The Psychological Treatment of Obsessive-Compulsive Disorder

Jonathan S Abramowitz, PhD'

The psychological treatment of obsessive-compulsive disorder (OCD) with exposure and

response prevention (ERP) methods is one ofthe great success stories within the field of men-
tal health. Within the span of about 20 years, the prognosis for individuals with OCD has
changed from poor to very good as a result ofthe development of ERP, This success not with-
standing, the procedures are far from perfect because a substantial minority of patients still
either refiise treatment, drop out prematurely, or fail to benefit, I begin this article with a
review of the development of ERP from early animal research on avoidance learning con-
ducted during the 1950s, Next, I discuss the mechanisms of ERP, The bulk ofthe article
reviews the treatment-outcome literature on ERP for OCD and includes comparisons with
cognitive therapythe "new kid on the block" with respect to psychological treatments for
OCD,
(Can J Psychiatry 2006;51:407-416)
Information on fiinding and support and author affiliations appears at the end ofthe article.

Highlights
Basic experimental reserch on humans and animal analogues of OCD serve as the basis for the
development of effective psychological treatments for this complex disorder,
Once considered treatment-resistant, OCD responds well to CBT,
This article also discusses factors that influence treatment outcome.

Key Words: ohsessive-compulsive disorder, cognitive-behavioural therapy, psychological

treatment, exposure, response prevention, cognitive therapy, treatment outcome

bsessive-compulsive disorder is characterized, first, by

O recurrent, unwanted, and seemingly bizarre thoughts,
impulses, or doubts that evoke affective distress (obsessions,
that are combined with reassurance-seeking rituals (for
example, compulsive checking); and unwanted, repugnant,
aggressive-violent, sexual, or blasphemous obsessional
for example, that one has struck a pedestrian with an automo- thoughts that are combined with covert compulsions or neu-
bile); and, second, by repetitive behavioural or mental rituals tralizing strategies such as mental rituals, praying, repeating
performed to reduce this distress (compulsions, for example, routine aetions, and thought suppression. Some patients also
constantly checking the rear-view mirror for injured individu- have excessive concems about lucky or unlucky numbers or
als). Obsessional fears tend to be about issues related to uncer- worries about orderliness and symmetry.
tainty about personal safety or the safety of others.
Compulsions are deliberately performed to reduce this uncer- Although many OCD sufferers recognize that their obses-
tainty, A large study on the various presentations of OCD sional fears and rituals are senseless and excessive, others
found that certain types of obsessions and compulsions occur strongly believe that their rituals serve to prevent the occur-
together within patients (1), These include obsessions regard- rence of disastrous consequences; in other words, they have
ing contamination that are combined with decontamination poor insight (2), Clinical observations suggest that a patient's
rituals (for example, inappropriate washing and cleaning); degree of insight can vary over time as well as across symptom
obsessions regarding responsibility for harm or catastrophe categories. For example, one patient evaluated in our clinic

Can J Psychiatry, Vol 51, No 7, June 2006 407

The Canadian Journal of PsychiatryIn Review

realized that her fear of causing her husband to die in a plane behaviourally oriented clinicians and researchers who looked
crash just by thinking about it was unrealistic (although she back to important animal-based research conducted in the
tried to prevent such thoughts, just to be on the safe side), yet 1950s to search for an animal analogue of OCD from which
she was strongly convinced that she would develop AIDS if they could conceptualize and develop behaviourally based
she did not shower after using a public washroom. It is impor- therapies, I discuss this historical event in some detail because
tant to ascertain patients' degree of insight because this can it highlights the methods used to derive behavioural therapy
affect treatment outcome, as I review further below. for OCD from experimental findings. Despite the positive
The lifetime prevalence rate of OCD in adults is 2% to 3% (3), impact that this approach has had on the management of OCD
Although symptoms typically wax and wane as a function of and other anxiety disorders, this method of deriving treatment
general life stress, a chronic and deteriorating course is the from experimental research remains, in the field of mental
norm if adequate treatment is not sought. In many cases, fears, health, unique to behaviourally oriented therapies.
avoidance, and rituals impair various areas of functioning,
including job or academic performance, social functioning,
Early Laboratory Research
and leisure activities. Many individuals with OCD also expe-
rience other Axis I disorders, such as mood and anxiety prob- The early work of Richard Solomon and his colleagues pro-
lems (4), This article provides an overview of the vides an elegant, yet often overlooked, animal behavioural
development of effective psychological treatments for OCD, model of OCD (6), Solomon and others worked with dogs in
together with a review of the latest treatment-outcome shuttle boxes (which were small rooms divided in 2 by a hur-
research, CBT is the most effective form of treatment for dle over which the animal could jump). Each half of the shuttle
OCD, Therefore, I focus on this intervention. box was separately furnished with an electric grate that could
be independently electrified to give the dog an electric shock
The Development of Effective Treatments for through its paws. In addition, a light served as a conditioned
stimulus. The procedure for producing the compulsive ritual-
OCD: From the Laboratory to the Clinic
like behaviour was to pair the light with an electric shock (the
The Needfor an Effective OCD Treatment shock occurred 10 seconds after the light was turned on). The
Prior to the 1970s and 1980s, treatment for OCD consisted dog soon learned to jump into the other compartment of the
largely of psychodynamic psychotherapy derived from psy- shuttle box, which was not electrified, once he had received
choanalytic ideas of unconscious motivation. Unfortunately, the shock. After several trials, the dog learned to successfully
there are virtually no scientific studies assessing the efficacy avoid the shock by jumping to the nonelectrified compartment
of such an approach. However, the general consensus among in response to the light (that is, within 10 seconds). In other
clinicians of that era was that OCD was an unmanageable con- words, the experimenter produced a conditioned response to
dition with a poor prognosiswhich clearly demonstrates the light, namely, jumping from one compartment ofthe box
how much confidence (or perhaps how little) clinicians placed to the other.
in psychodynamic psychotherapy for treating OCD, Indeed,
available reports suggest that the effects of psycho- Once this conditioned response was established, the electric-
dynamically oriented therapies are neither robust nor durable ity was disconnected, and the dog never received another
for OCD or for anxiety disorders in general. shock. Nevertheless, the animal continued to jump across the
By the last quarter ofthe 20th century, however, the prognos- hurdle each time the conditioned stimulus (that is, the light)
tic picture for OCD had improved drastically. This was was turned on. This continued for hundreds (and in some
largely because of the work of Victor Meyer (5) and other cases, thousands) of trials, despite no actual risk of shock.
Apparently, the dog had acquired an obsessive-compulsive
Abbreviations used In this article habitjumping across the hurdlewhich reduced his condi-
CBT cognitive-behavioural therapy
tioned fear of shock and thus was maintained by negative rein-
forcement (the removal of an aversive stimulus such as
CT cognitive therapy
emotional distress). This serves as an animal analogue to
EE expressed emotion
human OCD, where compulsive behaviour is triggered by fear
ERP exposure and response prevention
associated with situations or stimuli such as toilets, fioors, or
OCD obsessive-compulsive disorder obsessional thoughts (conditioned stimuli) that pose little or
RCT randomized controlled trial no actual risk of harm. This fear is then reduced by avoidance
SD standard deviation and compulsive rituals (for example, washing) that serve as an
Y-BOCS Yale-Brown Obsessive Compulsive Scale escape from distress and, in doing so, are negatively
reinforced (that is, they become habitual).

