Filetype PDF | Posted on 26 Sep 2022 | 3 years ago
Authentication
digital resources
109x Filetype PDF File size 0.15 MB Source: i-cbt.org.ua
Journal of Consulting and Clinical Psychology Copyright 2005 by the American Psychological Association
2005, Vol. 73, No. 5, 953–964 0022-006X/05/$12.00 DOI: 10.1037/0022-006X.73.5.953
Randomized Trial of Prolonged Exposure for Posttraumatic Stress Disorder
With and Without Cognitive Restructuring: Outcome at Academic and
Community Clinics
Edna B. Foa, Elizabeth A. Hembree, Norah C. Feeny
Shawn P. Cahill, Sheila A. M. Rauch, and Case Western Reserve University
David S. Riggs
University of Pennsylvania
Elna Yadin
University of Pennsylvania
Female assault survivors (N 171) with chronic posttraumatic stress disorder (PTSD) were randomly
assigned to prolonged exposure (PE) alone, PE plus cognitive restructuring (PE/CR), or wait-list (WL).
Treatment, which consisted of 9–12 sessions, was conducted at an academic treatment center or at a
community clinic for rape survivors. Evaluations were conducted before and after therapy and at 3-, 6-,
and 12-month follow-ups. Both treatments reduced PTSD and depression in intent-to-treat and completer
samples compared with the WL condition; social functioning improved in the completer sample. The
addition of CR did not enhance treatment outcome. No site differences were found: Treatment in the
hands of counselors with minimal cognitive–behavioral therapy (CBT) experience was as efficacious as
that of CBT experts. Treatment gains were maintained at follow-up, although a minority of patients
received additional treatment.
Keywords: exposure therapy, posttraumatic stress disorder, cognitive restructuring, rape
Despite the progress that has been made in the development of used by clinicians in the community who treat patients with
efficacious psychosocial treatments for chronic posttraumatic trauma-related disturbances (Becker, Zayfert, & Anderson, 2004).
stress disorder (PTSD; Foa, Keane, & Friedman, 2000), on the In the current study, we address these issues by examining two
average, treated patients remain somewhat symptomatic (Cahill & strategies for improving outcome and by comparing outcome in an
Foa, 2004). For example, the mean posttreatment score of the academic center and a community clinic.
Clinicians Administered PTSD Scale (Blake et al., 1990) was The wide range of PTSD symptoms has led some experts to
greater than 30 (a score of 20 defines remission) in the studies by suggest that treatment programs with multiple techniques will
Marks, Lovell, Noshirvani, Livanou, and Thrasher (1998) and be more efficacious than any single approach (e.g., Kilpatrick,
Bryant, Moulds, Guthrie, Dang, and Nixon (2003). Moreover, the Veronen, & Resick, 1982). Accordingly, most cognitive–
treatments developed in academic clinical centers are not widely behavioral therapy (CBT) programs for PTSD include several
techniques (e.g., Blanchard et al., 2003). Foa, Dancu, et al.
(1999) examined the hypothesis that prolonged exposure (PE)
Edna B. Foa, Elizabeth A. Hembree, Shawn P. Cahill, Sheila A. M. combined with stress inoculation training (SIT) would be su-
Rauch, David S. Riggs, and Elna Yadin, Department of Psychiatry, Uni- perior to either PE or SIT alone. Contrary to the predictions, the
versity of Pennsylvania; Norah C. Feeny, Department of Psychiatry, Case three treatments performed equally well on most measures. PE
Western Reserve University. alone, which focuses on exposure to trauma-related memories
Sheila A. M. Rauch is now at the Psychiatry Service, Veterans Affairs and situations, was superior on anxiety and global social ad-
Medical Center, Ann Arbor, Michigan. justment and yielded larger effect sizes on severity of PTSD,
This study was supported by National Institute of Mental Health Grant depression, and anxiety at posttreatment and follow-up. The
MH42178 awarded to Edna B. Foa. We wish to acknowledge our colleague authors suggested that the combined treatment put an excessive
Constance V. Dancu, who supervised the therapists of Women Organized demand on patients (SIT alone involved seven techniques) and
Against Rape throughout the study. We thank David M. Clark, who provided thus attenuated its potential superiority. The first aim of this
training and consultation in cognitive restructuring. We also thank the study study was to compare the efficacy of PE alone with a program
therapists: Lee Fitzgibbons, Lisa Jaycox, Kelly Johnson, Ann Gaulin, Deborah that combined PE plus cognitive restructuring (PE/CR), a po-
Callahan, Latonda Redick, Jane Folk, Katherine Rehm, and Mary Long. tent technique for ameliorating anxiety disorders.
