Toc Epr

Journal of Obsessive-Compulsive and Related Disorders 31 (2021) 100684
Contents lists available at ScienceDirect
Journal of Obsessive-Compulsive and Related Disorders

journal homepage: www.elsevier.com/locate/jocrd
A systematic review and meta-analysis on the effectiveness of exposure and

response prevention therapy in the treatment of
Obsessive-Compulsive Disorder
Cloe Ferrando *, Caroline Selai
Department of Clinical and Movement Neurosciences, University College London, Queen Square Institute of Neurology, Queen Square London, WC1N 3RX, UK
A R T I C L E I N F O A B S T R A C T
Keywords: Exposure and Response Prevention (ERP) is considered the most effective psychotherapeutic treatment for
Obsessive-compulsive disorder Obsessive-Compulsive Disorder (OCD). The literature supports its adoption yet results vary and vagueness
Exposure with response prevention regarding therapy protocol exists. We present an updated review and meta-analysis to provide clarity in the
Meta-analysis
comparison of strict ERP protocol to control or active therapy groups. Moderator analyses were conducted to
investigate treatment effect of cognitive elements, hours of therapy and duration of OCD.
A systematic literature search, concluded in January 2021, identified twenty-four studies, published between
1997 and 2018, including a total of 1134 patients. The main analysis assessed the difference between pre-
treatment and post-treatment scores compared amongst ERP and the other groups.
We found a statistically significant different reduction in pre-treatment to post-treatment Yale-Brown
Obsessive Compulsive Scale scores between ERP therapy versus other groups. We also found a statistically sig
nificant moderation effect of cognitive elements. The other two moderators, hours of therapy and OCD duration,
were non-significant.
Our review suggests that ERP was superior to the other groups, including both neutral and active treatments,
in reducing OCD symptomatology and should therefore be recommended as an optimal therapy. Future research
should focus on tailoring ERP to the individual and investigating further refinements.
1. Introduction et al., 2015). Although all those with clinical diagnoses will experience
these major symptoms, OCD is heterogeneous in nature; various
1.1. Background sub-types of the disorder will reflect the individual’s personal idiosyn
cratic worries (Robbins et al., 2019). These generally centre around
Obsessive-Compulsive Disorder is frequently considered to refer to specific themes: harm, accompanied by checking compulsions; symme
an exaggerated preoccupation with perfectionism and cleanliness, try, accompanied by ordering rituals; contamination, accompanied by
which is often even encouraged (Durna et al., 2019). Albeit rates of washing routines; sex, violence and religion, accompanied by mental
sub-clinical obsessive and compulsive tendencies are seen in the general rituals, which are also seen in the other sub-types (Abramowitz et al.,
population (Stein, 2002), these traits do not capture the reality of living 2010; McKay et al., 2004). These symptoms are highly time-consuming,
the disorder. Clinical OCD is rather common, with a lifetime prevalence distressing and detrimental to daily life, leading to impairments in
of around 2% (Mayerovitch et al., 2003; Ruscio et al., 2010), presenting global Quality of Life (QoL) outcomes (Coluccia et al., 2016), thereby
an often chronic and debilitating progression. Patients experience un differing from subclinical presentations. Finally, OCD is often comorbid
wanted obsessions, compulsions or both – respectively defined as: with other psychiatric disorders, such as Generalised Anxiety Disorder
persistent thoughts, urges or impulses that cause anxiety or distress and and Major Depressive Disorder (MDD) (Brakoulias et al., 2017; Nestadt
repetitive behaviours performed in response to an obsession or rigid et al., 2001).
rules (American Psychiatric Association, 2013). Avoidance, the physical Once thought to be intractable, with available psychoanalytic ther
evasion of triggering stimuli, also considerably impacts sufferers (McKay apies generally ineffective, OCD is now considered to be manageable
* Corresponding author.
E-mail address: [email protected] (C. Ferrando).
https://doi.org/10.1016/j.jocrd.2021.100684
Received 20 February 2021; Received in revised form 30 August 2021; Accepted 31 August 2021
Available online 8 September 2021
2211-3649/© 2021 Elsevier Inc. All rights reserved.
C. Ferrando and C. Selai Journal of Obsessive-Compulsive and Related Disorders 31 (2021) 100684
thanks to a multitude of psychotherapeutic and pharmacological treat model, which is the most extensively supported (Abramowitz et al.,
ment alternatives (Abramowitz, 1998; Foa, 2010; Foa & McLean, 2016). 2009). Dysfunctional cognitive beliefs lead to salience of intrusive
Cognitive-Behavioural Therapy (CBT) based on Exposure with Response thoughts thereby causing anxiety and distress which become condi
Prevention (ERP) is widely considered the go-to treatment with the most tioned. In seeking to reduce the discomfort via ritualising or avoiding
empirical support (Abramowitz, 2006b; Abramowitz & Arch, 2014), triggers, these coping mechanisms become negatively reinforced and
capable of significantly reducing obsessive, compulsive, depressive and prevent the opportunity to learn that the stimuli is neutral. Further, both
anxious symptoms (Abramowitz, 1996; McKay et al., 2015; Olatunji obsessions and compulsions act as triggers, perpetuating the broken
et al., 2013; Öst et al., 2015). Generally, ERP-based psychotherapy is system.
more or equally as effective as Selective Serotonergic Reuptake In
hibitors and clomipramine, the pharmacotherapies of choice for OCD 1.3. ERP in practice
(Kobak et al., 1998; Skapinakis et al., 2016; Sousa et al., 2006). A
combination of the two therapies, especially for severe cases, is also a ERP therapy developed from Meyer’s (1966) ground-breaking
recommended course (Eddy et al., 2004; Kobak et al., 1998; Romanelli approach which incorporated previous behaviour research (Abramo
et al., 2014). Further, ERP manifests advantages in terms of relapse; 12% witz, 2006a) based on prolonged exposure to distressing stimuli to
compared with clomipramine 45–89% (Abramowitz et al., 2009; modify “patients expectations”. Today, therapy is guided by a widely
Simpson et al., 2004). Other types of psychotherapy, such as Cognitive used and comprehensive treatment manual (Foa et al., 2012), developed
Therapy (CT) also lead to treatment gains (Fisher et al., 2020). Com via evidence-based research supporting necessity of: in vivo and imaginal
parisons of psychological therapies, though, have provided mixed re exercises (Foa et al., 1980; Foa & Goldstein, 1978); combined exposure
sults: some studies finding that ERP is more effective (Abramowitz et al., with response prevention (Foa et al., 1984); daily up to weekly sessions
2002; Fisher & Wells, 2005; Ponniah et al., 2013) and others that it is as (Abramowitz, Foa, & Franklin, 2003; Foa et al., 2012); and therapist
effective (Gava et al., 2007; Rosaalcazar et al., 2008; Skapinakis et al., guidance (Abramowitz, 1996; Tolin et al., 2007).
2016) as other interventions. Moreover, these findings are additionally Together, the therapist and patient construct a hierarchy of in vivo
ambiguous as ERP and CT delivery in the clinic greatly overlap and the and imaginal exposure exercises designed to evoke patient-specific
differences between them are often undiscernible (Abramowitz, 1998; anxiety and distress. The former involves physical interaction with
Gava et al., 2007; McMillan & Lee, 2010). ‘fearful’ stimuli (leaving objects in the ‘wrong’ place or touching
Despite the breadth and efficacy data of treatment options, treatment something ‘contaminated’). Imaginal exposure consists of mental
remains a pertinent issue for patients with OCD and clinicians alike. For engagement with situations that are difficult to be exposed to in real life.
the former, concerns including stigma, embarrassment and loss of con Exposures are initially conducted in-session and subsequently assigned
trol often lead to treatment deferral (Newth & Rachman, 2001) – many as homework. The therapist may also perform home visits. Following
avoid seeking treatment altogether (Mayerovitch et al., 2003; Torres exposures, patients are encouraged to refrain from acting upon the
et al., 2007), or delay between 9 and 17 years (García-Soriano et al., compulsions that succeed the exercises, known as ritual prevention. A
2014; Ruscio et al., 2010). Complete remission and treatment adherence discussion then ensues regarding the feelings experienced during
are also concerns; patient relapse or treatment discontinuation rates exposure, providing an opportunity for the patient to learn interactively.
range between 12% and 50% (Abramowitz & Arch, 2014; Simpson et al., Due to the highly distressing nature of the therapy, prior to its
2004). Continuation of research to maximise therapeutic potential is commencement the patient is informed about the purpose of the exer
necessary. To do so, we must turn to the pathological models underlying cises during a ‘psychoeduction’ session, instrumental in maintaining the
the disorder, and subsequent examination of treatment delivery. patient’s motivation (Abramowitz, 2006a). Relapse prevention is also
tackled once the hierarchy has been completed. This helps patients fully
1.2. Pathological models of OCD understand their progress, make a list of learnt strategies and prepare
them for the possibility of relapse. Overall, the manual envisions be
The functional relationship between obsessions and compulsions can tween 17 and 20 sessions lasting from 90 to 120 min, carried out weekly
be elucidated from a behavioural learning perspective. Originating from up to daily.
Mowrer’s two-factor theory of anxiety disorders (Mowrer, 1956, 1960), The mechanisms of change include a behavioural aspect – extinction
it posits that classical conditioning evokes acquisition of fear towards a learning of the conditioned fear response – combined with a cognitive
previously neutral stimulus. The fear association is subsequently main element – where patients dysfunctional beliefs are disconfirmed – and a
tained by operant conditioning via avoidance or performance of com self-efficacy factor – where patients master their fears by not relying on
pulsions. Insofar as compulsions are performed to momentarily reduce avoidance or rituals (Abramowitz, 2006a). Yet there are competing
fear and anxiety brought by obsessions, they become negatively rein theories as to how this occurs. Emotional Processing Theory (EPT) (Foa
forced and repeated, creating a vicious cycle and preventing extinction & Kozak, 1986; Foa & McNally, 1996), posits that during ERP a
learning of the conditioned fearful response (Abramowitz, 2006a). competing, more easily retrievable, non-pathological fear structure is
Cognitive theories illustrate the mechanisms behind the attribution created. The theory emphasizes that within-session and between-session
of fear and anxiety. Foa and Kozak (1985) first suggested that erroneous habituation of the fear response, or decreased anxiety, is a marker of
cognitions were driving OCD behaviour. Salkovskis (1985) proposes success. Critiques of EPT highlight that habituation is not always a pre
that, as a result, obsessions become problematic due to their role as dictor of treatment outcome or learning (Craske et al., 2008; Kircanski
stimuli themselves in inciting negative automatic thoughts, or in et al., 2012), and could encourage patients to think anxiety is a negative
trusions, idiosyncratic to the individual’s personal belief system. Thus, state rather than a normal behaviour (Jacoby & Abramowitz, 2016).
intrusive thoughts become pathological by virtue of the salience These authors instead propose an inhibitory learning approach (Abra
attributed to them through erroneous cognitions. These include: 1) mowitz & Arch, 2014), whereby fear tolerance and the acceptance of
thinking something is the same as acting upon it; 2) failing to prevent intrusive thoughts is favoured over habituation. This model emphasizes
harm is the same as causing harm; 3) inflated responsibility; 4) failure to the role of expectancy violation, focused on disproving immediate ex
perform compulsions is the same as wanting harm to occur; and 5) pectations around experiencing obsessional thoughts, uncertainty about
needing to control one’s thoughts. Finally, research shows that certain fears, anxiety around stimulus confrontation and the ability to tolerate
beliefs relate to specific symptoms, such as ‘perfectionism’ with check negative states associated with exposure at a higher intensity and length
ing and ‘needing to control ones thoughts’ with washing (Taylor et al., (Jacoby & Abramowitz, 2016). Exposure exercises are designed to
2010). violate said expectations, although cognitive restructuring still occurs
The two theories come together to form the cognitive-behavioural indirectly as a consequence (Abramowitz et al., 2005a), which
2
highlights the challenge in separating cognitive elements from ERP We included studies involving patients with comorbidities, who have
interventions. been shown to respond equally well to treatment, in order to be repre
sentative of the OCD population (Brakoulias et al., 2017; Franklin et al.,
1.4. Rationale 2000).
Next, we conducted a pre-treatment to post-treatment efficacy
We have elucidated the pathological models that underly OCD and comparison, as we wanted to compare within-subjects symptom
how ERP tackles the behavioural relationship between obsessions and reduction depending on group allocation. This design provides an
compulsions to alleviate symptoms. We have also seen how OCD is advantage to others, such as post-test only, where evaluation occurs only
underscored by a substantial cognitive component, which upholds the after the treatment (Morris, 2008).
comprehensive pathological model of the disorder, the adoption of ERP- We also wanted to address the lack of clarity regarding the inclusion
based CBT treatments or pure CT. The significant overlap in CBT and CT of cognitive elements in the ERP procedure (Abramowitz, 1998; Abra
and the often obscured integration of cognitive elements in what is re mowitz et al., 2002; Abramowitz and Houts, 2005; McMillan & Lee,
ported as ERP, has led to confusion regarding treatment protocol in the 2010). This could lead to differences in the way that therapists between
review literature on psychotherapy for OCD (Abramowitz, 1998; Gava trials deliver the therapy, potentially generating varying treatment ef
et al., 2007; McMillan & Lee, 2010). Indeed, cognitive techniques tar fects. Hence, we conducted a further analysis to investigate whether the
geting dysfunctional beliefs are regularly integrated into ERP, which is explicit adjunct of cognitive elements would have a moderating effect on
often called just CBT (Öst et al., 2015). Previous meta-analyses treatment effectiveness. Other moderators of interest that could influ
comparing such treatments, as discussed above, have shown mixed re ence practice and efficiency of treatment were hours of therapy and
sults and have not compared specific interventions with pharmaco duration of OCD. The former served to demonstrate whether fewer hours
therapies but have compared all psychological, all pharmacological or are as effective as a higher amount, which is significant for clinical
both therapies separately (Fisher & Wells, 2005; Gava et al., 2007; practice. Moreover, previous reviews have reported conflicting results
Rosaalcazar et al., 2008; Skapinakis et al., 2016). Further, the issue of regarding the role of this moderator (Abramowitz et al., 2002; Rosaal
overlapping treatments during comparison, instigating further research cazar et al., 2008). Next, assessment of treatment effectiveness
into specific intervention effects, has been called upon (Knopp et al., depending on the number of years spent with the disorder is another
2013; Ponniah et al., 2013). This was one of the primary motivations for noteworthy investigation due to patient’s lengthy treatment deferral.
the current investigation, prompting an updated literature review and Furthermore, the relationship between duration of OCD and treatment
meta-analysis of ERP treatment with very strict criteria. This would outcome is very complex (Dell’Osso et al., 2009), with some studies
ensure a highly targeted investigation to elucidate effectiveness of a having found it predicts outcome (Eisen et al., 2013), and others that it
specified treatment. does not (Steketee & Shapiro, 1995), thus warranting further research.
We also sought to build on previous meta-analyses via various Our review and meta-analysis consisted of an extensive literature
means. Firstly, we would only consider studies that followed the ERP search, after which our pre-defined criteria were applied to identify
protocol outlined above. Protocols assessing ERP as a booster to another studies for inclusion. We deemed twenty-four studies fit to be explored
therapy or ERP with added elements such as family-based or in the meta-analysis, where our main investigation compared the
mindfulness-based ERP, were all excluded. Further, we would not reduction between pre-treatment and post-treatment in YBOCS scores
consider studies assessing self-help ERP, internet-based ERP or intensive for ERP treatment, versus absence of intervention or other psychother
ERP protocols such as the Bergen-4-Day-Treatment. It was important to apeutic or pharmacological therapies.
include studies that maintained sufficient therapist-patient contact since
research has been inconclusive in determining the number of hours 2. Methods
needed for treatment effect and, importantly, the therapist holds an
essential role during treatment, particularly in maintaining motivation 2.1. Protocol
(Pearcy et al., 2016; Tolin et al., 2007). Previous reviews have included
studies with a variety of ERP protocols and included ERP designs with “Preferred Reporting Items for Systematic Reviews and Meta-Ana
limited therapist contact or exposure as homework only in one analysis lyses” (PRISMA) protocol was adopted for methodology and reporting
(Fisher et al., 2020; Ponniah et al., 2013). (Liberati et al., 2009; Moher et al., 2009). Specific objectives and
We focused solely on studies adopting the Yale-Brown Obsessive- methods for this research were delineated prior to its commencement.
Compulsive Scale (YBOCS) as the main outcome measure; it is widely See Fig. 1 for the PRISMA Flow diagram.
considered to be the ‘gold-standard’ for measuring OCD symptoms
(Abramowitz, 2006a) and has excellent validity (Kim et al., 1990; 2.2. Inclusion and exclusion criteria
Woody et al., 1995). Further, various meta-analyses on the topic have
combined multiple outcome measures (Abramowitz, 1996; Eddy et al., Exclusively Randomized Controlled Trials (RCT) were included, with
2004; Gava et al., 2007; Rosaalcazar et al., 2008), which could lead to the following additional criteria for inclusion. Studies accepting subjects
standardization bias (Morris & Deshon, 2002; Öst et al., 2015). In view over 18 years of age and with a primary diagnosis of OCD as defined in
of the YBOC’s strong reliability and validity and the statistical discrep either the International Classification of Disorders (ICD) (World Health
ancies that arise from combining outcome measures, we adopted the Organization, 2018) or the Diagnostic and Statistical Manual of Mental
YBOCS as the single outcome measure. Disorders (DSM) (American Psychiatric Association, 2013). Studies that
We included self-reported YBOCS, which has been shown to be as included patients with comorbid disorders were also considered as long
reliable and consistent as the clinician-reported scale (Federici et al., as the primary diagnosis was OCD. Trials assessing ERP as the main
2010), while others excluded (Öst et al., 2015). intervention, compared with placebo, waiting-list or other active in
We only included studies conducted with patients over 18 years-of- terventions with pre-treatment and post-treatment YBOCS scores as the
age, whereas other reviews had mixed adults and children trials (Gava primary outcome (Castro-Rodrigues et al., 2018). Either interview or
et al., 2007; Kobak et al., 1998; Olatunji et al., 2013; Romanelli et al., self-report YBOCS scales were accepted as both are comparatively reli
2014). Adult (late-onset) and paediatric (early-onset) OCD have been able and consistent (Federici et al., 2010). Both individual or group
found to differ significantly and to represent distinct etiological therapy studies were included as per their demonstrated equivalence
sub-types of the disorder, thus inclusion of both could moderate treat (Pozza & Dèttore, 2017; Öst et al., 2015). Treatment had to include both
ment evaluation results (Geller et al., 1998; Mancebo et al., 2008; exposure and response prevention elements (Foa et al., 1980). Finally,
Taylor, 2011). remote ERP interventions were considered provided there was ongoing
3
Fig. 1. Adapted PRISMA Flow Diagram showing the procedure followed for study selection into meta-analysis according to PRISMA guidelines (Moher et al., 2009).
24 studies met all the criteria and were included in the final meta-analysis.
patient-therapist contact equivalent to a non-internet-based therapy, 2.3. Information sources