408 - Can J Psychiatry, Vol 51, No 7, June 2006

 The Psychological Treatment of Obsessive-Compulsive Disorder

Solomon and his colleagues also attempted to reduce the com- only 2 of those who were successfully treated had
pulsive jumping behaviour of their "obsessive-compulsive" relapsed (8).
dogs, using various techniques, the most effective of which
Contemporary ERP entails therapist-guided, systematic,
involved a combination of procedures now known as ERP.
repeated, and prolonged exposure to situations that provoke
Specifically, the experimenter turned on the conditioned stim-
obsessional fear, along with abstinence from compulsive
ulus (light), an in vivo exposure technique, and increased the
behaviours. This can occur in the form of repeated actual con-
height of the hurdle in the shuttle box so that the dog was
frontations with feared low-risk situations (that is, in vivo
unable to jump (response prevention). When this was done,
exposure) or in the form of imagined confrontation with the
the dog immediately showed signs of a strong fear response by
feared disastrous consequences of low-risk situations (that is,
running around the chamber, jumping on the walls, defecat-
imaginal exposure). For example, an individual who fears
ing, urinating, and yelping. Clearly, he expected to receive a
being held responsible for harm if he or she writes the number
shock. Gradually, however, this emotional reaction subsided
666, or for a robbery if he or she leaves home without double-
until finally, the dog displayed calmness without the slightest
checking that the door is locked, would practise writing 666 or
hint of distress. In behavioural terms, this experimental para-
practise leaving home after rapidly closing and locking the
digm produced extinction. After several such extinction trials,
front door. The individual would also practise imagining
the entire emotional response (that is, fear of shock) was
being held responsible for harming others or causing a
extinguished, such that, even when the light was turned on and
burglary because of these exposure tasks.
the height of the hurdle was lowered, the dog did not jump.
Refraining from compulsive rituals (response prevention) is a
During the 1960s and 1970s, behaviourally oriented research- vital component of treatment because the performance of such
ers became interested in adapting similar treatment paradigms rituals to reduce obsessional anxiety would prematurely dis-
to human beings with OCD (7). Of course, no electric shocks continue exposure and rob the patient of learning, first, that
were used, but the adaptation was as follows: After they pro- the obsessional situation is not truly dangerous and, second,
vided informed consent, patients with OCD handwashing rit- that anxiety subsides on its own even if the ritual is not per-
uals were seated at a table with a container of dirt and formed. Thus successful ERP requires the patient to remain in
miscellaneous garbage. Placing his own hands in the mixture, the exposure situation until the obsessional distress decreases
the experimenter asked the patient to do the same and spontaneously, without attempting to reduce the distress by
explained that he or she would not be permitted to wash his or withdrawing from the situation or by performing compulsive
her hands for some length of time. When the patient began the rituals or neutralizing strategies.
procedure, an increase in anxiety, fear, and urges to wash his
or her hands, was (of course) observed. This increase in dis- Contemporary ERP can be delivered in several ways. One
tress was conceptualized as akin to the dogs' response when highly successful format comprises a few hours of assessment
the light was turned on and the hurdle had been increased in and treatment planning followed by 16 twice-weekly treat-
height to make jumping impossible. However, like the dogs, ment sessions lasting about 90 to 120 minutes each and spaced
the patients eventually demonstrated a substantial reduction over about 8 weeks (9). Generally, the therapist supervises the
in fear and in the urge to wash, thus demonstrating therapeutic exposure sessions and assigns self-exposure practice to be
extinction. This procedure was repeated on subsequent days, completed by the patient between sessions. Depending on the
the hypothesis predicting that, after some time, extinction patient's symptom presentation and the practicality of con-
would be complete and OCD symptoms would be reduced. fronting actual feared situations, treatment sessions involve
varying amounts of actual and imaginal exposure practice.
From the Laboratory to the Clinic A course of ERP typically begins with the assessment of
Meyer is credited with being the first to report a study of the obsessional thoughts, ideas, and impulses; of stimuli that trig-
effects of ERP treatment for OCD (5). He persuaded patients ger the obsessions; of rituals and avoidance behaviour; and of
hospitalized with OCD to deliberately confront, for 2 hours the anticipated harmful consequences of confronting feared
each day, situations and stimuli they usually avoided (for situations without performing rituals (that is, the cognitive
example, floors or bathrooms). The purpose of confrontation links between obsessions and compulsions). Before actual
was to induce obsessional fears and urges to ritualize. The treatment commences, the therapist socializes the patient to a
patients were also instructed to refrain from performing com- psychological model of OCD based on the principles of learn-
pulsive rituals (for example, washing or checking) after expo- ing and emotion (for example, 10). The patient is also given a
sure. Ten of Meyer's 15 patients responded extremely well to clear rationale for how ERP is expected to help reduce OCD.
this therapy, and the remainder showed partial improvement. This psychoeducational component is an important step in
Follow-up studies conducted several years later found that therapy because it helps to motivate the patient to tolerate the

