Filetype PDF | Posted on 26 Sep 2022 | 4 years ago
Authentication
digital resources
315x Filetype PDF File size 0.56 MB Source: coping.us
Acta Psychiatr Scand 2008: 1–9 Copyright 2008 The Authors
All rights reserved ACTAPSYCHIATRICA
DOI: 10.1111/j.1600-0447.2008.01270.x SCANDINAVICA
Review
Group cognitive behavioural therapy for
obsessive–compulsive disorder: a systematic
review and meta-analysis
´ 1,2 1
Jonsson H, Hougaard E. Group cognitive behavioural therapy for H. Jnsson , E. Hougaard
obsessive–compulsive disorder: a systematic review and meta-analysis. 1
Department of Psychology, University of Aarhus,
2
Aarhus and Clinic for Obsessive Compulsive Disorder,
Objective: Behaviour therapy with exposure and response prevention Aarhus University Hospital, Risskov, Denmark
(ERP) or cognitive behavioural therapy (CBT) including ERP are
considered the psychological treatments of choice for obsessive–
compulsive disorder (OCD), but group CBT⁄ERP has received
relatively little research attention in the treatment of OCD. The aim of
this study was to provide a meta-analysis of the effectiveness of group
CBT⁄ERPfor OCD.
Method: A systematic literature search was conducted and studies
were meta-analysed by means of the Cochrane Review Manager
Program with measures of i) pre- to post-effect sizes (ES) and
ii) between-group ES in comparison with different control conditions.
Outcome was primarily measured on the Y-BOCS and ES was
calculated in the form of Cohens d.
Results: Thirteen trials were included in the meta-analysis. The overall
pre–post-ES of these trials of 1.18 and a between-group ES of 1.12
compared with waiting list control in three randomized controlled
studies indicate that group CBT⁄ERP is an effective treatment for Key words: obsessive–compulsive disorder; meta-
OCD. Group CBT achieved better results than pharmacological analysis; cognitive behavioural therapy; group therapy
treatment in two studies. One study found no significant differences Hjalti Jnsson, Department of Psychology, University of
between individual and group CBT. Aarhus, Jens Chr. Skous Vej 4, 8000 Aarhus, Denmark.
Conclusion: Group CBT is an effective treatment for OCD, but more E-mail: hjalti@psy.au.dk
studies are needed to compare the effectiveness of group and individual
treatment formats. Accepted for publication August 21, 2008
Summations
• Group cognitive behavioural therapy including exposure and response prevention is an effective
treatment for obsessive–compulsive disorder.
• There are insufficient data on the relative effectiveness of individual and group cognitive behavioural
therapy including exposure and response prevention for obsessive–compulsive disorder.
Considerations
• There are few studies of the effectiveness of group cognitive behavioural therapy including exposure
and response prevention for obsessive–compulsive disorder and only four randomized control trials.
Introduction debilitatingdisorderwithachroniccourseifuntreated.
Epidemiological studies have found lifetime preva-
Obsessive–compulsive disorder (OCD) is character- lence estimates of OCD to be about 1–2% (2, 3).
ized by persistent, intrusive thoughts (obsessions) Overthelasttwodecades,researchershavemade
and⁄or stereotyped repetitive behaviours carried out progress in identifying effective treatments includ-
in a ritualistic fashion (compulsions) (1). It is a ing psychotherapy, pharmacotherapy and their
1
´
Jonsson and Hougaard
combination.ThepsychologicaltreatmentforOCD of Science and The National Research Register,
with the highest degree of empirical support is from the first available year to 01.02.07, using the
individualexposureandresponseprevention(ERP) keywords [(obsess* or compul* or ocd) AND
(4, 5). Most clinicians today, however, supplement (group next therap*) OR (group next treatment*)].
the behavioural methods with cognitive methods, In addition, the reference lists of other reviews and
although there is at present no empirical evidence selected articles were inspected for further relevant
showing that cognitive behavioural therapy (CBT) studies.