Correspondence concerning this article should be addressed to Edna B.
Foa, Center for the Treatment and Study of Anxiety, Department of Customarily, CBT programs used in research include a fixed
Psychiatry, University of Pennsylvania, 3535 Market Street, Suite 600N, numberoftreatmentsessions. In typical clinical practice, however,
Philadelphia, PA 19104. E-mail: foa@mail.med.upenn.edu treatment duration is determined by the patient’s response to
953
954 FOAETAL.
treatment. The second aim of the study was to examine whether ization. Thus, our intent-to-treat sample consisted of 179 women who
additional sessions would enhance outcome for patients who did signed consent, were randomized to a condition, and were not removed by
not reach an excellent response after eight sessions. the investigators. Twenty-six were assigned to WL, 74 to PE/CR, and 79
The vast majority of knowledge about the efficacy of CBT for to PE. A total of 105 were treated at the CTSA, and 74 were treated at
PTSD has been derived from studies conducted in academic cen- WOAR.
ters, but CBT has not been commonly used in community clinics Treatment completers were 121 participants (44 in PE/CR and 52 in
that specialize in treating trauma survivors, such as rape counsel- PE) who attended at least eight therapy sessions and completed a
ing centers. Becker et al. (2004) found that most community posttreatment assessment. Twenty-five participants in the WL condition
completed a posttreatment assessment. The overall dropout rate was
therapists do not use imaginal exposure with PTSD sufferers 32.4% and was lower for WL (3.8%) than PE/CR (40.5%), 2(1, N
primarily because they lack training. Thus, the third aim of the 100) 12.1, p .001, and PE (34.2%), 2(1, N 105) 9.2, p .01.
study was to provide training in the use of PE and CR to rape Dropout rates did not differ between the two treatments, 2(1, N
counselors and to compare the outcome of patients treated by these 153) 1, ns, or across sites (CTSA 33.3% and WOAR 31.1%),
counselors with those treated in the academic center where the PE 2(1, N 179) 1, ns.
protocol was developed. Demographic information for the intent-to-treat sample is summa-
To achieve the aims of the study, we enrolled women with rized in Table 1. Participants had a mean age of 31 years and were
chronic PTSD resulting from rape, nonsexual assault, or childhood predominately Caucasian or African American, single, with at least
sexual abuse in a treatment outcome study. Enrollment was either some college education. Nearly half of the participants reported annual
through the Center for the Treatment and Study of Anxiety incomes of $15,000 or less, and one third were not working or on
disability. Sexual assault during adulthood was the most prevalent
(CTSA), an academic center, or through Women Organized index trauma, defined as the one experienced by the patient as currently
Against Rape (WOAR), a Philadelphia community clinic for rape most distressing or most frequently reexperienced or both. The average
survivors where therapists had no prior experience with CBT. time since the index trauma was 9 years. Almost all participants either
Participants at each site were randomly assigned to PE alone, witnessed or directly experienced at least one traumatic event in addi-
PE/CR, or wait-list (WL) control. Participants in active treatment tion to the index trauma, and more than 80% directly experienced at
who, at the end of eight sessions, reached at least 70% improve- least one additional incident of interpersonal violence. Psychiatric
ment in self-reported PTSD symptoms completed treatment after comorbidity was common, with 67% of the sample having at least one
Session 9. The rest were offered up to 12 sessions. We hypothe- comorbid Axis I disorder. The most common comorbid conditions were
sized the following: (a) There would be greater reduction in PTSD, as follows: major depression (41.2%), social anxiety disorder (20.4%),
depression, and social and work dysfunction in PE and PE/CR than specific phobias (20.4%), generalized anxiety disorder (13.9%), and
panic disorder (11.9%). All other disorders were present at rates of 6%
in WL; (b) PE/CR would be superior to PE alone on all four or less.
measures; (c) Outcome at the CTSA would be superior to that at Significant site differences were found on five demographic variables:
WOARonallfourmeasures;(d)Participantswhofailedtoachieve age, index trauma, relationship status, employment, and overall comorbid-
excellent response on self-reported PTSD at Session 8 would ity, although sites did not differ on any specific disorder. There was also a
further improve after additional sessions; and (e) Participants in trend for a difference in ethnicity (see Table 1).