and all essential steps of ERP were covered.
Criteria for exclusion were as follows. Observational, case studies or The following electronic databases were used for the study search:
reports, case control and cohort or studies assessing children or ado PubMed, Google Scholar and Science Direct. An additional study search
lescents were not considered. Trials lacking a detailed account of was conducted by manually examining individual studies and past meta-
treatment protocol or employing limited treatment therapist coaching or analyses and systematic reviews on the topic. All articles available in
self-controlled exposure were also excluded. ERP assessed in conjunc English were included.
tion or as a booster to another therapy or with any additional elements,
were considered to have invalid control groups. We could not include
studies that compared ERP alone versus ERP with added elements as 2.4. Search
then any statistical effect would be due to the added elements and not
ERP. Lastly, we excluded studies lacking sufficient data such as lack of An initial electronic search was conducted using a defined and
individual pre-treatment and post-treatment YBOCS scores and Stan limited search in June 2019. The search was limited to studies published
dard Deviations (SD), as per our statistical design. since 1971, when the first RCT on ERP therapy occurred (Rachman et al.,
There were no limitations as to number of participants, population, 1971). An advanced PubMed search was carried out, using the keywords
severity of OCD symptoms or OCD domain (sub-type). *OCD*, *ERP* and *randomized* and excluding the keyword *paedi
atric*, resulting in 42 articles. A Science Direct search was conducted
using the keywords *OCD*, *ERP*, *randomized*, resulting in 739 ar
ticles. Finally, an advanced Google Scholar search was conducted using
4
the keywords *OCD*, *ERP* and *RCT* and excluding the keyword post-treatment scores for all groups. The variable specifying the raw
*paediatric*, resulting in 990 articles. The total number of articles ac correlation coefficients between pre-test and post-test scores (ri) was not
quired from the database search was 1771. An additional manual obtainable. As per the guidelines (The Metafor Package, 2017), we
investigation provided a further 121 articles. An updated search was searched for and selected estimated values based on known properties
carried out in January 2021, following the same procedure but limited and long-term test-retest reliability of the YBOCS. 0.61 was thus selected
to the years 2019–2021. The PubMed, Science Direct and Google Scholar as a representative estimate for the raw correlation coefficient variable
searches resulted in 11, 163 and 231 articles respectively and 3 articles (Woody et al., 1995). Next, we conducted a sensitivity analysis to ensure
were found manually, totalling 408. that altering the correlation would not affect the results of the
meta-analysis. Hence one value in each direction was selected: 0.71 and
2.5. Study selection 0.51, and the meta-analysis carried out for each. Next, the difference in
the standardized mean changes between the two groups was calculated
The 1892 articles found during the initial search, and the 408 articles for each study. Finally, the main meta-analysis was conducted, to ach
found in the updated search, were scanned for duplicates, leaving 2226 ieve an effect size of the reduction in YBOCS depending on allocation.
articles to be screened for basic acceptability. The titles and abstracts We then assessed the influence of the selected moderator variables:
were inspected for eligibility criteria and a total of 144 articles were addition of cognitive elements, number of hours of therapy and years of
subsequently read in full and evaluated for adherence to inclusion and OCD duration. Variables were separately evaluated as moderators to
exclusion criteria. 120 articles were excluded for the following reasons: symptom reduction within a mixed-effects model meta-analysis
not RCT, invalid control group, not on ERP, invalid ERP protocol, (Viechtbauer, 2010). First, the standardised mean change between
insufficient therapist coaching, no YBOCS, follow-up, not OCD, included pre-treatment and post-treatment for the treatment group was calcu
under 18’s, missing data, not original data. 24 studies met all of the lated, using the method previously outlined (Morris, 2008). Next, the
criteria and were included in the final analysis. cognitive elements moderator, a categorical variable, was coded
dichotomously (0 if it included cognitive elements or 1, if it did not). The
2.6. Data collection other two variables, hours of therapy and years of OCD duration, were
coded as continuous.
Data was manually extracted from the published articles. Where Finally, a heterogeneity of variance assessment was made. For the
possible, data from treatment completers was used; alternatively, in preliminary and main meta-analysis the Q-statistic was used. This value
formation from the intent-to-treat population was entered. See Table 1 indicates the similarity in outcome between studies beyond chance, or
for basic patient demographics and study information. the variability in intervention effects, which should not be significant
Where data was not available, authors were contacted to request its (Higgins & Green, 2011). For the moderator variables, we used the I^2
provision, particularly with regards to the inclusion of cognitive ele statistic to represent the residual heterogeneity contributing to the un
ments in the ERP procedure. Where no information was attainable, Not accounted variability, and the H^2 statistic to indicate the ratio of un
Applicable (NA) was noted. In two cases (Gomes et al., 2016; Visser accounted variability to sampling variability.
et al., 2015) SDs were extrapolated from other data following Cochrane
guidance for systematic reviews and meta-analyses (Higgins & Green, 3. Results
2011). See Appendix 1 for the data extrapolation process.
3.1. Risk of bias assessment
2.7. Risk of bias
Our risk of bias report is summarised in Table 2. As per our criteria,
Following PRISMA guidelines a risk of bias evaluation was con all studies randomized participant allocation ensuring a low risk of se
ducted, aimed at highlighting possible methodological and clinical lection bias in this regard. Only five studies reported on allocation
sources of bias (Liberati et al., 2009), ensuring an accurate assessment of concealment, with three adhering and two not. All studies were
intervention effects (Higgins & Green, 2011). We therefore followed assessing an active psychotherapeutic treatment therefore blinding of
Cochrane guidelines to assess sources of potential bias in each study, participants and personnel was not feasible. Blinding of outcome
reported in the Results (Higgins & Green, 2011). assessment was adhered to by fourteen of the twenty-four studies.
Fourteen out of twenty-four studies explained procedures for dealing
2.8. Statistical analysis with incomplete data. Other sources of bias included expectancy bias in
Simpson et al. (2004), as patients and clinicians were aware of when the
Statistical analyses were performed using Rstudio with metafor treatment sessions would come to an end; no independent assessor for
package for meta-analysis (Viechtbauer, 2010). the outcome measures (Vogel et al., 2004); lack of inter-rater reliability
Prior to the main analysis, to ensure there were no significant dif assessments in Foa et al. (2005) and Visser et al. (2015) indicating that
ferences between the groups across studies in initial OCD severity, a the outcome scores were not cross-referenced; baseline differences in
Standardized Mean Difference analysis between the pre-treatment group demographics in Fineberg et al. (2005); men-only in Khodarahimi
YBOCS scores was conducted. (2009); no protocol integrity checks in Belotto-Silva et al. (2012); and
The main analyses were conducted following pretest-posttest-control use of estimates for calculating duration of therapist contact and females
procedure (Morris, 2008). We aimed to assess the difference in inter only in Fineberg et al. (2018).
vention effect by comparing within-group changes in pre-treatment and
post-treatment scores between groups (ERP and ‘other’). This design is 3.2. Preliminary tests
advantageous in ensuring regulation of pre-existing differences as each
participant acts as their own control (see Morris (2008)). Thus, the As a preliminary test we compared pre-treatment YBOCS scores be
primary outcome of difference between treatment and ‘other’ in YBOCS tween the treatment and ‘other’ groups. There was no significant dif
change was assessed to quantify the impact of the ERP therapy. This ference in pre-treatment YBOCS scores, with an estimate of 0.03 (95%
consisted of a quantitative within-group and between-group analysis CI = [-0.09 – 0.15]; SE = 0.06; z = 0.46, p = .644). The test for het
using a linear fixed-effects model within the metafor package (The erogeneity was not significant, (Q-statistic = 27.41, p = .239), indicating
Metafor Package, 2017). First, the within-group component was calcu that variation was not attributable to studies being diverse from one-
lated; for each study, we obtained the standardized mean change values another. Thus, there was no difference between the groups on this
(yi) and sampling variances (vi) between pre-treatment and measure.
5
C. Ferrando and C. Selai
Table 1
All twenty-four included studies and relevant demographic and study information.
Study Type of Therapy N % Female Age (SD) OCD Total Cognitive YBOCS Minimum OCD sub-type Country
duration therapy element report (YBOCS) score
(SD) (hours) for inclusion
Author ERP C/OT ERP C/ ERP C/ ERP C/OT ERP ERP ERP Interview/ all all all
OT OT self-report
Lindsay et al. ERP AM 9 9 44 89 31.6 34 9 (8.7) 15 N Interview NA NA Australia