Can J Psychiatry, Vol 51, No 7, June 2006 409

The Canadian Journal of Psychiatryin Review

distress that typically accompanies exposure practice. A help- From a cognitive perspective, ERP is effective because it cor-
ful rationale includes information about how ERP involves rects dysfunctional beliefs (such as overestimates of threat)
the provocation and reduction of distress during prolonged that underlie OCD symptoms by presenting patients with
exposure. Information gathered during the assessment ses- information that disconfirms these beliefs. For example, when
sions is then used to plan, collaboratively with the patient, the a patient confronts feared situations and refrains from rituals,
specific exposure exercises that will be pursued. he or she finds out that obsessional fear declines naturally
(habituation) and that feared negative consequences are
In addition to explaining and planning a hierarchy of exposure unlikely to occur. This evidence is processed and incorpo-
exercises, the educational stage of ERP must also acquaint rated into the patient's belief system. Thus compulsive rituals
patients with response prevention procedures. Importantly, to reduce anxiety and prevent feared disasters become
the term "response prevention" does not imply that therapists unnecessary (redundant).
actively prevent patients from performing rituals. Instead,
therapists must convince patients to resist urges to perform rit- Finally, ERP helps patients gain self-efficacy by helping them
uals on their own. Self-monitoring of rituals is often used in to master their fears without having to rely on avoidance or
support of this goal. safety behaviours. The importance of this sense of mastery is
an often-overlooked effect of ERP.
The exposure exercises typically begin with moderately dis-
tressing situations, stimuli, and images and escalate to the
Foa and Kozak have drawn attention to 3 indicators of change
most distressing situations, which must be confronted during
during exposure-based treatment (11). First, physiological
treatment. Beginning with exposure tasks that evoke less anx-
arousal and subjective fear must be evoked during exposure.
iety increases the likelihood that patients will learn to manage
Second, fear responses gradually diminish during the expo-
their distress and complete the exposure exercise success-
sure session (within-session habituation). Third, the initial
fully. Moreover, success with initial exposures increases con-
fear response at the beginning of each exposure session
fidence in the treatment and helps motivate patients to
declines across sessions (between-session habituation).
persevere during later, more difficult exercises. At the end of
each treatment session, the therapist instructs the patient to
continue exposure for several hours alone and in different
Assessment of OCD Symptoms:
environmental contexts. Exposure to the most anxiety-
The Yale-Brown Obsessive Compulsive Scale
evoking situations is not left to the end of the treatment but,
rather, is practised about mid-way through treatment. This The use of assessment instruments that are psychometrically
tactic allows patients ample opportunity to repeat exposure to reliable, valid, and sensitive to change is important in assuring
the most difficult situations in different contexts that allow that any improvement in symptoms is really attributable to the
generalization of treatment effects. During the later treatment treatments and not to fiuctuations in a poor assessment instru-
sessions, the therapist emphasizes the importance of the ment. The Y-BOCS (12,13), a semistructured clinical inter-
patient's continuing to apply the ERP procedures learned view, is considered the gold standard measure of OCD
during treatment. symptoms. Owing to its respectable psychometric proper-
ties (14), the Y-BOCS is widely used in OCD treatment-out-
come research; it thus provides an excellent measure by which
ERP Mechanisms of Action to compare the results of treatment across studies. Therefore,
Three mechanisms are thought to be involved in the reduction it is important to briefly discuss what scores on the Y-BOCS
of obsessions and compulsions during ERP; a behavioural indicate clinically before I review the treatment-outcome
mechanism, a cognitive mechanism, and changes in self- literature.
efficacy. From a behavioural perspective, ERP is effective
because it provides an opportunity for the extinction of condi- When administering the Y-BOCS, the interviewer rates the
tioned fear responses. Specifically, repeated and uninter- following parameters for both obsessions (items 1 to 5) and
rupted exposure to feared stimuli produces habituationan compulsions (items 6 to 10); time, interference with function-
inevitable natural decrease in conditioned fear. Response pre- ing, distress, resistance, and control. Items are rated on a scale
vention fosters habituation by blocking the performance of ranging from 0 (no symptoms) to 4 (extreme). The total score
anxiety-reducing rituals that would foil the habituation pro- is the sum ofthe 10 items and therefore ranges from 0 to 40.
cess. Extinction of conditioned anxiety occurs when the Y-BOCS scores of 0 to 7 indicate subclinical OCD, 8 to 15
once-feared obsessional stimulus is repeatedly paired with the indicate mild symptoms, 16 to 25 indicate moderate symp-
nonoccurrence of feared consequences and the eventual toms, 26 to 35 indicate severe symptoms, and 36 to 40 indicate
reduction of anxiety. extreme severity.

