Elsevier

Clinical Psychology Review

Volume 40, August 2015, Pages 156-169
Clinical Psychology Review

Cognitive behavioral treatments of obsessive–compulsive disorder. A systematic review and meta-analysis of studies published 1993–2014

https://doi.org/10.1016/j.cpr.2015.06.003Get rights and content

Highlights

  • CBT yielded very large effect sizes compared to wait list and placebo.

  • CBT was significantly better than antidepressants.

  • The addition of antidepressants did not potentiate the effect of CBT.

  • There was no significant difference between ERP and cognitive therapy.

  • There was no significant difference between individual and group treatment.

Abstract

Obsessive–compulsive disorder is ranked by the WHO as among the 10 most debilitating disorders and tends to be chronic without adequate treatment. The only psychological treatment that has been found effective is cognitive behavior therapy (CBT). This meta-analysis includes all RCTs (N = 37) of CBT for OCD using the interview-based Yale–Brown Obsessive Compulsive Scale, published 1993 to 2014. The effect sizes for comparisons of CBT with waiting-list (1.31), and placebo conditions (1.33) were very large, whereas those for comparisons between individual and group treatment (0.17), and exposure and response prevention vs. cognitive therapy (0.07) were small and non–significant. CBT was significantly better than antidepressant medication (0.55), but the combination of CBT and medication was not significantly better than CBT plus placebo (0.25). The RCTs have a number of methodological problems and recommendations for improving the methodological rigor are discussed as well as clinical implications of the findings.

Introduction

Obsessive–compulsive disorder (OCD) is characterized by anxiety evoking intrusive thoughts, images or urges (obsessions) and repetitive behaviors aimed at reducing the discomfort (compulsions). The lifetime prevalence of OCD has been estimated to approximately 2% (Kessler et al., 2005) and OCD has been ranked by the WHO among the 10 most debilitating disorders. Untreated OCD tends to be chronic, causing significant functional impairment and reduced quality of life (Koran, Thienemann, & Davenport, 1996).

The most common treatment for OCD is pharmacological (Blanco et al., 2006), primarily with selective serotonin reuptake inhibitors (SSRIs). A meta-analysis by Soomro, Altman, Rajagopal, and Oakley-Browne (2008) found that SSRIs were significantly better than placebo and that the weighted mean difference on the Yale–Brown Obsessive Compulsive Scale (Goodman et al., 1989) was 3.21 points in favor of SSRIs.

The recommended treatment of choice for OCD is cognitive-behavioral therapy (CBT; National Collaborating Centre for Mental Health, 2006) which refers to exposure and response prevention (ERP) with or without the inclusion of cognitive therapy strategies. The most recent meta-analysis on the effects of psychological treatments for OCD was carried out by Olatunji, Davis, Powers, and Smits (2013). However, in this analysis only 16 studies were included, and three of these were on the treatment of pediatric OCD. The main reason for the small sample size with only 13 adult treatment studies may be the application of strict exclusion criteria, as the authors report that 21 studies were excluded due to having active treatment as control condition.

With the exception of Olatunji et al. (2013), the last extensive meta-analysis on the treatment of adult OCD was conducted by Rosa-Alcázar, Sánchez-Meca, Gómez-Conesa, and Marín-Martínez (2008), who included 19 controlled studies published in the period 1980–2006. Of the sample, which also included studies with a non-randomized design, 18 studies were published in the period from 1993 to 2006, thus overlapping with the inclusion period of the present meta-analysis. Since 2006 more than 20 controlled trials have been published on the treatment of OCD and it is evident that a significant proportion of the latest research on OCD has yet to be analyzed by meta-analytic procedures, and potentially twice the number of studies analyzed by Rosa-Alcázar et al. (2008) can be evaluated for inclusion in the present study. Eight of the studies included in the present meta-analysis overlap with the Rosa-Alcázar et al. (2008) study. Nine studies included in the Rosa-Alcázar et al. (2008) analysis were excluded from the present analysis as they failed to meet our inclusion criteria of randomized design (n = 8) or using Yale–Brown Obsessive–Compulsive Severity Scale (Y-BOCS; Goodman et al., 1989) as the primary outcome measure (n = 1). In summary, this warrants an updated meta-analysis on the cognitive behavioral treatment of adult OCD.

Despite randomization, RCTs may vary substantially in methodological stringency. In a recent meta-analysis of 60 RCTs on Acceptance and Commitment Therapy (Öst, 2014) a significant association between low methodological stringency and high effect size was found. In the previously published meta-analyses on OCD, no systematic efforts were made to evaluate the methodological qualities of the included studies. To investigate whether methodological stringency and outcome are related in the treatment of OCD this issue will be explored in the present meta-analysis.