(1997) (8.9) (9.3)
Freeston et al. ERP WL 15 14 NA NA NA NA NA 40.5 Y Interview NA Obsessions only France
(1997)
O’Connor et al. CBT WL 6 6 33 50 33 41.5 10.4 (3.9) 22 Y Interview NA NA Canada
(1999) (6.7) (16)
Cottraux et al. ERP CT 32 30 66 83 34.8 36.8 11.4 (8.7) 20 NA Interview ≥16 NA France
(2001) (11.4) (9.8)
McLean et al. Group Group 32 31 NA NA NA NA NA 30 N Interview NA Washing/cleaning, Checking, Harm, Canada
(2001) ERP CT Sexual and miscellaneous (Ordering,
Hoarding, Counting, Repeating, Mental
Rituals)
Cordioli et al. Group WL 23 24 NA NA NA NA NA 24 Y Interview ≥16 NA Brazil
(2003) CBT
Simpson et al. ERP PP 18 2 50 100 33.1 33.5 13.6 (12.2) 36 NA Interview ≥16 NA USA
(2004) (10.5) (12)
Vogel et al. CBT WL 16 12 56 42 31.4 37.8 NA 24 Y Interview NA Washing/cleaning, Checking, Covert Norway
(2004) (10.4) (13.2) rituals, Hoarding
O’Connor et al. ERP IBA 12 16 NA NA NA NA NA NA N Interview NA NA Canada
(2005)
Foa et al. (2005) ERP PP 29 26 62 38 33.8 34.3 14.4 (11.5) 36 Y Interview ≥16 NA USA
6
(8.9) (11.4)
Nakatani et al. ERP AT 10 8 70 62 32.5 35.9 10 (7.1) 9 NA Interview >16 Aggression, Contamination, Sex, Japan
(2005) (11.2) (8.7) Hoarding, Religion, Symmetry,
Miscellaneous, Somatic
Fineberg et al. Group Group 24 17 75 76 37.5 41.4 NA 24 Y Interview ≥16 NA UK
(2005) CBT RT (11.1) (13)
Whittal et al. ERP CBT 29 30 52 73 34.2 35.6 11.1 (9.9) 11 N Interview NA NA Canada
Journal of Obsessive-Compulsive and Related Disorders 31 (2021) 100684

(2005) (11.3) (9.7)
Sousa et al. Group Sertraline 25 25 NA NA NA NA NA 24 Y NA NA NA Brazil
(2006) CBT
Anderson and ERP WL 17 14 65 64 32.2 34.4 11.4 (9) 12 Y Interview NA NA Australia
Rees (2007) (7.6) (10.2)
Belloch et al. ERP CT 13 16 62 60 34.2 30.2 6.8 (6.8) 20 N Interview NA Pure obsessions (Aggressive, Sexual, Spain
(2008) (13) (5.7) Moral/Religious obsessions);
Obsessions with overt compulsive rituals
(Checking, Cleaning, Superstition); Both
pure obsessions and obsessions with
overt compulsive rituals
Jaurrieta et al. ERP WL 19 19 NA NA 24.2 23.3 NA 15 Y Interview >16 Aggressive, Contamination/Cleaning, Spain
(2008) (6.7) (6.6) Doubting/Checking, Sexual, Collecting,
Religious, Ordering/Symmetry, Somatic,
Repetition, Slowness
Khodarahimi ERP WL 20 20 0 0 NA NA NA 18 NA Interview NA NA Iran
(2009)
Belotto-Silva Group Fluoxetine 70 88 56 55 33.9 34.1 NA 24 Y Interview ≥16 NA Brazil
et al. (2012) ERP (11.1) (10.6)
Visser et al. CBT IBA 47 43 60 72 33.7 35.9 NA 18 Y Interview ≥16 NA Netherlands
(2015) (8.5) (10.6)
(continued on next page)
3.3. Main analysis
C/OT=Control/Other Therapy; AM = Anxiety Management; WL=Waiting-List; CT=Cognitive Treatment; PP=Pill Placebo; IBA=Inference-Based Approach; AT = Autogenic Training; RT = Relaxation Training; TAU =
For the main meta-analysis, we conducted a sensitivity exploration
England
Country
Brazil
by altering the correlation variable (ri). We conducted the analysis using
UK
UK
all
0.51 (estimate = − 0.76; SE = 0.09; z = − 8.50; p < .01), 0.61 and 0.71
(estimate = − 0.74; SE = 0.08; z = − 9.30; p < .01), and consistently
achieved the same results; the sensitivity analysis was therefore satis
factory and altering this variable did not impact our findings. Since the
true values of ri for each study were not attainable, our sensitivity
exploration demonstrates that our results were not impacted and are
consistent. Here are reported the results from the test using the 0.61
correlation as was established in the literature (Woody et al., 1995). Our
meta-analysis revealed a statistically significant difference between ERP
OCD sub-type
and the ‘other’ groups in YBOCS score reduction at pre-treatment versus

post-treatment. This was indicated by a mean effect size, effect size =
− 0.75, significant at p < .01 (95% CI = [-0.92 to − 0.59]; SE = 0.09, z =
NA
NA
NA
NA
all
− 8.87). The results can be visualized in Fig. 2.

Our test for heterogeneity was significant (Q-statistic = 185.66, p <
(YBOCS) score
.01), indicating that one study could be driving the results as there is
for inclusion
Minimum
Not clear
some unexplained heterogeneity between studies. Cochrane guidelines,

used throughout this piece, report that when using a fixed-effects model
≥16
>16
NA
all
the heterogeneity result can be ignored (Higgins & Green, 2011).

Regardless, to ensure that any individual study was not driving the
– No; NA=No information; SD=Standard Deviation.
heterogeneity, a leave1out test was conducted. This computation runs

Interview/
self-report
Interview
Interview
Interview
the main meta-analysis excluding each study successively, therefore

YBOCS
report
Self
assessing whether the results remain significant each time, eventually

confirming the reliability of the main analysis (Viechtbauer, 2010). The
p-values were significant for each test, therefore our main meta-analysis
Cognitive
was not affected by the heterogeneity between the studies. Additionally,

element
we ran influential case diagnostics on the studies to identify any po

ERP
tential influential cases, following guidelines used throughout the sta

tistical analysis (Viechtbauer, 2010). We used the influence function to
generate diagnostic plots (see Appendix 2). This showed that three
therapy
(hours)
Total
studies (19, 20 and 21) (Belotto-Silva et al., 2012; Visser et al., 2015 and
ERP
24
12
16
16
Gomes et al., 2016 respectively) were outliers and potential drivers of

the results. This can be deduced from their influence on the fit of the
30 (13.5)
duration
model (Cook.d and hat plot) and large residuals (rstudent plot), whereby
their removal would reduce the amount of residual heterogeneity (Tau2.
OCD
(SD)
ERP
NA
NA
NA
Treatment-As-Usual; EMDR = Eye Movement Desensitisation and Reprocessing; Y=Yes; N–
del). Thus, to test the robustness of our findings, we ran the

meta-analysis without the outliers. Our original findings were unaf
(13.3)
C/OT
(9.8)
(NA)
fected by the removal of these studies (effect size = − 0.97, significant at

37.1
32.7
30.9
NA
p < .01 (95% CI = [-1.18 to − 0.76]; SE = 0.11, z = − 8.94)). This,

Age (SD)
combined with the leave1out test, demonstrates that our original results
(14.3)
(15.4)
(NA)
can be considered valid.