410 Can J Psychiatry, Vol 51, No 7, June 2006

 The Psychological Treatment of Obsessive-Compulsive Disorder

The Efficacy of ERP and evidence-based challenging and correction of faulty and
Over the last 30 years, numerous investigations of ERP for dysfiinctional thoughts and beliefs thought to underlie obses-
treating OCD have been conducted worldwide. Studies com- sional fear (25), Specifically, cognitive models of OCD begin
pleted in England (15), Holland (16), Greece (17), and the with the well-established finding that intrusions (thoughts,
United States (18), with over 500 patients and numerous dif- images, and impulses that intrude into consciousness, such as
ferent therapists, have afftrmed the generalizability of ERP's unwanted thoughts of harming a loved one) are experienced
beneftcial effects, RCTs have provided particularly strong by most people (normal obsessions) but can develop into
evidence ofthe superiority of ERP over credible control thera- obsessions when appraised as posing a threat for which the
pies such as progressive muscle relaxation training (for exam- individual is personally responsible (10), For example, an
ple, 19), anxiety management training (20), and pill OCD patient might think that "Having thoughts about harm-
placebo (21), Intensive ERP has also been found more effec- ing Mother means I'm a dangerous person who must take
tive than the antidepressant clomipramine, believed to be the extra care to ensure that I don't lose control," Such appraisals
most effective form of pharmaeotherapy for OCD (21), For evoke distress and motivate the individual to try to suppress or
patients receiving ERP, Y-BOCS reductions typically exceed remove the unwanted intrusion (for example, by replacing it
50% to 60%, and posttreatment scores average between 9 and with a "good" thought) and to try to prevent any harmful
13, indicating mild residual symptoms. Importantly, despite events associated with the intrusion (for example, by avoiding
this clinically significant improvement in symptoms, patients driving).
rarely achieve complete symptom reduction with ERP (22),
According to the cognitive model, compulsions are conceptu-
Treatment of Pure Obsessionals alized as efforts to remove intrusions and to prevent any per-
Whereas many individuals with OCD exhibit overt compul- ceived harmful consequences, Salkovskis advanced 2 main
sive rituals (for example, washing or checking), a substantial reasons to explain why compulsions become persistent and
subset report mental rituals and other subtle anxiety-reduction excessive (10), First, they are reinforced by immediate dis-
strategies that are difficult to distinguish from obsessional tress reduction and by temporary removal of the unwanted
(anxiety-evoking) phenomena. These patients, often labelled thought (negative reinforcement, as in the conditioning mod-
as "pure obsessionals," were once considered nonresponsive els of OCD), Second, they prevent the individual from learn-
to cognitive-behavioural treatments (23), Recently, however, ing that his or her appraisals are unrealistic (for example, the
Freeston and colleagues conducted an RCT in which they individual fails to leam that unwanted harm-related thoughts
compared subjects receiving a form of ERP, developed spe- do not lead to acts of harm). Compulsions infiuence the fre-
cifically for OCD without overt rituals, with a waiting-list quency of intrusions by serving as reminders of intrusions and
control group (24), This treatment primarily involved thereby triggering their reoccurrence. For example, compul-
repeated exposure to descriptions of obsessional thoughts (via sive handwashing can remind the individual that he or she
audiotapes) and abstinence from mental ritualizing. Mean may have become contaminated. Attempts at distracting one-
pre- and posttreatment Y-BOCS scores for the ERP group self from unwanted intrusions may paradoxically increase
were 25,1 and 12,2, respectively. As expected, there was no their frequency, possibly because the distractors become
improvement in the wait-list group, whose Y-BOCS pre- and reminders (retrieval cues) ofthe intrusions. Compulsions can
posttreatment scores were 21,2 and 22,0, respectively. Impor- strengthen one's perceived responsibility. That is, the absence
tantly, at 3-month follow-up, ERP patients had maintained of the feared consequence after performing the compulsion
their gains, as evidenced by a mean Y-BOCS score of 10,8, reinforces the belief that the individual is responsible for
These results demonstrate that ERP procedures may be effec- removing the threat.
tively varied to accommodate the heterogeneous phenomen-
ology of OCD patients, including those with obsessional Although Salkovskis emphasizes the importance of responsi-
symptoms and mental rituals. bility appraisals and beliefs (10), several cognitive-
behavioural theorists have proposed that other types of
Cognitive Therapy for OCD dysflinctional beliefs and appraisals are also important in
OCD (25), Thus contemporary cognitive-behavioural theo-
Basis ofCT ries have extended the work of Salkovskis to propose that var-
Given the challenges of ERP (that is, high levels of anxiety ious types of dysfiinctional beliefs and appraisals, in addition
produced during exposure), some clinicians and researchers to those pertaining to responsibility, play an important role in
have turned to CT approaches that incorporate less prolonged OCD's etiology and maintenance. Although contemporary
exposure to fear cues and that have led to advances in the treat- belief and appraisal models differ from one another in some
ment of other anxiety disorders. The basis of CT is the rational ways, their similarities generally outweigh their differences.