withERPachievesbetteroutcomesthanERPalone
(4). Meta-analyses suggest that ERP and CBT with Inclusion criteria
ERPachievelarge effects in pre–post-conditions or
compared with waitlist or placebo conditions. For The following criteria were used for inclusion of
example,Eddyet al.(6)foundanuncontrolled,pre– studies: i) participants aged 18 years or above, ii) a
post-effect size (ES) of 1.52 (Cohens d) in a meta- primary diagnosis of OCD according to a stan-
analysis of 13 randomized controlled trials (RCTs) dardized diagnostic classification system (e.g.
andacontrolled,between-groupESof1.12basedon DSM-III or later editions), iii) interventions in
three of these studies. In the same review, a meta- the form of group ERP or group CBT, iv)
analysis of 32 RCTs of pharmacological treatment outcomes reported with means and standard devi-
for OCD reported an uncontrolled pre–post-ES of ations on the Yale-Brown Obsessive–Compulsive
1.18 and a controlled, between-group ES of 0.83. Scale (Y-BOCS) [clinical rating form or self-report
Acomparisonofpre–post-ESforERPtreatmentof version (only one study (9) used self-report ver-
110patientsinanaturalistictreatmentsettingfound sion)], v) number of participants in each treatment
outcomessimilartothoseachievedinfourRCTs(7) condition ‡10, vi) studies available in English or
thusindicatingthatthemethodiseffectiveoruseful German language and vii) published in peer-
in general clinical practice. reviewed journals.
Group CBT⁄ERP for OCD has been proposed Exclusion criteria: i) studies limited to patients
as a cost-effective treatment format. In a qualita- with only hoarding symptoms or patients with
tive review of 12 studies of group CBT or ERP for obsessions only, ii) studies where patients received
adults Himle, Van Etten and Fischer (8) concluded combined individual and group therapy, and iii)
that there was some evidence of the effectiveness of studies where the duration of treatment was more
group CBT or ERP, although limitations in than 20 weeks.
quantity and quality of the research made conclu- All decisions on inclusion⁄exclusion criteria were
sions rather tentative. The meta-analysis by Eddy made a priori. Two authors were contacted
et al. (6) found somewhat larger uncontrolled, pre– because of missing statistical information; Fals-
post-ESs for individual therapy (1.48) than for Stewart (10), and Sousa (11), and the needed
group therapy (1.17). However, only two studies information from the second-mentioned author
on group treatment were included in their analysis was retrieved.
and the authors did not report whether the
difference reached statistical significance. Methodological quality of studies
Aims of the study Studies were ranked into three categories:
Theaimofthisstudywastoprovideameta-analysis i) Randomized controlled studies; i.e. studies
of group cognitive behavioural therapy (CBT) and comparing group ERP or group CBT to
exposure and response prevention (ERP) for OCD, placebo control, waitlist control or to other
whichhasnot,asfarasweknow,beendonebefore. active treatments.
Thereview primarily analyses the overall pre–post- ii) Controlled studies; i.e. studies with control
effect size (ES) of group CBT and ERP therapy for conditions but without randomized group
OCD and, secondarily, between-group ESs for allocation.
different control conditions. iii) Open clinical trials with outcome measures
before and after therapy but no control
conditions.
Material and methods Quality of individual studies was independently
Identification of studies assessed, by the two authors of the paper, on the
Cochrane Collaboration Depression, Anxiety and
Studies were located by searching the following Neurosis Group (CCDAN) quality rating scale a
databases: PsychInfo, EBSCO host, PubMed, Web 23-item scale with total scores from 0 to 46 (12).
2
Group CBT for OCD
The consistency between the two raters was Results
acceptable with a Cronbachs a value of 0.93. In Trial flow
case of a substantial disagreement (>3 points on
the scale), the differences were discussed and new The electronic search strategy yielded 1749 publi-
consensual ratings were applied. cations. After abstract screening, 37 studies were
retrieved for more detailed evaluation with two
Statistical analysis additional studies found from references. Based on
inspection of papers, 13 primary studies were
DatafromtheY-BOCSwereconsideredasprimary judged to fulfil the inclusion criteria; four RCTs
outcome measures. Data were entered into the (11, 18–20), four controlled studies (21–24) and five
computer software review manager 4.2 (RevMan), open studies (9, 25–28). A list of excluded studies
provided by The Cochrane Collaboration (13). For and reasons for exclusion are available on request
continuous outcomes, the software calculates stan- from the corresponding author.