44.5
32.4
33.3
ERP
NA
3.4. Moderator analyses

100
OT
67
47
59
C/
% Female
The first moderator analysis examined whether the addition of

ERP
100
58
62
65
cognitive elements within treatment would affect the significant

reduction in pre-treatment to post-treatment YBOCS scores after ERP
OT
46
17
15
29
C/
therapy. The test of moderators was significant (QM = 160.04; p <

.001). Although, both treatment with (estimate = − 1.74; 95% CI =
ERP
[-2.13 to − 1.35]; SE = 0.20; z= − 8.68, p < .001) and without (estimate

52
17
16
26
N
= − 2.87; 95% CI = [-3.48 to − 2.26]; SE = 0.31; z = − 9.20, p < .001)

cognitive elements lead to a significant reduction in YBOCS scores at
Sertraline
pre-treatment versus post-treatment. This suggests both treatment var

EMDR
C/OT
Type of Therapy
TAU
WL
iations significantly reduce OCD symptoms. The test of heterogeneity

was significant (QE = 113.24; p < .001). The residual heterogeneity that
was contributing to the unaccounted variability was, I^2 = 82%, and the
Group
CBT
CBT
CBT
ERP
ERP
ratio of unaccounted variability to sampling variability was, H^2 = 5.54.

Table 1 (continued )
These results can be disregarded as we had already conducted additional

heterogeneity tests for the main meta-analysis and the results remained
et al. (2017)
Fineberg et al.
Marsden et al.
Challacombe
Gomes et al.
unaffected.
(2016)
(2018)
(2018)
Next, we assessed whether number of hours of therapy affected the

Author
Study
reduction in YBOCS pre-treatment to post-treatment. The moderator

effect of hours of therapy was not significant, QM = 1.46, at p = .227,
7
Table 2
Cochrane risk of bias assessment tool. Each study has been quality assessed for the 6 potential sources of bias and
given a code and colour equivalent to the degree of bias. Green sections with ‘L’ indicate the study has scored a
low risk of bias; red sections with ‘H’ indicate the study has a high risk of bias in this area; and yellow sections
with ‘U’ indicate the study has an uncertain risk of bias on this measure.
indicating that number of hours did not predict treatment outcome. The analysis. Although the moderator effect was not significant (estimate =
output confirmed that there was a significant difference between pre- − 0.03; 95% CI = [-0.08 – 0.02]; SE = 0.03; z = − 1.21; p = .227),
treatment and post-treatment in general, indicated by the intercept indicating that number of hours did not alter the reduction in YBOCS
being significant, (estimate = − 1.49; 95% CI = [-2.57 to − 0.42]; SE = between pre-treatment and post-treatment. Again, the test of heteroge
0.55; z = − 2.72, p < .05), thus confirming the results of the main meta- neity was significant (QE = 128.68; p < .001; I^2 = 85%; H^2 = 6.87),
8
Fig. 2. Forest plot displaying the reduction in YBOCS scores from pre-treatment to post-treatment between ERP therapy and ‘other’ groups for each study included in
the meta-analysis. Each row represents a different study, with the authors and year on the left-side column; on the right are the effect size (ES) and CIs; in the middle
of the plot, across from each study, is the standardized mean change, with the points on the negative side of zero on the bottom key indicating a significant study
result and on the positive side of zero a non-significant study result. Each study also has a different size point, which represents the relative weighting to the overall
result. The final row before the bottom key indicates the over fixed-effects model result as stated above.
although we can discount it once again. 4. Discussion

The final moderator analysis assessed the impact of years of OCD
duration on pre-treatment to post-treatment score change. The moder 4.1. Results summary
ator was not significant, (QM = 0.11; p = .738), indicating that years
living with the disorder did not affect pre-treatment to post-treatment Our main aim was to determine whether ERP-based therapy was
score. Accordingly, the effect of the duration variable was non- more effective in reducing OCD symptoms compared to no treatment or
significant (estimate = 0.02; 95% CI = [-0.08 – 0.12]; SE = 0.05; z = other psychotherapeutic or pharmaceutical interventions. Our strict
0.33, p = .738). Again, we obtained significant heterogeneity (QE = study selection with particular attention to treatment protocol allowed
38.76; p < .001; I^2 = 79%; H^2 = 4.76), although we can discount it. us to determine results with a high degree of accuracy and specificity. In
such manner preventing other confounding variables, both in treatment
– such as self-controlled exposure, mindfulness-based CBT or self-help
programs – and in-review methodology – such as concomitant
9
inclusion of adults and children, from influencing the results. By 6 months and Simpson et al. (2004), Fineberg et al. (2005), Belloch et al.
building upon past meta-analyses and a multitude of trials, we attemp (2008) and Visser et al. (2015) reported no effect size comparisons at
ted to highlight valuable insights for clinicians and their patients. follow-up. Overall, from the effect sizes available, two studies report the
Ultimately, we included twenty-four RCTs carried out in 11 coun superiority of ERP over ‘other’ groups at follow-up, two studies found no
tries, published between 1997 and 2018 and concerning a total of 1134 difference between ERP and ‘other groups’ at follow-up and one study
patients in our analysis. The efficacy of ERP therapy in reducing OCD found the ‘other’ group (Sertraline) was superior to ERP at follow-up.
symptoms was compared to a control group, either waiting-list or pla From this wide range of findings, it is not possible to conclude
cebo, or another therapy, which included anxiety management, cogni whether the gains seen post-treatment in the ERP groups over ‘other’
tive therapy, IBA, autogenic training, relaxation therapy, fluoxetine, groups, as found in this meta-analysis, were maintained at follow-up.
sertraline and eye-movement desensitisation and reprocessing. Our
meta-analysis showed that there was a greater reduction in pre- 4.2. Implications
treatment to post-treatment OCD symptoms, as measured via the
YBOCS, for patients that received ERP than for any of the ‘other’ groups. According to these findings, ERP should be considered the treatment
Thus, ERP was more efficacious than no treatment or other available of choice for OCD. As treatment-seeking is an issue for patients, ongoing
alternatives included in the meta-analysis. Additionally, we can confirm research into the success of treatment is essential in order to first, pro
that our results were not impacted by pre-existing differences in OCD vide a safe environment where patients feel confident of the opportunity
severity, measured in the pre-treatment scores. This finding coincides for improvement and second, to increase awareness. Hopefully,
with past investigations (Abramowitz, 1996; Fisher & Wells, 2005; continuing to provide consistent results and building on current
McKay et al., 2015; Olatunji et al., 2013; Ponniah et al., 2013; Öst et al., knowledge will encourage more people suffering to come forward and
2015), and thus new trials have added to the pool of significant ERP seek treatment. It is also noteworthy that no studies, but one (Jaurrieta
treatment outcome for OCD. et al., 2008), had an upper limit YBOCS score for inclusion, indicating
We subsequently performed additional analyses, to investigate the that even severe cases were included in the trials. Further augmenting
role of specific moderator variables as predictors of ERP success. As has the generalisability of our findings, it is notable that the included studies
been discussed throughout, whether the ‘standard’ ERP procedure were carried out across 11 countries globally, all contributing to the pool
contains cognitive elements is unclear (Abramowitz and Houts, 2005), of our significant result.
and this could lead to differences in the effectiveness and patient Our moderator analyses also have valuable implications. Even
response to ERP. One of our aims was to clarify its effect. Treatments though ERP with cognitive elements did not surpass ‘standard’ ERP, its
both with and without cognitive elements resulted in statistically sig efficacy is notable regardless. Our results could also suggest that benefit
nificant changes in pre-treatment to post-treatment YBOCS scores. This from the addition of cognitive elements is dependent on the individual.
was unexpected due to the ample support for the cognitive-behavioural It has been shown that dysfunctional beliefs play a particularly salient
theory of OCD (Abramowitz et al., 2009), and the prominence given to role in certain sub-types of OCD (Abramowitz et al., 2006) and therefore
the role of dysfunctional beliefs as drivers to pathology (Abramowitz cognitive adjuncts would not aid those with less prominent dysfunc
et al., 2006; Salkovskis, 1985), suggesting that additional cognitive tional beliefs. Additionally, attempting to disaggregate cognitive tech
components would lead to enhanced treatment results. This finding niques completely from ERP interventions may prove difficult,
should be considered with caution as, although we only allocated studies considering discussions around dysfunctional beliefs take place (Abra
to the cognitive elements group if they had explicitly stated this was part mowitz et al., 2005a) and therapists implement additions to varying
of their procedure, it was not possible to identify how often and to which degrees during ERP delivery (Jacoby & Abramowitz, 2016).
patients it was delivered. Nonetheless, treatment with cognitive ele Next, we demonstrated that hours of therapy did not affect pre-
ments resulted in a significant change between pre-treatment to treatment to post-treatment change. The hours of therapy offered in
post-treatment, suggesting patients benefitted from it. Further, efficacy the included studies ranged considerably: from nine to forty and a half
of Behavioural Therapy alone and CBT is also reported elsewhere hours. This is potentially enormously consequential, as it could suggest
(Skapinakis et al., 2016; Öst et al., 2015). that measurable reductions in OCD symptomatology are achievable with
Our next moderators, hours of therapy and years of OCD duration, fewer hours, thus putting less burden on public health services or self-
were non-significant, meaning that change in YBOCS scores between funded patients and offering patients faster relief. Although, this
pre-treatment and post-treatment were unaltered. should be considered with caution. Last, we found that years of OCD
Finally, a consideration on follow-up results. Fourteen out of the duration also did not affect pre-treatment to post-treatment change; this
twenty-four included studies did not have follow-up data available for should be highly reassuring for patients. Years spent with OCD in the
analysis: seven did not conduct follow-up assessments by design, one studies within our analysis ranged between seven and thirty years, hence
reported YBOCS scores at follow-up as a cumulative score for the whole our results suggest that ERP can lead to improvements for a whole range
sample, four studies reported YBOCS scores at follow-up only for the of patients.
active treatment groups (for these, the control group had been used as
the comparison group in the present analysis) and two studies did not 4.3. Limitations
report YBOCS scores at follow-up. Ten studies remained for examina
tion, a significantly lower number than those included in the main Our study is not without drawbacks, and these should be taken into
analysis and thus insufficient to supplement the pre-to post-analyses consideration. The first limitation is regarding treatment protocol am
conducted with further insights from an analysis of the limited follow-up biguity of adjunct cognitive elements. In the literature this is also
data available. The follow-up results from the ten studies are hereby evident: some use ERP interchangeably with CBT, thereby including the
reported descriptively. Cottraux et al. (2001) report an effect size (d = cognitive aspect (McKay et al., 2015; Öst et al., 2015), and the indis
− 0.44) in favour of BT over CT at week 52; McLean et al. (2001) report tinctness around this is well recognised (Abramowitz and Houts, 2005).
an effect size (η2 = 0.11) in favour of ERP over CT at 3 months; Whittal Although we performed a moderator analysis specifically to investigate
et al. (2005) report an effect size (η2 = 0.018) with no significant dif this and paid attention to carefully selecting studies with clear protocols,
ference between ERP and CBT at 3 months; Khodarahimi et al. (2009) we cannot be certain about what occurred in every therapy session.
only report effect sizes for analyses conducted on all three treatment Second, we conducted a risk of bias assessment, and the most critical
groups; Fineberg et al. (2018) report an effect size (d = 0.56) in favour of criteria were met by at least half of the studies (random sequence gen
Sertraline over CBT at week 52; Marsden et al. (2018) report an effect eration, blinding of outcome data and dealing with incomplete outcome
size (d = − 0.03) with no significant difference between ERP and CBT at data). As discussed, allocation concealment and blinding of participants
10
and personnel was not strictly possible as studies were assessing psy investigating specific elements and how they can impact treatment ef
chotherapeutic interventions. A source of potential bias was possibility fect. These could be based on variables that have been related to suc
of selective reporting. Cochrane guidelines indicate this could lead to cessful treatment in the literature, which include: in vivo and imaginal
‘within-study publication bias’ (Higgins & Green, 2011). Nonetheless, exposure, psychoeducation and relapse prevention (Abramowitz &
due to the strength of the literature supporting ERP, we can assume that Arch, 2014; Abramowitz, Franklin, & Cahill, 2003; Foa et al., 1980; Foa
our results are representing a true treatment effect. & McLean, 2016). Other variables of interest could include patient prior
Another limitation was that many studies had small sample sizes, and treatment, session frequency, and initial depression and initial QoL
further, not all conducted a power estimation. This calculation ensures scores (Franklin et al., 2000). Additionally, the treatment effect of
that enough participants are selected in a clinical trial in order to avoid additional elements such as those discussed in the rationale,
Type I and II errors that mean the true effect of the intervention is mindfulness-based or family-based ERP, should be explored further
overestimated, as with the former, or, in the latter’s case, under individually.
estimated (Jones et al., 2003). As with other limitations, due to the Another intriguing area for further research is to examine the
abundance of literature supporting our findings, we can still consider treatment effect of delivering ERP treatment within the two mechanism-
them to be reflecting a true result. of-change frameworks elucidated in the introduction. The literature
Next, we did not analyse changes in depression, anxiety and QoL seems to support an inhibitory learning approach of fear tolerance and
scores. As with our main outcome measure (YBOCS), we wanted to avoid extinction; delivering ERP within this framework could increase its ef
combining different measures in order prevent bias (Morris & Deshon, ficacy (Craske, 2015; Craske et al., 2008). Delivering ERP within an EPT
2002; Öst et al., 2015). Indeed, a variety of depression, anxiety and QoL framework which encourages fear habituation could also be of interest,
were adopted by each study, if any (see Appendix 3 for other outcome although some have suggested this method could create more oppor
scales used by the studies). We therefore opted not to investigate these tunities for relapse (Abramowitz & Arch, 2014; Jacoby & Abramowitz,
variables, regrettably because the life of an OCD patient is impacted in a 2016). Studies could directly compare these frameworks, in order to
multitude of ways and it is important to observe disease progression as decipher which leads to superior and longer-lasting symptom reduction.
such. As previously discussed, depression (Abramowitz, 2004), anxiety Another avenue to yield increasingly consistent treatment effects is
(Nutt & Malizia, 2006) and a decline in QoL (Coluccia et al., 2016) are to investigate personalised approaches to therapy. This could be
highly prominent in OCD, hence any therapy should aim to measurably implemented at an individual level, investigating a variety of cognitive
change these too. We also could not evaluate the long-term treatment styles and tailored approaches that focus on the idiosyncratic nature of
effect of the ERP intervention versus the ‘other’ groups, which would obsessions and compulsions. Additionally, there could be an increas
constitute a worthy investigation due to the rates of relapse following ingly specific approach to OCD subtypes (Abramowitz et al., 2005b).
treatment interventions and in continuing to ensure patient recovery This has begun to take place for sub-types which are now considered
and sustained QoL following treatment. This investigation was not clinically distinct from OCD such as hoarding, a subset of patients which
feasible as fourteen out of the twenty-four included studies either did not has been found to respond poorly to traditional OCD treatment (Steketee
extend measures to include follow-up assessments or did not report raw et al., 2010). Furthermore, ameliorating treatment for patients with
outcome scores for each group, a caveat to be addressed by future study prominent comorbidities should also be part of the effort in tailoring
designs. We descriptively summarised the follow-up findings of the treatment. Indeed there are successful case studies on ERP for OCD
remaining ten studies to provide some assessment of these crucial results patients with MDD (Abramowitz, 2004) and trials assessing a version of
since a separate meta-analysis could not be carried out. Out of ten ERP with incorporated therapy for MDD (Rector et al., 2009). More
studies though, five did not report adequate effect size comparisons; this studies investigating other common comorbid disorders, such as affec
further showcases the need for more consistent follow-up assessments. tive disorders (Nestadt et al., 2001) are needed.
In order to ensure optimal reports of the literature which ultimately
inform treatment decisions, future studies should aim to effectively
conduct and report on follow-up assessments, so that additional analyses 4.5. Concluding comments
may take place.
Finally, a note on the limitations of RCTs more broadly. Albeit the Once thought to be rare and intractable, OCD is now recognised as a
gold-standard design in research, criticisms of RCTs, particularly in common disorder with a positive prognosis. Yet, it is also considered to
relation to psychological intervention studies, question the general be one of the most devastating conditions with the most complex psy
isability of findings. These include: poor sampling practices, which chopathology of emotional disorders (Abramowitz, 2006b), therefore it
restrict patient inclusion to those without comorbidities or high severity is vital to keep on researching and updating the field. Although phar
scores thereby limiting the applicability to the general population; macological options are effective and less intensive than psychological
similarity between groups after randomization not being measured with interventions, relapse after discontinuation is higher (Simpson et al.,
relevant metrics (such as intelligence or cognitive flexibility) but with 2005), and almost the majority of patients do not benefit (Fineberg et al.,
standard demographic measures (such as age); various assumptions 2012). Moreover, it is noteworthy that patients themselves prefer psy
including therapy standardisation across patients, to name a few (Carey chological therapy or a combination (Patel & Simpson, 2010). Our re
& Stiles, 2015; Shean, 2014). While considering these limitations, it is sults support the evidence for the most popular psychological therapy
noteworthy that various studies evaluated here included patients with available, ERP, in that it should be offered to patients as a reliable and
comorbidities (eg. Belloch et al., 2008 or; Challacombe et al., 2017), all effective treatment. There is still progress to be made, particularly with
but one had no upper YBOCS score for inclusion and ERP generally is regards to refining elements of the therapy itself and tailoring to pa
tailored around the individual thereby ensuring a more robust repre tient’s needs; we hope to have contributed to this advancement.
sentation of the general population base.
Acknowledgments
4.4. Future directions
We would like to thank Dr Matilde Vaghi, for supporting our research
A majority of the literature, including this review, supports the ef into ERP therapy for OCD, particularly for assistance on statistical
ficacy of ERP in consistently reducing OCD symptomatology; future methods and advice on the field of OCD. Our deepest gratitude to our
studies and reviews should focus on optimizing the therapy itself by family and friends.
11
Appendix 1. Additional data extrapolation process
Gomes et al. (2016) provided Standard Errors (SE) from which SD’s were calculated using Equation (1) by multiplying the SE by the square root of
the sample size (N) in each group:
√̅̅̅̅
SD = SE × N Equation 1
Visser et al. (2015) provided 95% Confidence Intervals (CI), which were used to calculate SD’s, using Equation (2).
√̅̅̅̅
N × (upper limit − lower limit)
SD = Equation 2
x
As the sample size for this study was small (<60 in each group) x represents the SE width of the 95% CI, which can be found in t distribution tables
with the appropriate degrees of freedom (N – 1) for each group. This value is then multiplied by 2 to get x. For the treatment group (ERP) this was
4.026; for the other group, Inference-Based Approach (IBA) this was 4.036. Prior to the calculation, it was ensured that the CI was symmetrical to the
mean.
Appendix 2. Plots of influential case diagnostics
Plot of the externally standardized residuals, DFFITS values, Cook’s distances, covariance ratios, leave-one-out estimates of τ2 and test statistics for
residual heterogeneity, hat values, and weights for the 24 studies included in the meta-analysis comparing ERP treatment versus ‘other’ groups.
12
Appendix 3. Scales used to measure depression, anxiety and QoL
Author Year Depression Scale Anxiety Scale QoL scale
Lindsay et al. 1997 BDI STAI Interference Rating Scale