Can J Psychiatry, Vol 51, No 7, June 2006 < 411

The Canadian Journal of PsychiatryIn Review

Table 1 Domains of dysfunctional beliefs associated with OCD

Belief domain Description

Excessive responsibility Belief that one has the special power to

cause, and (or) the duty to prevent,
negative outcomes
Overimportance of thoughts Belief that the mere presence of a
thought indicates that the thought is
significant (for example, the belief that
the thought has ethical or moral
ramifications or that thinking the thought
increases the probability of the
corresponding behaviour or event)
Need to control thoughts Belief that complete control over one's
thoughts is both necessary and possible
Overestimation of threat Belief that negative events are
especially likely and would be especially
awful
Perfectionism Belief that mistakes and imperfection are
intolerable
Intoierance for uncertainty Belief that it is necessary and possible to
be completely certain that negative
outcomes will not occur

A major contemporary cognitive model is that developed by patients recognize and correct dysfunctional thinking pat-
the Obsessive Compulsive Cognitions Working Group (26- tems. Behavioural experiments, in which patients enter and
28), This intemational group of over 40 investigators shares observe situations that exemplify their fears, are often used to
an interest in understanding the role of cognitive factors in facilitate the collection of information that will allow patients
OCD, Extending the work of Salkovskis and others, they have to revise their judgments about the degree of risk associated
reached a consensus regarding the most important underlying with obsessions. Although the rationale for behavioural
beliefs in OCD (26), They identified responsibility beliefs and experiments in CT is somewhat different fi-om the rationale
other belief domains (listed in Table 1) that were said to give for exposure exercises in ERP, there is often procedtiral over-
rise to corresponding appraisals. Two self-report measures lap, and fundamental differences between the 2 techniques
the Obsessional Beliefs Questionnaire and the Interpretations may be difficult to discem,
of Intmsions Inventorywere developed to assess these
domains (27), A few specific cognitive techniques used in the treatment of
OCD are as follows: Where patients overestimate personal
Erom Theory to Practice
responsibility, the "pie technique" (25) has them give an ini-
Typically, at the beginning of CT, the therapist presents a
tial estimate ofthe percentage of responsibility that would be
rationale for treatment incorporating the notion that intmsive
attributable to them if a feared consequence were to occur.
obsessional thoughts are normal experiences and not harmftil
The patient then generates a list ofthe parties (other than him-
or indicative of anything important. Rather, OCD arises
self or herself) who would also have some responsibility for
because the patient appraises the intrusions as significant in a
the feared consequence. The patient then draws a pie chart,
way that is distressing (for example, "Thoughts of violence
each slice of which represents one ofthe responsible parties
are equivalent to committing violent acts"), Misappraisal of
identified. Next, the patient labels all parties' slices according
intrusions in this way leads to preoccupation with the
to their percenage of responsibility and labels his or her own
unwanted thought as well as with responses, such as avoid-
slice last. By the exercise's end, it is generally clear to patients
ance and compulsive rituals, that unwittingly maintain the
that most of the responsibility for the feared event would not
obsessional preoccupation and anxiety (10),
be their own. For patients with difficulty discriminating
Various techniques are used to help patients correct their erro- between unwanted obsessional thoughts and actions, the cog-
neous beliefs and appraisals, such as didactic presentation of nitive continuum technique has them rate how immoral they
educational material and Socratic dialogue aimed at helping perceive themselves to be for having the intrusive obsessional