2
active treatment would maintain their gains at follow-up on all four Completers differed from noncompleters on level of education, (4,
measures. N177)11.8,p.05,being more likely to have a bachelor’s degree
or higher (34% vs. 12%) and less likely to have not completed high school
Method (8% vs. 17%). There were trends for completers to be older (M 32.2,
SD9.7)than noncompleters (M 29.3, SD 10.0), t(176) 1.9, p
Participants .064, and to be employed full time (43% vs. 33%) or to be students (22%
vs. 14%) rather than unemployed (17% vs. 35%), 2
(4, N 176) 7.9,
Participants were referred by police departments, victim advocacy p .096. There was a trend for completion rates to differ across traumas,
workers, and other professionals, or they were recruited through adver- 2
(2, N 179) 4.6, p .099, with 63% of survivors of adult rape, 76%
tisements in city newspapers and flyers. Recruitment occurred between of nonsexual assault, and 81% of childhood sexual abuse completing
January 1995 and September 2000. Eligible participants were adult treatment. Notably, comorbidity, exposure to additional trauma, or direct
women with a primary diagnosis of PTSD related to a sexual or experience of additional interpersonal violence was not associated with
nonsexual assault that occurred at least 3 months prior to the evaluation dropout status (all 2 values 1, ns).
or to childhood sexual abuse (i.e., the index trauma). Exclusion criteria Twelve serious adverse events led to termination in the study, six of
were as follows: being in an abusive relationship; current diagnosis of which are included in the postrandomization removal category in Figure 1
organic mental disorder, schizophrenia, or psychotic disorder; unmedi- (4 participants reassaulted, 1 developing a life threatening illness, and 1
cated, symptomatic bipolar disorder; substance dependence; and illit- death). The remaining six serious adverse events were classified as drop-
eracy in English. Women deemed at high risk for suicidal behavior (i.e., outs (4 had severe depression and suicidal ideation that required immediate
with intent or plan or both) or with recent history of serious self- intervention, 2 of which were hospitalized, and 2 exhibited extreme dis-
injurious behavior (i.e., cutting) were also excluded. Women taking sociative symptoms).
psychiatric medication (e.g., antidepressants) were required to have
been on a stable dose for at least 3 months prior to entry, and they were Measures
asked to maintain this regimen during treatment.
Figure 1 summarizes participants’ flow from the intake evaluation to Diagnostic Interview
treatment completion. A total of 285 women were evaluated. Fifty-six did
not meet study criteria, 8 were eligible but not interested in the study, and TheStructured Clinical Interview for DSM–IV Axis I Disorders With
11 were lost to contact after evaluation but prior to consenting. Of the 210 Psychotic Screen (First, Spitzer, Gibbon, & Williams, 1995) is a
eligible women who signed consent, 20 withdrew before being assigned a semistructured interview used to assess major Axis I disorders as well
treatment condition, and 11 were removed from the study after random- as to screen for the presence of psychotic symptoms. In the current
PE AND PE/CR FOR PTSD AT CLINICS 955
Figure 1. Flowofparticipants from intake evaluation through completion of treatment. PTSD posttraumatic
stress disorder; WL wait-list; PE/CR prolonged exposure plus cognitive restructuring; PE prolonged
exposure.
study, it was used to assess comorbid conditions and some exclusion Secondary Outcome Measures
criteria.
Beck Depression Inventory (BDI). The BDI (Beck, Ward, Mendelson,
Primary Outcome Measure Mock, & Erbaugh, 1961) is a 21-item self-report inventory measuring
depression severity. Split-half reliability was .93. Correlations with clini-
ThePTSDSymptomScale—Interview(PSS–I;Foa,Riggs,Dancu,& cian ratings ranged from .62 to .66.
Rothbaum, 1993) is a semistructured interview that consists of 17 items Social Adjustment Scale (SAS). The SAS (Weissman & Paykel, 1974)
corresponding to the Diagnostic and Statistical Manual of Mental is a semistructured interview assessing functioning in eight specific areas
Disorders (4th ed.; DSM–IV; American Psychiatric Association, 1994) on separate 7-point scales, with higher scores indicating more severe
PTSDsymptoms.Itemsareratedon0–3scalesforcombinedfrequency maladjustment. We used only the Social and Work scales.
and severity in the past 2 weeks (0 not at all,3 5 or more times PTSD Symptom Scale—Self-Report (PSS–SR). The PSS–SR
per week/very much). Interrater reliability for PTSD diagnosis ( .91) (Foa et al., 1993) is a self-report version of the PSS–I. It is both
and overall severity (r .97) are excellent (Foa et al., 1993). Of the internally consistent ( .91) and stable over a period of 1 month
audiotaped PSS–I interviews in the current study, 5% were randomly (r .74). Symptoms were rated for frequency/severity in the past
selected for rating by a second evaluator. The interrater reliability was .94. week.