Freeston et al. 1997 BDI BAI CFA
O’Connor et al. 1999 BDI STAI NA
Cottraux et al. 2001 BDI FQ QOL
McLean et al. 2001 BDI NA NA
Cordioli et al. 2003 17-HamD HamA WHOQOL-BREF
Simpson et al. 2004 HamD NA NA
Vogel et al. 2004 BDI STAI NA
O’Connor et al. 2005 BDI BAI NA
Foa et al. 2005 HamD NA NA
Nakatani et al. 2005 17-HamD HamA GAF
Fineberg et al. 2005 MÅDRS HAS SASS
Whittal et al. 2005 BDI NA NA
Sousa et al. 2006 BDI BAI WHOQOL-BREF
Anderson and Rees 2007 BDI NA Q-LES-Q
Belloch et al. 2008 BDI STAI-State NA
Jaurrieta et al. 2008 HamD HamA NA
Khodarahimi 2009 NA NA NA
Belotto-Silva et al. 2012 NA NA NA
Visser et al. 2015 BDI BAI EuroQoL
Gomes et al. 2016 BDI BAI NA
Challacombe et al. 2017 DASS-Depression DASS-Anxiety NA
Fineberg et al. 2018 MÅDRS NA EuroQol
Marsden et al. 2018 PHQ-9 GAD-7 WSAS
BDI= Beck Depression Inventory; HamD = Hamilton Rating Scale for Depression; MÅDRS = Montgomery-Åsberg Depression Rating Scale; DASS =
Depression Anxiety Stress Scales – Depression; PHQ-9 = Patient Health Questionnaire 9; STAI= State-Trait Anxiety Inventory; BAI= Beck Anxiety
Inventory; FQ = Mark’s Fear Questionnaire; HamA = Hamilton Rating Scale for Anxiety; HAS= Hamilton Anxiety Scale; GAD-7 = Generalised Anxiety
Disorder Assessment; CFA= Current Functioning Assessment; QOL = Quality Of Life Scale; WHOQOL-BREF= World Health Organization Quality of Life
Assessment; GAF = Global Assessment of Functioning, SASS= Social Adjustment and Self-Evaluation Scale; Q-LES-Q = Quality of Life Enjoyment and
Satisfaction Questionnaire; EuroQoL = EuroQoL scale; WSAS= Work and Social Adjustment Scale.
Statement of contribution
Cloe Ferrando: Data curation, Formal analysis, Investigation, Project administration, Writing – original draft, Writing – review and editing.
Caroline Selai: Conceptualisation, Methodology, Supervision.
Role of funding source
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Declaration of interest
None.
References Abramowitz, J. S., Foa, E. B., & Franklin, M. E. (2003). Exposure and ritual prevention for
obsessive-compulsive disorder: Effects of intensive versus twice-weekly sessions.
Journal of Consulting and Clinical Psychology, 71(2), 394–398. https://doi.org/
Abramowitz, J. S. (1996). Variants of exposure and response prevention in the treatment
10.1037/0022-006X.71.2.394
of obsessive-compulsive disorder: A meta-analysis. Behavior Therapy, 27(4),
Abramowitz, J. S., Franklin, M. E., & Cahill, S. P. (2003). Approaches to common
583–600. https://doi.org/10.1016/S0005-7894(96)80045-1
obstacles in the exposure-based treatment of obsessive-compulsive disorder.
Abramowitz, J. S. (1998). Does cognitive-behavioral therapy cure obsessive-compulsive
Cognitive and Behavioral Practice, 10(1), 14–22. https://doi.org/10.1016/S1077-
disorder? A meta-analytic evaluation of clinical significance. Behavior Therapy, 29
7229(03)80004-4
(2), 339–355. https://doi.org/10.1016/S0005-7894(98)80012-9
Abramowitz, J. S., Franklin, M. E., & Foa, E. B. (2002). Empirical status of cognitive-
Abramowitz, J. S. (2004). Treatment of obsessive-compulsive disorder in patients who
behavioral therapy for obsessive-compulsive disorder: A meta-analytic review.
have comorbid major depression. Journal of Clinical Psychology, 60(11), 1133–1141.
Romanian Journal of Cognitive and Behavioral Psychotherapies, 2(2), 89–104.
https://doi.org/10.1002/jclp.20078
Abramowitz, J. S., & Houts, A. C. (Eds.). (2005). Concepts and controversies in obsessive-
Abramowitz, J. S. (2006a). The psychological treatment of obsessive—compulsive
compulsive disorder. Boston, MA: Springer US. https://doi.org/10.1007/b101087.
disorder. Canadian Journal of Psychiatry, 51, 10. https://doi.org/10.1177/
Abramowitz, J. S., Khandker, M., Nelson, C. A., Deacon, B. J., & Rygwall, R. (2006). The
070674370605100702
role of cognitive factors in the pathogenesis of obsessive–compulsive symptoms: A
Abramowitz, J. S. (2006b). Understanding and treating obsessive-compulsive disorder: A
prospective study. Behaviour Research and Therapy, 44(9), 1361–1374. https://doi.
cognitive behavioral approach. Mahwah, NJ: Lawrence Erlbaum Associates.
org/10.1016/j.brat.2005.09.011
Abramowitz, J. S., & Arch, J. J. (2014). Strategies for improving long-term outcomes in
Abramowitz, J. S., McKay, D., & Taylor, S. (2005). Special series subtypes of obsessive-
cognitive behavioral therapy for obsessive-compulsive disorder: Insights from
compulsive disorder: Introduction. Behavior Therapy, 36(4), 367–369. https://doi.
learning theory. Cognitive and Behavioral Practice, 21(1), 20–31. https://doi.org/
org/10.1016/S0005-7894(05)80118-2
10.1016/j.cbpra.2013.06.004
Abramowitz, J., Taylor, S., & McKay, D. (2005). Potentials and limitations of cognitive
Abramowitz, J. S., Deacon, B. J., Olatunji, B. O., Wheaton, M. G., Berman, N. C.,
treatments for obsessive-compulsive disorder. Cognitive Behaviour Therapy, 34(3),
Losardo, D., Timpano, K. R., McGrath, P. B., Riemann, B. C., Adams, T.,
140–147.
Björgvinsson, T., Storch, E. A., & Hale, L. R. (2010). Assessment of obsessive-
Abramowitz, J. S., Taylor, S., & McKay, D. (2009). Obsessive-compulsive disorder. The
compulsive symptom dimensions: Development and evaluation of the dimensional
Lancet, 374, 9.
obsessive-compulsive scale. Psychological Assessment, 22(1), 180–198. https://doi.
American Psychiatric Association. (2013). In Diagnostic and statistical manual of mental
org/10.1037/a0018260
disorders (DSM-V) (5th ed.). Arlington, VA: APA.
13
Anderson, R. A., & Rees, C. S. (2007). Group versus individual cognitive-behavioural treatment: A clinical significance analysis of manualised psychological interventions.
treatment for obsessive-compulsive disorder: A controlled trial. Behaviour Research Journal of Affective Disorders, 275, 94–108. https://doi.org/10.1016/j.
and Therapy, 45(1), 123–137. https://doi.org/10.1016/j.brat.2006.01.016 jad.2020.06.019
Belloch, A., Cabedo, E., & Carrió, C. (2008). Empirically grounded clinical interventions: Fisher, P. L., & Wells, A. (2005). How effective are cognitive and behavioral treatments
Cognitive versus behaviour therapy in the individual treatment of obsessive- for obsessive–compulsive disorder? A clinical significance analysis. Behaviour
compulsive disorder: Changes in cognitions and clinically significant outcomes at Research and Therapy, 43(12), 1543–1558. https://doi.org/10.1016/j.
post-treatment and one-year follow-up. Behavioural and Cognitive Psychotherapy, 36, brat.2004.11.007
521. https://doi.org/10.1017/S1352465808004451 Foa, E. B. (2010). Cognitive behavioral therapy of obsessive-compulsive disorder.
Belotto-Silva, C., Diniz, J. B., Malavazzi, D. M., Valério, C., Fossaluza, V., Borcato, S., Dialogues in Clinical Neuroscience, 12(2), 199–207.
Seixas, A. A., Morelli, D., Miguel, E. C., & Shavitt, R. G. (2012). Group cognitive- Foa, E. B., & Goldstein, A. (1978). Continuous exposure and complete response
behavioral therapy versus selective serotonin reuptake inhibitors for obsessive- prevention in the treatment of obsessive-compulsive neurosis. Behavior Therapy, 9
compulsive disorder: A practical clinical trial. Journal of Anxiety Disorders, 26(1), (5), 821–829. https://doi.org/10.1016/S0005-7894(78)80013-6
25–31. https://doi.org/10.1016/j.janxdis.2011.08.008 Foa, E. B., & Kozak, M. J. (1985). Treatment of anxiety disorders: Implications for
Brakoulias, V., Starcevic, V., Belloch, A., Brown, C., Ferrao, Y. A., Fontenelle, L. F., psychopathology. In Anxiety and the anxiety disorders (pp. 421–452). Hillsdale, NJ,
Lochner, C., Marazziti, D., Matsunaga, H., Miguel, E. C., Reddy, Y. C. J., do US: Lawrence Erlbaum Associates, Inc.
Rosario, M. C., Shavitt, R. G., Shyam Sundar, A., Stein, D. J., Torres, A. R., & Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: Exposure to corrective
Viswasam, K. (2017). Comorbidity, age of onset and suicidality in information. Psychological Bulletin, 99(1), 20–35. https://doi.org/10.1037/0033-
obsessive–compulsive disorder (OCD): An international collaboration. Comprehensive 2909.99.1.20
Psychiatry, 76, 79–86. https://doi.org/10.1016/j.comppsych.2017.04.002 Foa, E. B., Liebowitz, M. R., Kozak, M. J., Davies, S., Campeas, R., Franklin, M. E.,
Carey, T., & Stiles, W. (2015). Some problems with randomized controlled trials and Huppert, J. D., Kjernisted, K., Rowan, V., Schmidt, A. B., Simpson, H. B., & Tu, X.
some viable alternatives. Clinical Psychology & Psychotherapy, 23(1), 87–95. (2005). Randomized, placebo-controlled trial of exposure and ritual prevention,
Castro-Rodrigues, P., Camacho, M., Almeida, S., Marinho, M., Soares, C., Barahona- clomipramine, and their combination in the treatment of obsessive-compulsive
Corrêa, J. B., & Oliveira-Maia, A. J. (2018). Criterion validity of the Yale-Brown disorder. American Journal of Psychiatry, 162(1), 151–161.
obsessive-compulsive scale second edition for diagnosis of obsessive-compulsive Foa, E. B., & McLean, C. P. (2016). The efficacy of exposure therapy for anxiety-related
disorder in adults. Frontiers in Psychiatry, 9, 431. https://doi.org/10.3389/ disorders and its underlying mechanisms: The case of OCD and PTSD. Annual Review
fpsyt.2018.00431 of Clinical Psychology, 12, 1–28. https://doi.org/10.1146/annurev-clinpsy-021815-
Challacombe, F. L., Salkovskis, P. M., Woolgar, M., Wilkinson, E. L., Read, J., & 093533
Acheson, R. (2017). A pilot randomized controlled trial of time-intensive Foa, E. B., & McNally, R. J. (1996). Mechanisms of change in exposure therapy. In Current
cognitive–behaviour therapy for postpartum obsessive–compulsive disorder: Effects controversies in the anxiety disorders (p. 984). New York: Guilford Press.
on maternal symptoms, mother–infant interactions and attachment. Psychological Foa, E. B., Steketee, G., Grayson, J. B., Turner, R. M., & Latimer, P. R. (1984). Deliberate
Medicine, 47, 1478–1488. https://doi.org/10.1017/S0033291716003573 exposure and blocking of obsessive-compulsive rituals: Immediate and long-term
Coluccia, A., Fagiolini, A., Ferretti, F., Pozza, A., Costoloni, G., Bolognesi, S., & effects. Behavior Therapy, 15(5), 450–472. https://doi.org/10.1016/S0005-7894(84)
Goracci, A. (2016). Adult obsessive–compulsive disorder and quality of life 80049-0
outcomes: A systematic review and meta-analysis. Asian Journal of Psychiatry, 22, Foa, E. B., Steketee, G., Turner, R. M., & Fischer, S. C. (1980). Effects of imaginal
41–52. https://doi.org/10.1016/j.ajp.2016.02.001 exposure to feared disasters in obsessive-compulsive checkers. Behaviour Research
Cordioli, A. V., Heldt, E., Bochi, D. B., Margis, R., de Sousa, M. B., Tonello, J. F., and Therapy, 18(5), 449–455. https://doi.org/10.1016/0005-7967(80)90010-8