412 Can J Psychiatry, Vol 51, No 7, June 2006

 The Psychological Treatment of Obsessive-Compulsive Disorder

thoughts. Next, patients rate the morality level of other indi- outcome was likely attenuated by the use of suboptimal proce-
viduals who have committed acts of varying degrees of immo- dures (for example, lack of therapist-supervised exposure),
rality (for example, a serial rapist or abusive parents). Then, and CT programs were possibly enhanced by behavioural
patients again rate themselves and reevaluate how immoral experiments that likely have effects similar to unsupervised
they are for simply experiencing intrusive thoughts. exposure. Using metaanalytic methods, we found that behav-
ioural experiments improve the efficacy of CT for OCD (34).
CT Compared With ERP In later studies that incorporate in-session exposure within
Four studies that used the Y-BOCS have compared CT with ERP protocols, the treatment based on behavioural theory
variations of ERP. van Oppen and others randomly assigned appears to be superior to CT.
patients to either 16 sessions of CT or 16 sessions of self-
controlled ERP (all exposure was conducted by the patient Adding CT to ERP
without therapist supervision) (29). Both treatments led to an To examine whether adding elements of CT would improve
improvement in OCD symptoms, although CT was more response to ERP, Vogel, Stiles, and Gotestam (35) conducted
effective than ERP (Y-BOCS reductions were 53% and 43%, a controlled study in which 35 individuals with OCD were
respectively). Importantly, the brief and infrequent therapist randomly assigned to receive either ERP plus CT ( = 16) or
contact (weekly 45-minute sessions), along with reliance on ERP plus relaxation therapy {n = 19). Relaxation therapy was
patients to manage all exposure practice on their own, likely added as a placebo procedure to control for the effects of add-
accounted for the relatively modest effects of ERP in this ing additional techniques to ERP. Results indicated that both
study. Moreover, CT involved behavioural experiments that therapy programs were superior to the wait-list condition.
resembled exposure, which blurred the distinction between Among treatment completers, Y-BOCS scores were reduced
the 2 treatments. Only after behavioural experiments were from 25.1 (ERP + CT) and 23.4 (ERP + relaxation) at pretreat-
introduced (at the sixth session) did symptom reduction in the ment to 16.4 (ERP + CT) and 11.3 (ERP + relaxation) at
CT group approach that of ERP. Thus it is possible that the posttreatment and to 13.3 (ERP + CT) and 10.2 (ERP + relax-
exposure component of behavioural experiments is key to the ation) at 1-year follow-up. Statistical analyses indicated a
efficacy of CT. Using a sample that overlapped with the van nonsignificant trend toward superiority of ERP + relaxation
Oppen and others study, van Balkom and others found no sig- therapy at posttreatment, but this difference disappeared at the
nificant difference between CT with behavioural experiments follow-up assessment. In addition, the inclusion of CT was
and self-controlled ERP (30). useful in reducing dropout. Thus there appear to be benefits to
incorporating CT techniques along with ERP.
In a study by Cottraux and others , CT involving 20 hours of
therapist contact over 16 weeks was compared with a similar Clinically speaking, just as exposure adds to the benefits of
regimen of therapist-supervised and homework ERP (31). CT (34), CT probably adds to the effects of ERP. It is unfortu-
The 2 treatments produced comparable outcomes at nate that most published accounts of ERP (for example, 36)
posttreatment (Y-BOCS reductions of 42% to 44%). At fail to fully describe the informal cognitive procedures that
1-year follow-up, patients treated with ERP improved further likely contribute to its efficacy. For example, during ERP,
from their posttreatment status (follow-up mean Y-BOCS patients often need to be persuaded that exposure to fear cues
score, 11.1), whereas this was not the case with CT (follow-up will be beneficial for them. This typically requires discussion
mean Y-BOCS score, 15.0). of fear-related beliefs and assumptions (for example, overes-
timates of danger). Nonetheless, although CT techniques are
Finally, McLean and others compared the 2 treatment
important during ERP, the research suggests that CT should
approaches as conducted in group settings (32). Patients
accompany, rather than replace, systematic prolonged and
received 12 weekly, 2.5-hour group sessions (with 6 to 8 par-
repeated therapist-supervised ERP. That is, cognitive inter-
ticipants per group) of either CT or ERP involving in-session
ventions are best used to "tenderize" distorted cognitions that
and homework exposures. Both treatments were more effec-
underlie obsessional fears, thereby creating the conditions for
tive than a wait-list condition, and ERP was associated with
patients to comply with ERP procedures.
greater improvement than CT at both posttreatment (40% and
27% Y-BOCS score reductions, respectively) and follow-up
Effectiveness Research
(21% and 41% Y-BOCS score reductions, respectively).
Although RCTs have yielded sound evidence that ERP
Although some earlier studies suggest that ERP and CT have reduces OCD symptoms, these studies employ highly selec-
similar efficacy for OCD (for example, 33), interpretation of tive patient samples that are not necessarily representative of
these results as indicating equivalent success for these 2 types the typical treatment-referred OCD patient. For example,
of treatment is questionable because both treatments yielded despite the high frequency with which comorbid conditions
minimal improvements in most of these early studies. ERP exist in patients with OCD, individuals with comorbid