956 FOAETAL.
Table 1
Demographics of the Intent-to-Treat Sample (Total and at Each Site)
Variable Total CTSA WOAR Statistic
Age (years)
a
M(SD) 31.3 (9.8) 33.4 (10.2) 28.2 (8.5) t(170.5) 3.7, p .001
N 178 105 73
Years since index trauma
M(SD) 9.0 (11.3) 8.8 (10.2) 9.3 (12.9) t(135) 1.0, ns
N 137 81 56
Index trauma, n (%)
Sexual assault 123 (68.7) 63 (60.0) 60 (81.1) 2
(2, N 179) 17.2, p .001
Nonsexual assault 25 (14.0) 24 (22.9) 1 (1.4)
Childhood sex abuse 31 (17.3) 18 (17.1) 13 (17.6)
N 179 105 74
Witnessed or experienced 171 (96.6) 98 (95.1) 73 (98.6) Fisher’s Exact Test, ns
other (nonindex) traumatic
event, n (%)
N 177 103 74
Experienced other 145 (82.9) 82 (80.4) 63 (86.3) 2
(1, N 177) 1.0, ns
interpersonal violence, n (%)
N 175 102 73
Ethnicity, n (%)
African American 78 (43.6) 40 (38.1) 38 (51.4) 2
(2, N 179) 5.1, p .08
Caucasian 88 (49.2) 59 (56) 29 (39.2)
Other 13 (7.3) 6 (5.7) 7 (9.5)
N 179 105 74
Relationship, n (%)
Single 109 (61.6) 52 (50.0) 57 (78.1) 2
(2, N 177) 16.6, p .001
Married/cohabiting 38 (21.5) 26 (25.0) 12 (16.4)
Divorced/separated 30 (16.9) 26 (25.0) 4 (5.5)
N 177 104 73
Employment, n (%)
Not working 40 (22.7) 26 (25.0) 14 (19.4) 2
(4, N 176) 12.1, p .05
Part time 19 (10.8) 9 (8.7) 10 (13.9)
Full time 70 (39.8) 45 (43.3) 25 (34.7)
Disability 13 (7.4) 11 (10.6) 2 (2.8)
Student 34 (19.3) 13 (12.5) 21 (29.2)
N 176 104 72
Education, n (%)
Some high school 19 (10.7) 13 (12.5) 6 (8.2) 2
(4, N 177) 4.9, ns
High school/GED 34 (19.2) 22 (21.2) 12 (16.4)
AAor some college 77 (43.5) 40 (38.5) 37 (50.7)
BA/BS 25 (14.1) 13 (12.5) 12 (16.4)
Greater than BA/BS 22 (12.4) 16 (15.4) 6 (8.2)
N 177 104 73
Income, n (%)
Less than or equal to 15,000 82 (47.4) 48 (47.5) 34 (47.2) 2
(3, N 173) 2.1, ns
15,001–30,000 42 (24.3) 24 (23.8) 18 (25.0)
30,001–50,000 28 (16.2) 19 (18.8) 9 (12.5)
Greater than 50,001 21 (12.1) 10 (9.9) 11 (15.3)
N 173 101 72
Any current comorbid
Axis I condition, n (%) 103 (67.3) 55 (61.1) 48 (76.2) 2
(1, N 153) 3.8, p .05
N 153 90 63
Note. CTSA participants treated at the university-based Center for the Treatment and Study of Anxiety; WOAR participants treated at the
community-based WomenOrganizedAgainstRape;GEDgeneralequivalencydiploma;AAAssociateofArtsdegree;BABachelorofArtsdegree;
BS Bachelor of Science degree.
a Degree of freedom adjusted because of unequal variances.
both the CTSA and the WOAR participants. Participants were evaluated at
Procedure
their treatment site. The Structured Clinical Interview for DSM–IV Axis I
Evaluations Disorders With Psychotic Screen was administered only at pretreatment;
Independent evaluations were conducted at pretreatment and posttreat- PSS–I and SAS were administered at all evaluations. At each assessment
ment and 3-, 6-, and 12-month posttreatment. All evaluations were con- point, participants completed the PSS–SR and BDI. Participants were
ducted by trained doctoral or master’s level CTSA clinicians who were instructed by their therapists and the evaluators to not reveal any informa-
blind to study condition. The same evaluators conducted assessments for tion that might unblind the evaluator to treatment condition. After the
no reviews yet
Please Login to review.