In the randomized trials on the treatment of OCD a wide variation in outcome measures have been applied and the differences in outcome measures represent a challenge when comparing studies. Although it may be assumed that by standardizing the outcome measures and comparing effect sizes the outcome will not be influenced, it can be argued that the standardization and comparison of multiple outcome measures may bias the calculated effect sizes because the standard deviations (SDs) may vary substantially between measures (Morris & DeShon, 2002). In order to avoid this potential bias we decided to apply a common outcome measure as criterion for inclusion in the present meta-analysis. The Y-BOCS (Goodman et al., 1989), which has been widely used as primary outcome measure in research on the treatment of OCD and has been established as the “gold standard” of OCD symptom measures, was chosen. Research has demonstrated only a moderate relationship between the interview and self-report version of Y-BOCS in a clinical sample of OCD patients (Federici et al., 2010). In the present analysis we therefore decided to only include studies that applied clinician administered Y-BOCS interview as outcome measure.

The calculation of effect sizes is a widely used means when comparing the results of outcome measures across studies. Effect sizes nicely demonstrate statistically significant changes following treatment and thus provide a basis for comparison of treatment outcome between studies; however, the effect sizes do not provide a way of determining clinically significant improvement. To overcome this limitation, Jacobson and Truax (1991) have recommended procedures for calculating clinical improvement and classifying patients accordingly in the categories “recovered”, “improved”, “unchanged” and “deteriorated”. To be classified as recovered, the patient must a) show a change that is larger than the measurement error (the reliable change index; RCI) from pre- to post-treatment, and b) be in the range of the non-clinical population after treatment. In the present analysis the treatment outcome of the included studies will be analyzed and compared both in terms of effect size as well as clinically significant change; an approach not previously applied in meta-analyses of OCD.

ERP has often been described as a challenging treatment due to the confrontation to anxiety provoking cues and it has been estimated that approximately 25% of patients refuse the offer of treatment (Franklin & Foa, 1998), and this number is suggested to reflect that patients “find [behavior therapy] too frightening” (p. 353). Many studies have referred to this refusal rate, even though Franklin and Foa (1998) did not provide a reference for their estimate. It is thus unclear whether the estimate is valid and we will investigate this issue empirically by calculating the refusal rates in the included studies in our meta-analysis. A related question is how many patients drop out of treatment prematurely. An attrition rate of 25–30% has often been referred to (Abramowitz, 2006, Kozak et al., 2000), however, Kozak et al. (2000) based their estimate on only one study and Abramowitz (2006) did not provide a reference for his estimate. In general there is a huge variation in attrition rates across studies and in order to calculate a valid estimate it is necessary to include a large number of research trials. In the present meta-analysis we will therefore provide a calculation of attrition rates from all included randomized controlled trials published from 1993 to 2014.

ERP has been demonstrated as a treatment that can be disseminated in different modes of therapy and a relevant issue is whether exposure assignments are equally effective when self-administered as when administered by a therapist. In addition there is the question if type of exposure (in vivo versus imaginal exposure) has relevance for the outcome. Abramowitz (1996) published a meta-analysis evaluating different variants of ERP for OCD and the influence on outcome. He concluded that therapist assisted ERP was superior to self-administered ERP. Furthermore, he evaluated the type of exposure and concluded that the combination of in-vivo and imaginal exposure produced better outcome than in-vivo exposure alone. In addition the amount of therapy is known to vary substantially across different treatment formats, both with respect to length of sessions, frequency of sessions and length of the course of therapy. ERP has been disseminated across a range of different formats, e.g. group therapy, family-based interventions, etc., and the potential relation between treatment format and outcome will be analyzed in the present meta-analysis.

To sum up, there are many reasons that suggest the need for an update of the empirical basis for the psychological treatment of OCD and potential moderators of treatment. In the present meta-analysis we aim to provide an updated analysis of several important questions which are of relevance for the treatment of OCD in accordance with the following goals:

  • a)

    To provide an updated review and meta-analysis regarding the efficacy of cognitive-behavioral treatments of OCD from 1993 until 2014 using meta-analytic procedures.

  • b)

    To evaluate potential moderators for treatment outcome.

  • c)

    To evaluate the included studies according to the methodological criteria proposed by Öst (2008) and calculate if there is any difference between the first and the second 10 year period.

  • d)

    To provide recommendations for enhanced methodological stringency in future research on the basis of the methodological evaluation of OCD studies.

Section snippets

Literature search

PsycINFO and PubMed were searched from 1993 to February 2015 with the following search words: obsessive–compulsive disorder or OCD, and exposure and response prevention or ERP or behavior therapy or cognitive therapy or cognitive behavior therapy, and Randomized controlled trial or RCT or random*. The reason to use 1993 as the starting year is that the first RCTs that used the Y-BOCS as outcome measure was published that year.