Can J Psychiatry, Vol 51, No 7, June 2006 413

The Canadian Journal of PsychiatryIn Review

disorders (for example, Axis II disorders or major depression) articulated fears of disasters tended to improve more than
are usually excluded from RCTs. Effectivetiess studies are those who did not articulate these kinds of fears
designed to address these methodological shortcomings and (posttreatment Y-BOCS mean scores were 8.2 and 14.9,
examine the effects of treatments in highly representative respectively). Although this difference was not statistically
patient samples treated in general clinical settings. The aim of significant (P = 0.06), the authors concluded that the inability
effectiveness research is therefore to bridge the gaps between to articulate feared consequences of exposure decreases the
research and clinical practice. therapist's ability to contrive exposure exercises that provide
disconfirming information. This, in turn, could hinder treat-
The largest effectiveness study of ERP included 110 patients
ment with ERP.
treated on a fee-for-service hasis in an outpatient setting (37).
Treatment included 3 weeks of daily 2-hour ERP sessions. No Of the 11 patients with articulated feared consequences in the
patients were excluded for reasons of age, comorhidity, previ- Foa and others study, 5 showed poor insight into the irratio-
ous treatment failure, or medical problems. In fact, one-half of nality of their obsessional fears (39). A comparison of out-
the sample had comorbid Axis I or Axis II diagnoses, and 61 % come indicated that these individuals showed a poorer
were also using serotonergic medication. Patients were only response to ERP, relative to patients who showed good insight
excluded from the study if they were suffering from active that their obsessional fears were senseless. To explain this
psychosis, substance abuse, or suicidal ideation (all these con- finding, Foa and others speculated that patients with poor
ditions are contraindications for ERP in any setting; 38). An insight have difficulty learning information that is inconsis-
intent-to-treat analysis indicated considerable improvement: tent with their OCD beliefs. Alternatively, because of their
mean Y-BOCS scores improved from 26.79, SD 4.89, to extreme fear, these patients may not adhere to ERP instruc-
11.81, SD 7.30. This is equivalent to a 60% reduction in OCD tions as closely as do patients with better insight. The findings
symptoms. This study suggests that the effects of ERP extend from this small study highlight the importance of assessing
beyond the highly selected patient samples treated in research OCD patients' degree of insight into the senselessness of their
studies. symptoms. Although more studies with larger samples are
necessary, insight is likely an important prognostic indicator
Predictors of Improvement of response to ERP.
While ERP is effective for most OCD patients who receive
this treatment, about 25% to 30% of patients who begin ERP Comorbid Depression
drop out of therapy prematurely. Among those who remain in Depression often coexists with OCD (4). Using a large sample
treatment, about 80% respond well, yet 20% or more do not. of 87 patients, Abramowitz and colleagues examined the
Therefore, about 50% of patients referred with OCD are not effects of comorbid depressive symptoms on ERP out-
significantly improved with ERP, and it is important to con- come (40). They divided patients into groups without depres-
sider this alongside the impressive data for ERP's effective- sion, with mild depression, with moderate depression, and
ness. Substantial effort has recently gone into investigating with severe depression on the basis of pretreatment scores on
factors that might predict poor treatment response. Below, I the Beck Depression Inventory (41). Results indicated attenu-
describe some of the recent research on predictors of ERP out- ated outcome only for the group with the most severe depres-
come and consider the following variables: insight into OCD sion. The authors suggested that, because of their high
symptoms, depression, and family expressed emotion. emotional reactivity, individuals with severe depression fail
to undergo the decrease in anxiety-distress that occurs fol-
Insight
lowing extended exposure to feared stimuli. Thus they do not
In the DSM-IV Field Trial for OCD, Foa and colleagues found
have the therapeutic experience of feeling comfortable in the
that some patients hold strongly fixed beliefs that their obses-
presence of feared stimuli and therefore fail to leam that
sional fears are realistic and that compulsive rituals are neces-
obsessive doubts are unrealistic. Motivational difficulties,
sary to prevent disastrous consequences (2). Foa,
which often accompany depression, may also account for
Abramowitz, Franklin, and Kozak examined whether the
poor treatment outcome.
presence of such fixed beliefs is related to treatment outcome
with ERP (39). In their study, 20 OCD patients received an Abramowitz and Foa compared outcome of ERP for OCD
intensive (daily) 3-week ERP program. Eleven patients artic- patients with and without a comorbid diagnosis of major
ulated specific obsessional fears of disastrous consequences depression (42). They found that, although the presence of
(for example, that if they did not touch all 4 walls, their parents major depression was not related to treatment failure per se,
would die), whereas 9 did not. At pretreatment, mean patients without depression had significantly lower
Y-BOCS scores did not differ between these 2 groups (overall posttreatment and follow-up Y-BOCS scores than did patients
mean 25.20). However, at posttreatment, patients with with an additional diagnosis of depression.