Manfro, G. G., & Kapczinski, F. (2003). Cognitive-behavioral group therapy in Foa, E. B., Yadin, E., & Lichner, T. K. (2012). In Exposure and response (ritual) prevention
obsessive-compulsive disorder: A randomized clinical trial. Psychotherapy and for obsessive compulsive disorder: Therapist guide (2nd ed.). Oxford University Press.
Psychosomatics, 72(4), 211–216. https://doi.org/10.1159/000070785 Franklin, M. E., Abramowitz, J. S., Kozak, M. J., Levitt, J. T., & Foa, E. B. (2000).
Cottraux, J., Note, I., Yao, S. N., Lafont, S., Note, B., Mollard, E., Bouvard, M., Effectiveness of exposure and ritual prevention for obsessive-compulsive disorder:
Sauteraud, A., Bourgeois, M., & Dartigues, J.-F. (2001). A randomized controlled Randomized compared with nonrandomized samples. Journal of Consulting and
trial of cognitive therapy versus intensive behavior therapy in obsessive compulsive Clinical Psychology, 68(4), 594–602. https://doi.org/10.1037/0022-006X.68.4.594
disorder. Psychotherapy and Psychosomatics, 70(6), 288–297. https://doi.org/ Freeston, M. H., Ladouceur, R., Laval, U., Gagnon, F., Thibodeau, N., Laval, C. H. D.
10.1159/000056269 Universite, Rheaume, J., Letarte, H., Bujold, A., Laval, U., & Doren, V. (1997).
Craske, M. (2015). Optimizing exposure therapy for anxiety disorders: An inhibitory Cognitive-Behavioral treatment of obsessive thoughts: A controlled study. Journal of
learning and inhibitory regulation approach. Verhaltenstherapie, 25(2), 134–143. Consulting and Clinical Psychology, 65(3), 405–413.
https://doi.org/10.1159/000381574 García-Soriano, G., Rufer, M., Delsignore, A., & Weidt, S. (2014). Factors associated with
Craske, M. G., Kircanski, K., Zelikowsky, M., Mystkowski, J., Chowdhury, N., & Baker, A. non-treatment or delayed treatment seeking in OCD sufferers: A review of the
(2008). Optimizing inhibitory learning during exposure therapy. Behaviour Research literature. Psychiatry Research, 220(1,2), 1–10. https://doi.org/10.1016/j.
and Therapy, 46(1), 5–27. https://doi.org/10.1016/j.brat.2007.10.003 psychres.2014.07.009
Dell’Osso, B., Buoli, M., Hollander, E., & Altamura, A. (2009). Duration of untreated Gava, I., Barbui, C., Aguglia, E., Carlino, D., Churchill, R., Vanna, M. D., & McGuire, H.
illness as a predictor of treatment response and remission in obsessive–compulsive (2007). Psychological treatments versus treatment as usual for obsessive compulsive
disorder. World Journal of Biological Psychiatry, 11(1), 59–65. disorder (OCD). Cochrane Database of Systematic Reviews, 2, 50.
Durna, G., Yorulmaz, O., & Aktaç, A. (2019). Public stigma of obsessive compulsive Geller, D., Biederman, J., Jones, J., Park, K., Schwartz, S., Shapiro, S., & Coffey, B.
disorder and schizophrenic disorder: Is there really any difference? Psychiatry (1998). Is juvenile obsessive-compulsive disorder a developmental subtype of the
Research, 271, 559–564. https://doi.org/10.1016/j.psychres.2018.12.065 disorder? A review of the pediatric literature. Journal of the American Academy of
Eddy, K. T., Dutra, L., Bradley, R., & Westen, D. (2004). A multidimensional meta- Child & Adolescent Psychiatry, 37(4), 420–427. https://doi.org/10.1097/00004583-
analysis of psychotherapy and pharmacotherapy for obsessive-compulsive disorder. 199804000-00020
Clinical Psychology Review, 24(8), 1011–1030. Gomes, J. B., Cordioli, A. V., Bortoncello, C. F., Braga, D. T., Gonçalves, F., & Heldt, E.
Eisen, J. L., Sibrava, N. J., Boisseau, C. L., Mancebo, M. C., Stout, R. L., Pinto, A., & (2016). Impact of cognitive-behavioral group therapy for obsessive-compulsive
Rasmussen, S. A. (2013). Five-year course of obsessive-compulsive disorder: disorder on family accommodation: A randomized clinical trial. Psychiatry Research,
Predictors of remission and relapse. Journal of Clinical Psychiatry, 74(3), 233–239. 246(30), 70–76. https://doi.org/10.1016/j.psychres.2016.09.019
https://doi.org/10.4088/JCP.12m07657 Higgins, J. P. T., & Green, S. (2011). Cochrane handbook for systematic reviews of
Federici, A., Summerfeldt, L. J., Harrington, J. L., McCabe, R. E., Purdon, C. L., Rowa, K., interventions. John Wiley & Sons.
& Antony, M. M. (2010). Consistency between self-report and clinician-administered Jacoby, R. J., & Abramowitz, J. S. (2016). Inhibitory learning approaches to exposure
versions of the Yale-Brown obsessive–compulsive scale. Journal of Anxiety Disorders, therapy: A critical review and translation to obsessive-compulsive disorder. Clinical
24(7), 729–733. https://doi.org/10.1016/j.janxdis.2010.05.005 Psychology Review, 49, 28–40. https://doi.org/10.1016/j.cpr.2016.07.001
Fineberg, N. A., Baldwin, D. S., Drummond, L. M., Wyatt, S., Hanson, J., Gopi, S., Jaurrieta, N., Jimenez-Murcia, S., Menchón, J. M., Alonso, M. D. P., Segalas, C., ÁLvarez-
Kaur, S., Reid, J., Marwah, V., Sachdev, R. A., Pampaloni, I., Shahper, S., Moya, E. M., Labad, J., Granero, R., & Vallejo, J. (2008). Individual versus group
Varlakova, Y., Mpavaenda, D., Manson, C., O’Leary, C., Irvine, K., Monji-Patel, D., cognitive–behavioral treatment for obsessive–compulsive disorder: A controlled
Shodunke, A., … Wellsted, D. (2018). Optimal treatment for obsessive compulsive pilot study. Psychotherapy Research, 18(5), 604–614. https://doi.org/10.1080/
disorder: A randomized controlled feasibility study of the clinical-effectiveness and 10503300802192141
cost-effectiveness of cognitive-behavioural therapy, selective serotonin reuptake Jones, S., Carley, S., & Harrison, M. (2003). An introduction to power and sample size
inhibitors and their combination in the management of obsessive compulsive estimation. Emerg. Med. J. EMJ, 20, 453–458. https://doi.org/10.1136/emj.20.5.453
disorder. International Clinical Psychopharmacology, 33(6), 334–348. https://doi.org/ Khodarahimi, S. (2009). Satiation therapy and exposure response prevention in the
10.1097/YIC.0000000000000237 treatment of obsessive compulsive disorder. Journal of Contemporary Psychotherapy,
Fineberg, N. A., Brown, A., Reghunandanan, S., & Pampaloni, I. (2012). Evidence-based 39, 203–207. https://doi.org/10.1007/s10879-009-9110-z
pharmacotherapy of obsessive-compulsive disorder. International Journal of Kim, S. W., Dysken, M. W., & Kuskowski, M. (1990). The Yale-Brown obsessive-
Neuropsychopharmacology, 15(8), 1173–1191. https://doi.org/10.1017/ compulsive scale: A reliability and validity study. Psychiatry Research, 34(1), 99–106.
S1461145711001829 https://doi.org/10.1016/0165-1781(90)90061-9
Fineberg, N. A., Hughes, A., Gale, D. T. M., & Roberts, A. (2005). Group cognitive Kircanski, K., Mortazavi, A., Castriotta, N., Baker, A. S., Mystkowski, J. L., Yi, R., &
behaviour therapy in obsessive-compulsive disorder (OCD): A controlled study. Craske, M. G. (2012). Challenges to the traditional exposure paradigm: Variability in
International Journal of Psychiatry in Clinical Practice, 9(4), 257–263. https://doi.org/ exposure therapy for contamination fears. Journal of Behavior Therapy and
10.1080/13651500500307180 Experimental Psychiatry, 43(2), 745–751. https://doi.org/10.1016/j.
Fisher, P. L., Cherry, M. G., Stuart, T., Rigby, J. W., & Temple, J. (2020). People with jbtep.2011.10.010
obsessive-compulsive disorder often remain symptomatic following psychological
14
Knopp, J., Knowles, S., Bee, P., Lovell, K., & Bower, P. (2013). A systematic review of Olatunji, B. O., Davis, M. L., Powers, M. B., & Smits, J. A. J. (2013). Cognitive-behavioral
predictors and moderators of response to psychological therapies in OCD: Do we therapy for obsessive-compulsive disorder: A meta-analysis of treatment outcome
have enough empirical evidence to target treatment? Clinical Psychology Review, 33 and moderators. Journal of Psychiatry Research, 47(1), 33–41. https://doi.org/
(8), 1067–1081. 10.1016/j.jpsychires.2012.08.020
Kobak, K. A., Greist, J. H., Jefferson, J. W., Katzelnick, D. J., & Henk, H. J. (1998). Öst, L.-G., Havnen, A., Hansen, B., & Kvale, G. (2015). Cognitive behavioral treatments of
Behavioral versus pharmacological treatments of obsessive compulsive disorder: A obsessive–compulsive disorder. A systematic review and meta-analysis of studies
meta-analysis. Psychopharmacology, 136, 205–216. https://doi.org/10.1007/ published 1993–2014. Clinical Psychology Review, 40, 156–169. https://doi.org/
s002130050558 10.1016/j.cpr.2015.06.003
Liberati, A., Altman, D. G., Tetzlaff, J., Mulrow, C., Gøtzsche, P. C., Ioannidis, J. P. A., Patel, S. R., & Simpson, H. B. (2010). Patient preferences for obsessive-compulsive
Clarke, M., Devereaux, P. J., Kleijnen, J., & Moher, D. (2009). The PRISMA statement disorder treatment. Journal of Clinical Psychiatry, 71(11), 1434–1439. https://doi.
for reporting systematic reviews and meta-analyses of studies that evaluate health org/10.4088/JCP.09m05537blu
care interventions: Explanation and elaboration. Journal of Clinical Epidemiology, 62 Pearcy, C. P., Anderson, R. A., Egan, S. J., & Rees, C. S. (2016). A systematic review and
(10), 1–34. https://doi.org/10.1371/journal.pmed.1000100 meta-analysis of self-help therapeutic interventions for obsessive–compulsive
Lindsay, M., Crino, R., & Andrews, G. (1997). Controlled trial of exposure and response disorder: Is therapeutic contact key to overall improvement? Journal of Behavior
prevention in obsessive–compulsive disorder. The British Journal of Psychiatry, 171, Therapy and Experimental Psychiatry, 51, 74–83.
135–139. https://doi.org/10.1192/bjp.171.2.135 Ponniah, K., Magiati, I., & Hollon, S. D. (2013). An update on the efficacy of
Mancebo, M. C., Garcia, A. M., Pinto, A., Freeman, J. B., Przeworski, A., Stout, R., psychological therapies in the treatment of obsessive-compulsive disorder in adults.
Kane, J. S., Eisen, J. L., & Rasmussen, S. A. (2008). Juvenile-onset OCD: Clinical Journal of Obsessive-Compulsive and Related Disorders, 2(2), 207–218. https://doi.org/
features in children, adolescents and adults. Acta Psychiatrica Scandinavica, 118(2), 10.1016/j.jocrd.2013.02.005
149–159. https://doi.org/10.1111/j.1600-0447.2008.01224.x Pozza, A., & Dèttore, D. (2017). Drop-out and efficacy of group versus individual
Marsden, Z., Lovell, K., Blore, D., Ali, S., & Delgadillo, J. (2018). A randomized cognitive behavioural therapy: What works best for obsessive-compulsive disorder?
controlled trial comparing EMDR and CBT for obsessive–compulsive disorder. A systematic review and meta-analysis of direct comparisons. Psychiatry Research,
Clinical Psychology & Psychotherapy, 25(1), e10–e18. https://doi.org/10.1002/ 258, 24–36. https://doi.org/10.1016/j.psychres.2017.09.056
