauma, 6,
217–236.
Sweet, A. (1995). A theoretical perspective on the clinical
use of EMDR. The Behavior Therapist, 18, 5–6.
Taylor, S., Thordarson, D.S., Maxfield, L., Fedoroff, I.C.,
Lovell, K., & Ogrodniczuk, J. (2003). Comparative effi-
cacy, speed, and adverse effects of three PTSD treat-
ments: Exposure therapy, EMDR, and relaxation
training. Journal of Consulting and Clinical Psychology,
71, 330–338.
Thyer, B.A., Papsdorf, J.D., Davis, R., & Vallecorsa, S.
(1984). Autonomic correlates of the subjective anxiety
scale. Journal of Behavior Therapy and Experimental Psy-
chiatry, 15, 3–7.
Wilson, D., Silver, S.M., Covi, W., & Foster, S. (1996). Eye
movement desensitization and reprocessing: Effective-
ness and autonomic correlates. Journal of Behavior
Therapy and Experimental Psychiatry, 27, 219–229.
Yousefi, F. (1998). The normalization of Rutter test
(teacher form): Assessing the behavioral and emo-
tional difficulties of Iranian students. The Journal of
Society and Humanity Sciences of Shiraz University, 25 &
26, 171–194.