dardized weighted mean difference based on
Cohens (14) d and 95% confidence intervals. The Study characteristics
random-effects model was used, which yields supe-
rior results in terms of clinical interpretability and Table 1 summarizes selected characteristics of the
external generalizibility to other clinical contexts 13 studies. Number of participants in the studies
compared with analyses based on the fixed effects varied from 20 to 155 with a total of 828
model (15). The RevMan software calculates two participants. Of these, a total of 549 received
additional statistics for estimating heterogeneity of group therapy (395 group CBT and 154 group
2 2 ERP), 79 functioned as waitlist controls, 83
studies, the I-squared (I ) and Q-statistics. The I
statistics indicates the percentage of variance in a received pharmacological treatment, 25 received
pooledESthatcanbeattributedtoheterogeneityin individual CBT, 17 received group relaxation
the sample of studies (16). Values of 25% are training and 75 received other sorts of active
consideredlow,50%asmoderateand75%ashigh. treatments (50 received a mixture of group therapy
TheQ-statistics calculates P-values for heterogene- and pharmacotherapy and 20 received multifamily
ity of studies (i.e. P-values £0.05 indicates signif- group ERP).
icant heterogeneity). Mean age of the group treatment sample was
As a supplement to these analyses, Fail-Safe N 36.4 years, about 63% was females, and approxi-
was calculated as a measure of how vulnerable mately 54% was in pharmacological treatment at
meta-analysis ES findings are to the possibility of treatment onset. The overall mean on the Y-BOCS
undiscoveredstudies (17). Fail-Safe N estimates the scale at inclusion was 23.4 (SD = 1.62, range
number of undiscovered, approximately equally 21.2–26.7). There were no significant differences in
sized studies with an ES of zero needed to reduce Y-BOCS scores at treatment start between the
the overall ES to a certain criterion level, e.g. to 0.5 three categories of studies [F(2, 12) = 0.519
or 0.2, ESs classified as medium and small by P=0.608]. Treatment duration ranged from
Cohen (14). seven to 16 weekly sessions with an average of 11
ByusingtheRevMansoftware,pooledmeanpre– sessions with each session lasting 1½–2½ h (mean
post-ES, weighted according to numbers of partic- 120 min). Group sizes varied from four to 10
ipants in the studies, were computed for each of the participants (mean 6.8) with one or two therapist(s)
three methodological categories as well as for total in each group. Follow-up ratings were obtained in
number of studies. Between-group ES were calcu- nine of the included studies, and the follow-up
lated for comparisons of group CBT⁄ ERP to i) period ranged from 1 month to 4 years, with an
waitlist control group, ii) placebo control, iii) average of 12.3 months (median 3 months)
individual CBTandiv)pharmacologicaltreatment. (follow-up data are not analysed in this study).
In case of evidence of marked heterogeneity Drop-out rate was 13.5% in group treatment,
2 compared with 8.5% in other treatments and
(I > 50%) studies with markedly higher⁄lower
ES were tentatively considered as outliers and 11.4% in the waiting list control conditions.