cpp.2120 Rachman, S., Hodgson, R., & Marks, I. M. (1971). The treatment of chronic obsessive-
Mayerovitch, J. I., du Fort, G. G., Kakuma, R., Bland, R. C., Newman, S. C., & Pinard, G. compulsive neurosis. Behaviour Research and Therapy, 9(3), 237–247. https://doi.
(2003). Treatment seeking for obsessive-compulsive disorder: Role of obsessive- org/10.1016/0005-7967(71)90009-X
compulsive disorder symptoms and comorbid psychiatric diagnoses. Comprehensive Rector, N. A., Cassin, S. E., & Richter, M. A. (2009). Psychological treatment of obsessive-
Psychiatry, 44(2), 162–168. https://doi.org/10.1053/comp.2003.50005 compulsive disorder in patients with major depression: A pilot randomized
McKay, D., Abramowitz, J. S., Calamari, J. E., Kyrios, M., Radomsky, A., Sookman, D., controlled trial. Canadian Journal of Psychiatry, 54(12), 846–851. https://doi.org/
Taylor, S., & Wilhelm, S. (2004). A critical evaluation of obsessive–compulsive 10.1177/070674370905401208
disorder subtypes: Symptoms versus mechanisms. Clinical Psychology Review, 24(3), Robbins, T. W., Vaghi, M. M., & Banca, P. (2019). Obsessive-compulsive disorder: Puzzles
283–313. https://doi.org/10.1016/j.cpr.2004.04.003 and prospects. Neuron, 102(1), 27–47. https://doi.org/10.1016/j.
McKay, D., Sookman, D., Neziroglu, F., Wilhelm, S., Stein, D. J., Kyrios, M., Matthews, K., neuron.2019.01.046
& Veale, D. (2015). Efficacy of cognitive-behavioral therapy for Romanelli, R. J., Wu, F. M., Gamba, R., Mojtabai, R., & Segal, J. B. (2014). Behavioral
obsessive–compulsive disorder. Psychiatry Research, 225(3), 236–246. https://doi. therapy and serotonin reuptake inhibitor pharmacotherapy in the treatment of
org/10.1016/j.psychres.2014.11.058 obsessive–compulsive disorder: A systematic review and meta-analysis of head-to-
McLean, P. D., Whittal, M. L., Thordarson, D. S., Taylor, S., Söchting, I., Koch, W. J., head randomized controlled trials. Depression and Anxiety, 31(8), 641–652. https://
Paterson, R., & Anderson, K. W. (2001). Cognitive versus behavior therapy in the doi.org/10.1002/da.22232
group treatment of Obsessive-Compulsive disorder. Journal of Consulting and Clinical Rosaalcazar, A., Sanchezmeca, J., Gomezconesa, A., & Marinmartinez, F. (2008).
Psychology, 69(2), 205–214. https://doi.org/10.1037/0022-006X.69.2.205 Psychological treatment of obsessive–compulsive disorder: A meta-analysis. Clinical
McMillan, D., & Lee, R. (2010). A systematic review of behavioral experiments vs Psychology Review, 28(8), 1310–1325. https://doi.org/10.1016/j.cpr.2008.07.001
exposure alone in the treatment of anxiety disorders: A case of exposure while Ruscio, A. M., Stein, D. J., Chiu, W. T., & Kessler, R. C. (2010). The epidemiology of
wearing the emperor’s new clothes? Clinical Psychology Review, 30(5), 467–478. obsessive-compulsive disorder in the national comorbidity survey replication.
https://doi.org/10.1016/j.cpr.2010.01.003 Molecular Psychiatry, 15(1), 53–63. https://doi.org/10.1038/mp.2008.94
Meyer, V. (1966). Modification of expectations in cases with obsessional rituals. Salkovskis, P. M. (1985). Obsessional-compulsive problems: A cognitive-behavioural
Behaviour Research and Therapy, 4(4), 273–280. https://doi.org/10.1016/0005-7967 analysis. Behaviour Research and Therapy, 23(5), 571–583. https://doi.org/10.1016/
(66)90023-4 0005-7967(85)90105-6
Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & PRISMA group. (2009). Preferred Shean, G. (2014). Limitations of randomized control designs in psychotherapy research.
reporting Items for systematic reviews and meta-analyses: The PRISMA statement. Advances in Psychiatry, 1–5, 2014.
PLoS Medicine, 6(7), Article e1000097. https://doi.org/10.1371/journal. Simpson, H. B., Franklin, M. E., Cheng, J., Foa, E. B., & Liebowitz, M. R. (2005). Standard
pmed.1000097 criteria for relapse are needed in obsessive-compulsive disorder. Depression and
Morris, S. B. (2008). Estimating effect sizes from pretest-posttest-control group designs. Anxiety, 21(1), 1–8. https://doi.org/10.1002/da.20052
Organizational Research Methods, 11(2), 364–386. https://doi.org/10.1177/ Simpson, H. B., Liebowitz, M. R., Foa, E. B., Kozak, M. J., Schmidt, A. B., Rowan, V.,
1094428106291059 Petkova, E., Kjernisted, K., Huppert, J. D., Franklin, M. E., Davies, S. O., &
Morris, S. B., & Deshon, R. P. (2002). Combining effect size estimates in meta-analysis Campeas, R. (2004). Post-treatment effects of exposure therapy and clomipramine in
with repeated measures and independent-groups designs. Psychological Methods, 7 obsessive-compulsive disorder. Depression and Anxiety, 19(4), 225–233. https://doi.
(1), 105–125. https://doi.org/10.1037//1082-989X.7.1.105 org/10.1002/da.20003
Mowrer, O. H. (1956). Two-factor learning theory reconsidered, with special reference to Skapinakis, P., Caldwell, D. M., Hollingworth, W., Bryden, P., Fineberg, N. A.,
secondary reinforcement and the concept of habit. Psychological Review, 63(2), Salkovskis, P., Welton, N. J., Baxter, H., Kessler, D., Churchill, R., & Lewis, G. (2016).
114–128. Pharmacological and psychotherapeutic interventions for management of obsessive-
Mowrer, O. H. (1960). Revised two-factor theory and the concept of habit. In Learning compulsive disorder in adults: A systematic review and network meta-analysis. The
theory and behavior (pp. 212–252). Hoboken, NJ, US: John Wiley & Sons Inc. https:// Lancet Psychiatry, 3(8), 730–739. https://doi.org/10.1016/S2215-0366(16)30069-4
doi.org/10.1037/10802-007. Sousa, M. B., Isolan, L. R., Oliveira, R. R., Manfro, G. G., & Cordioli, A. V. (2006).
Nakatani, E., Nakagawa, A., Nakao, T., Yoshizato, C., Nabeyama, M., Kudo, A., A randomized clinical trial of cognitive-behavioral group therapy and sertraline in
Isomura, K., Kato, N., Yoshioka, K., & Kawamoto, M. (2005). A randomized the treatment of obsessive-compulsive disorder. Journal of Clinical Psychiatry, 67(7),
controlled trial of Japanese patients with obsessive-compulsive disorder – 1133–1139.
effectiveness of behavior therapy and fluvoxamine. Psychotherapy and Stein, D. J. (2002). Obsessive-compulsive disorder. The Lancet, 360(9330), 397–405.
Psychosomatics, 74(5), 269–276. https://doi.org/10.1159/000086317 https://doi.org/10.1016/S0140-6736(02)09620-4
Nestadt, G., Samuels, J., Riddle, M. A., Liang, K.-Y., Bienvenu, O. J., Hoehn-Saric, R., Steketee, G., Frost, R., Tolin, D., Rasmussen, J., & Brown, T. (2010). Waitlist-controlled
Grados, M., & Cullen, B. (2001). The relationship between obsessive–compulsive trial of cognitive behavior therapy for hoarding disorder. Depression and Anxiety, 27
disorder and anxiety and affective disorders: Results from the johns hopkins OCD (5), 476–484.
family study. Psychological Medicine, 31(3), 481–487. https://doi.org/10.1017/ Steketee, G., & Shapiro, L. (1995). Predicting behavioral treatment outcome for
S0033291701003579 agoraphobia and obsessive compulsive disorder. Clinical Psychology Review, 15(4),
Newth, S., & Rachman, S. (2001). The concealment of obsessions. Behaviour Research and 317–346.
Therapy, 39(4), 457–464. https://doi.org/10.1016/S0005-7967(00)00006-1 Taylor, S. (2011). Early versus late onset obsessive–compulsive disorder: Evidence for
Nutt, D., & Malizia, A. (2006). Anxiety and OCD – the chicken or the egg? Journal of distinct subtypes. Clinical Psychology Review, 31(7), 1083–1100. https://doi.org/
Psychopharmacology, 20(6), 729–731. https://doi.org/10.1177/0269881106068424 10.1016/j.cpr.2011.06.007
O’Connor, K. P., Aardema, F., Bouthillier, D., Fournier, S., Guay, S., Robillard, S., Taylor, S., Coles, M. E., Abramowitz, J. S., Wu, K. D., Olatunji, B. O., Timpano, K. R.,
Pélissier, M. C., Todorov, C., Tremblay, M., & Pitre, D. (2005). Evaluation of an McKay, D., Kim, S.-K., Carmin, C., & Tolin, D. F. (2010). How are dysfunctional
inference-based approach to treating obsessive-compulsive disorder. Cognitive beliefs related to obsessive-compulsive symptoms? Journal of Cognitive
Behaviour Therapy, 34(3), 148–163. https://doi.org/10.1080/16506070510041211 Psychotherapy, 24(3), 165–176. https://doi.org/10.1891/0889-8391.24.3.165
O’Connor, K., Todorov, C., Robillard, S., Borgeat, F., & Brault, M. (1999). Cognitive- The Metafor Package. (2017). Metaf. Package meta-anal. Package R. URL http://www.
behaviour therapy and medication in the treatment of obsessive-compulsive metafor-project.org/doku.php/analyses:morris2008 accessed 7.25.19.
disorder: A controlled study. Canadian Journal of Psychiatry, 44(1), 64–71. https:// Tolin, D. F., Hannan, S., Maltby, N., Diefenbach, G. J., Worhunsky, P., & Brady, R. E.
doi.org/10.1177/070674379904400108 (2007). A randomized controlled trial of self-directed versus therapist-directed
cognitive-behavioral therapy for obsessive-compulsive disorder patients with prior
15
medication trials. Behavior Therapy, 38(2), 179–191. https://doi.org/10.1016/j. Vogel, P. A., Stiles, T. C., & Götestam, K. G. (2004). Adding cognitive therapy elements to
beth.2006.07.001 exposure therapy for obsessive-compulsive disorder: A controlled study. Behavioural
Torres, A. R., Prince, M. J., Bebbington, P. E., Bhugra, D. K., Brugha, T. S., Farrell, M., and Cognitive Psychotherapy, 32(3), 275–290. https://doi.org/10.1017/
Jenkins, R., Lewis, G., Meltzer, H., & Singleton, N. (2007). Treatment seeking by S1352465804001353
individuals with obsessive-compulsive disorder from the British psychiatric Whittal, M. L., Thordarson, D. S., & McLean, P. D. (2005). Treatment of
morbidity survey of 2000. Psychiatric Services, 58(7), 977–982. https://doi.org/ obsessive–compulsive disorder: Cognitive behavior therapy vs. exposure and
10.1176/ps.2007.58.7.977 response prevention. Behaviour Research and Therapy, 43(12), 1559–1576.
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Woody, S. R., Steketee, G., & Chambless, D. L. (1995). Reliability and validity of the Yale-
Journal of Statistical Software, 36(3). https://doi.org/10.18637/jss.v036.i03 Brown obsessive-compulsive scale. Behaviour Research and Therapy, 33(5), 597–605.
Visser, H. A., van Megen, H., van Oppen, P., Eikelenboom, M., Hoogendorn, A. W., https://doi.org/10.1016/0005-7967(94)00076-V
Kaarsemaker, M., & van Balkom, A. J. (2015). Inference-based approach versus World Health Organization. (2018). International calssification of diseases (ICD) [WWW
cognitive behavioral therapy in the treatment of obsessive-compulsive disorder with Document]. URL https://icd.who.int/browse11/l-m/en accessed 8.7.19.
poor insight: A 24-session randomized controlled trial. Psychotherapy and
Psychosomatics, 84(5), 284–293. https://doi.org/10.1159/000382131
16