All abstracts were read and when there was an indication of a group

Background data

Background data of the studies included in the meta-analysis are displayed in Appendix A, Table A.1. The 37 studies (see Appendix B) originated from USA (n = 10), Canada (n = 7), Brazil (n = 4), Australia (n = 3), Spain (n = 3), Holland (n = 2), Norway (n = 2), and one each from Denmark, France, Germany, Great Britain, Japan, and Sweden. A total of 2414 participants started treatment or control conditions. The proportion of females ranged from 29 to 77% with a mean of 57.9%, and the mean age of the samples

Discussion

The primary aims of the present meta-analysis were to provide a methodological review of the randomized controlled trials of cognitive behavioral treatment of OCD in adult patients published from 1993 to 2014, and to assess their efficacy. A total of 37 studies met criteria for inclusion, which makes this the largest meta-analysis of randomized controlled trials for OCD. In accordance with previous, considerably smaller meta-analyses on cognitive behavioral treatment of OCD (e.g. Abramowitz,

Funding

No external funding was obtained for this meta-analysis.

Role of funding sources

The authors have not obtained any financial support for the preparation of the present manuscript.

Contributors

LGÖ designed the meta-analysis, wrote the coding schema, rated the studies, meta-analyzed the included studies, and wrote the first draft of method and results. AH did the literature search, rated the studies, and wrote the first draft of introduction and discussion. BH and GK rated the studies and co-wrote the first draft of introduction and discussion with AH. All authors participated in the revision and approved of the final manuscript.

Conflict of interest

None of the authors have any conflict of interest to report.

References (63)

  • A.I. Rosa-Alcázar et al.

    Psychological treatment of obsessive–compulsive disorder: a meta-analysis

    Clinical Psychology Review

    (2008)
  • H.B. Simpson et al.

    Challenges using motivational interviewing as an adjunct to exposure therapy for obsessive–compulsive disorder

    Behaviour Research and Therapy

    (2010)
  • D.F. Tolin et al.

    A randomized controlled trial of self-directed versus therapist-directed cognitive-behavioral therapy for obsessive–compulsive disorder patients with prior medication trials

    Behavior Therapy

    (2007)
  • M.L. Whittal et al.

    Treatment of obsessive–compulsive disorder: Cognitive behavior therapy vs. exposure and response prevention

    Behaviour Research and Therapy

    (2005)
  • M.L. Whittal et al.

    Treatment of obsessions: A randomized controlled trial

    Behaviour Research and Therapy

    (2010)
  • B.M. Wootton et al.

    Remote treatment of obsessive–compulsive disorder: A randomized controlled trial

    Journal of Obsessive-Compulsive and Related Disorders

    (2013)
  • J.S. Abramowitz

    The psychological treatment of obsessive–compulsive disorder

    Canadian Journal of Psychiatry

    (2006)
  • E. Andersson et al.

    Internet-based cognitive behaviour therapy for obsessive–compulsive disorder: A randomized controlled trial

    Psychological Medicine

    (2012)
  • J.P. Barber et al.

    Development of the cognitive therapy adherence and competence scale

    Psychotherapy Research

    (2003)
  • J.P. Barber et al.

    Assessing intervention competence and its relation to therapy outcome: A selected review derived from the outcome literature

    Professional Psychology: Research and Practice

    (2007)
  • A. Belloch et al.

    Cognitive versus behaviour therapy in the individual treatment of obsessive–compulsive disorder: Changes in cognitions and clinically significant outcomes at post-treatment and one-year follow-up

    Behavioural and Cognitive Psychotherapy

    (2008)
  • Biostat Inc.

    Comprehensive meta-analysis, version 2. Statistical software

    (2006)
  • C. Blanco et al.

    Treatment of obsessive–compulsive disorder by U.S. psychiatrists

    Journal of Clinical Psychiatry

    (2006)
  • E. Cabedo et al.

    Group versus individual cognitive treatment for obsessive–compulsive disorder: Changes in severity at post-treatment and one-year follow-up

    Behavioural and Cognitive Psychotherapy

    (2010)
  • D.L. Chambless et al.

    Update on empirically validated therapies, II

    The Clinical Psychologist

    (1998)
  • D.L. Chambless et al.

    An update on empirically validated therapies

    The Clinical Psychologist

    (1996)
  • J. Cohen

    Statistical power analysis for the behavioral sciences

    (1988)
  • A.V. Cordioli et al.

    Cognitive-behavioral group therapy in obsessive–compulsive disorder: A randomized clinical trial

    Psychotherapy and Psychosomatics

    (2003)
  • J. Cottraux et al.

    A randomized controlled trial of cognitive therapy versus intensive behavior therapy in obsessive compulsive disorder

    Psychotherapy and Psychosomatics

    (2001)
  • S. Duval et al.

    Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis

    Biometrics

    (2000)
  • M. Egger et al.

    Bias in meta-analysis detected by a simple, graphical test

    British Medical Journal

    (1997)
  • Cited by (375)

    View all citing articles on Scopus
    View full text