excluded from the analysis. A sensitivity analysis
was then conducted comparing meta-analyses with Quantitative data synthesis
and without the outliers checking for significant
differences. Thirteen studies were included in the meta-anal-
In addition, overall pooled mean pre–post-ESs ysis with 15 comparisons. Separate analyses were
were computed for other outcome measures of performed on CBT and ERP group therapy in
symptoms of depression and anxiety. McLean et al. (20) and on participants receiving
3
´
Jonsson and Hougaard
CCDAN*2929 32.5 28 26 19.5 18.5 20 20 24
-BOCS7.37.30.610.87.411.61.55.09.70.86.45.48.38.92.97.26.77.38.65.712.6
Meanchange ii) i)ii)iii)ii) )
Y i)ii)iii)i)ii)i) i)ii)iii)i)ii) i) i)ii) i i)
sizeof(2) (1) (2) (2) (1)(1)
5–7 5–8 7–8 6–8 7–94–75–85–8 7–85 R
Grouptherapists)) )
(numberi) i i) i) i)ii)iii)NR i)ii) N 5–8
+ 49.9
term
Follow-upmonth monthsmonthsmonths monthslong(meanmonths)monthsmonths
1 – 3 3 12 – 3 12 – 3
h
72 ses-
reatmentdurationweeklysessionsweeklysessionsweeklysessionsweeklysessionsweeklysessionsweeklysessionsorweeklysessionsweeklyhweeklysession
T h h h h h h h h
102 122 122 122 122 102 12weeklysessions10–122102.5sions122
prior months to
months 3 months during not
3 3 how
treatmentmonthtreatmenttreatmentalloweddosesstablefor
medicationstable1stablestablenot
On to to to kept stable-but
63% 0%stabletreatment44,7%48%34%0%60% 36%56.2%
i)priori)ii)toi)priori)priori)increasetreatmenti)i)NotNRKeptreportedlongi)i)
(SD) (NR) (11.57)(11.2) (NR)(10.3) (11.2)
Meandisorder13.623.521.8 18.517.4 23.6
durationofi)i) i) NR i) i) NR NR NR i)
of(SD)(NR)(7.58) (6.64) (7.2) (9.02)(6.64)
Meanageonset19.91514.7 19.6 10.715
i) i) i) NR NR i) NR NR i) i)
67.9 80 51 48 83 NR 62 67 78 69
Percentwomeni)i) i) i) i) i) i) i)
(SD)(NR)(11.79)(12.8)(NR) (NR)(NR) (10.86)(NR) (8.85)(12.8)
Meanage33.738.5 36.5 35 39.534.1 37 33.0 34.839.5
) )
i) i) i) i) i) i) i) i) i i
5 3 3 1 35 6 2 0 NR 1
4⁄ ⁄ 3⁄ 1⁄ 1⁄⁄ 1 ⁄ 10⁄⁄ NR⁄ 5 ⁄ 8 2
drop-⁄ 17⁄ ⁄ ⁄ 38 ⁄ ⁄ 55⁄ ⁄ ⁄ ⁄
⁄ outs21252828 2324 1919 24174555 2489 2220 53 32
n i)ii)iii)i)ii)i)ii)i)ii)iii)i)ii)i)ii)iii)i)ii)i)ii)i)i)
or brain
16),severe abuse
> risk; mentalsevereother
phobia,, 24 , suicidalsevere(borderline
mentalmoderate(BDIabuse,severeseverementaland>alcohol,
criteria psychotic physical diagnosis, withmental
intellectual,suicidalsocialdisorderdisorders,epilepsyI
disorderwithdisordersorganicMADRS disorderdisorder
organic substancenervosa,severepsychosisaxis
Exclusion,syndrome,depressiondisorder,retardation,,s,symptoms;
´s severethought physical depression
, bipolar
studies schizotypical),
disabilitydisorderouretteseverebipolardisorderpersonalityDepressioninjurymentalanorexiapersonalityActiveretardation,disorderdisordersourettedepressionschizophrenia,comorbidNRNRrisk,personalityorretardation
Schizophrenia,T T Severe PsychoticMajor
studies + – –
– controlled
include CBT CBT CBT CBTERPCBTRTCBTCBTERPERPERPERPCBTCBT
of aitlist aitlistaitlist weeksweeks
InterventionGroupIndividualCBTWGroupSertralineGroupWGroupGroupWdelayedGroupGroup-basedGroupGroupmedicationMedicationGroup12Group7GroupMultifamilygroupGroupGroup
i)ii)iii)i)ii)i)ii)i)ii)iii)i)ii)i)ii)iii)i)ii)i)ii)studiesi)i)
s
RCTand Non-randomized Open
Characteristics1:(18)(11)(19)(20)2:(21)(22)(23) (24)3:(25)(26)
1. al. al. al. al.al. al. Noppenal.al.al.
Rees et et et et et et n et et et
a
able AndersonSousaCordioliMcLeanFinebergAignerHimleVMeierCordioli
T StudyCategory Category Category
4
no reviews yet
Please Login to review.