Toc Epr

Uploaded by

Copyright:

Available Formats

Toc Epr

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Toc Epr

Uploaded by

Copyright:

Available Formats

Journal of Obsessive-Compulsive and Related Disorders 31 (2021) 100684

Contents lists available at ScienceDirect

Journal of Obsessive-Compulsive and Related Disorders

A systematic review and meta-analysis on the effectiveness of exposure and

patient-therapist contact equivalent to a non-internet-based therapy, 2.3. Information sources

Lindsay et al. ERP AM 9 9 44 89 31.6 34 9 (8.7) 15 N Interview NA NA Australia

Journal of Obsessive-Compulsive and Related Disorders 31 (2021) 100684

3.3. Main analysis

and the ‘other’ groups in YBOCS score reduction at pre-treatment versus

− 8.87). The results can be visualized in Fig. 2.

some unexplained heterogeneity between studies. Cochrane guidelines,

the heterogeneity result can be ignored (Higgins & Green, 2011).

heterogeneity, a leave1out test was conducted. This computation runs

the main meta-analysis excluding each study successively, therefore

assessing whether the results remain significant each time, eventually

was not affected by the heterogeneity between the studies. Additionally,

we ran influential case diagnostics on the studies to identify any po­

tential influential cases, following guidelines used throughout the sta­

Gomes et al., 2016 respectively) were outliers and potential drivers of

Treatment-As-Usual; EMDR = Eye Movement Desensitisation and Reprocessing; Y=Yes; N–

del). Thus, to test the robustness of our findings, we ran the

fected by the removal of these studies (effect size = − 0.97, significant at

p < .01 (95% CI = [-1.18 to − 0.76]; SE = 0.11, z = − 8.94)). This,

can be considered valid.

3.4. Moderator analyses

The first moderator analysis examined whether the addition of

cognitive elements within treatment would affect the significant

therapy. The test of moderators was significant (QM = 160.04; p <

[-2.13 to − 1.35]; SE = 0.20; z= − 8.68, p < .001) and without (estimate

= − 2.87; 95% CI = [-3.48 to − 2.26]; SE = 0.31; z = − 9.20, p < .001)

pre-treatment versus post-treatment. This suggests both treatment var­

iations significantly reduce OCD symptoms. The test of heterogeneity

ratio of unaccounted variability to sampling variability was, H^2 = 5.54.

These results can be disregarded as we had already conducted additional

Next, we assessed whether number of hours of therapy affected the

reduction in YBOCS pre-treatment to post-treatment. The moderator

although we can discount it once again. 4. Discussion

Appendix 1. Additional data extrapolation process

Appendix 2. Plots of influential case diagnostics

Appendix 3. Scales used to measure depression, anxiety and QoL

Author Year Depression Scale Anxiety Scale QoL scale

Lindsay et al. 1997 BDI STAI Interference Rating Scale

Role of funding source

You might also like

we ran influential case diagnostics on the studies to identify any po

tential influential cases, following guidelines used throughout the sta

pre-treatment versus post-treatment. This suggests both treatment var