414 Can J Psychiatry, Vol 51, No 7, June 2006

 The Psychological Treatment of Obsessive-Compulsive Disorder

Expressed Emotion Funding and Support

The way in which family members respond to a relative with No funding or other support was received for this review.
OCD (or any problem) is called "expressed emotion."
Researchers have conceptualized EE as including emotional References
overinvolvement, hostility, and perceived criticism. To deter-
1. Abramowitz JS, Franklin M, Schwartz S, Furr J. Symptom presentation and
mine whether these family interaction variables play a role as outcome of cognitive-behavioral therapy for obsessive-compulsive disorder.
predictors of response to ERP, Chambless and Steketee con- J Consult Clin Psychol 2003;71:1049-57.
2. Foa EB, Kozak MJ, Goodman WK, Hollander E, Jenike MA, Rasmussen S.
ducted a study in which patients with OCD ( = 60) and DSM-IV field trial: obsessive-compulsive disorder. Am J Psychiatry
patients with agoraphobia ( = 41), together with their fami- 1995;152:90-6.
3. Kamo M, Golding JM, Sorenson SB, Btimam MA. The epidemiology of
lies, were administered measures of EE prior to beginning obsessive-compulsive disorder in five U.S. communities. Arch Gen Psychiatry
treatment (43). The most consistent predictor of negative 1988;45{Suppl 3):3-14.
4. Ricciardi JN, McNally RJ. Depressed mood is related to obsessions but not
treatment outcome was hostility: when relatives were hostile compulsions in obsessive-compulsive disorder. J Anxiety Disord 1995;9:249-56.
to the identified patient, the odds of premature termination 5. Meyer V. Modification of expectations in cases with obsessional rituals. Behav
Res Ther 196;4:273-80.
were about 6 times greater than when relatives were not hos- 6. Solomon RL, Kamin LJ, Wynne LC. Tratimatic avoidance learning: the
tile. Hostility was also associated with poorer response in outcomes of several extinction procedures with dogs. J Abnorm Soc Psychol
195;48:291-302.
patients who completed treatment. Interestingly, once hostil- 7. Rachman S, Hodgson R. Obsessions and compulsions. Englewood Cliffs (NJ):
ity was statistically controlled, criticism had a positive effect. Prentice-Hall; 1980.
8. Meyer V, Levy R, Schnurer A. The behavioral treatment of
This suggests that, when relatives express dissatisfaction with obsessive-compulsive disorders. In: Beech HR, editor. Obsessional states.
patients' symptoms but do not express personal rejection, crit- London (UK): Methuen; 1974. p 233-58.
9. Abramowitz JS, Foa EB, Franklin ME. Exposure and ritual prevention for
icism may have motivational properties that enhance treat- obsessive-compulsive disorder: effects of intensive versus twice-weekly
sessions. J Consult Clin Psyehol 2003:71:394-8.
ment response. This underscores the importance of educating 10. Salkovskis PM. Cognitive-behavioral approaches to the understanding of
family members about OCD and about how to therapeutically obsessional problems. In: Rapee R, editor. Current controversies in the anxiety
disorders. New York (NY): Guilford: 1996. p 103-33.
assist with ERP exercises during treatment. 11. Foa E, Kozak M. Emotional processing of fear: exposure to corrective
information. Psychol Bull 1986;99:20-35.
12. Goodman WK, Price LH, Rasmussen SA, Mazure C, Fleischmann RL, Hill CL,
and others. The Yale-Brown Obsessive Compulsive Scale, I: development, use,
Summary and Future Directions and reliability. Arch Gen Psychiatry 1989;46:1006-l 1.
Recent research has demonstrated that CBT using ERP is the 13. Goodman WK, Price LH, Rasmussen SA, Mazure C, Delgado P, Heninger GR,
and others. The Yale-Brown Obsessive Compulsive Scale, II: validity. Arch Gen
most effective short- and long-term treatment for OCD. These Psychiatry 1989;46: 1012-6.
encouraging findings notwithstanding, full remission is not 14. Taylor S. Assessment of obsessions and compulsions: reliability, validity, and
sensitivity to treatment effects. Clin Psychol Rev 1995;15:261-96.
the standard. Evidence from recent effectiveness studies sug- 15. Rachman S, Hodgson R, Marks I. The treatment of chronic
obsessive-compulsive netirosis. Behav Res Ther 1971;9:237-47.
gests that ERP is transportable to nonresearch settings and 16. Boersma K, den Hengst S, Dekker J, Emmelkamp P. Exposure and response
therefore should be a first-line treatment modality for OCD in prevention in the naturla environment: a comparison with obsessive-compulsive
patients. Behav Res Ther 1976;14:19-24.
all settings. Although response to treatment is highly variable, 17. Rabavilas A, Boulougouris J, Stefanis C. Duration of flooding sessions in the
we are beginning to uncover factors that may reliably predict treatment of obsessive-compulsive patients. Behav Res Ther 1976; 14:349-55.
18. Foa EB, Goldstein A. Continuous exposure and complete response prevention in
poorer response, such as poor insight into the senselessness of the treatment of obsessive-compulsive neurosis. Behav Ther 1978;9:821-9.
obsessional fears, severe depression, and family hostility. 19. Fals-Stewart W, Marks AP, Schafer J. A comparison of behavioral group therapy
and individual behavior therapy in treating obsessive-compulsive disorder.
JNervMentDis 1993;181:189-93.
Although the research to date has addressed many critical 20. Lindsay M, Crino R, Andrews G. Controlled trial of exposure and response
issues in the treatment of OCD, important topics still require prevention in obsessive-compulsive disorder. Br J Psychiatry 1997; 171:135-9.
21. Foa EB, Liebowitz MR, Kozak MJ, Davies S, Campeas R, Franklin ME, and
further study. For example, a treatment program that includes others. Treatment of obsessive-compulsive disorder by exposure and ritual
training family members about OCD, its treatment, and how prevention, clomipramine, and their combination: a randomized, placebo
controlled trial. Am J Psychiatry 2005;162:151-6I.
to effectively assist with a loved one's therapy would be use- 22. Abramowitz JS. Does cognitive-behavioral therapy cure obsessive-compulsive
disorder? A meta-analytic evaluation of clinical significance. Behav Ther
ful, given the high prevalence of relational problems in fami- 1998:29:339-55.
lies of OCD patients. Motivation to begin treatment, 23. Jenike MA. Obsessive-compulsive disorder: efficacy of specific treatments as
assessed by controlled trials. Psychopharmacol Bull 1993;29:487-99.
especially given the anxiety-evoking nature of ERP, is often a 24. Freeston MH, Ladouceur R, Gagnon F, Thibodeau N, Rheaume J, Letarte H, and
problem. Thus readiness programs, in which patients read others. Cognitive-behavioral treatment of obsessive thoughts: a controlled study.
J Consult Clin Psychol 1997;65:405-13.
case histories or discuss treatment with former patients, might 25. Clark DA. Cognitive-behavioral therapy for OCD. New York (NY): Guilford
decrease refusal rates and increase treatment compliance. Press; 2004.
26. Obsessive Compulsive Cognitions Work Group. Cognitive assessment of
From the clinician's perspective, providing successful ERP obsessive-compulsive disorder. Behav Res Ther 1997;35:667-8I.
can be a challenge, and very few centres offer the training 27. Obsessive Compulsive Cognitions Working Group. Psychometric validation of
the Obsessive Beliefs Questionnaire and the Interpretation of Intrusions
needed to become proficient in these procedures. Therefore, InventoryPart 2: factor analyses and testing ofa brief version. Behav Res Ther
development of programs for psychology and psychiatry 2005;43:1527^2.
28. Frost R, Steketee G, editors. Cognitive approaches to obsessions and
trainees might also improve access to this effective therapy. compulsions. Amsterdam (NE): Pergamon; 2002.

Can J Psychiatry, Vol 51, No 7, June 2006 415

The Canadian Journal of Psychiatryin Review

29. van Oppen P, De Haan E, van Balkom AJLM, Spinhoven P, Hoogduin K, van 38. Abramowitz JS, Schwartz SA. Treatment for obsessive-compulsive disorder:
Dyck R. Cognitive therapy and exposure in vivo in the treatment of obsessive deciding what method for whom. Brief Treatment and Crisis Intervention
compulsive disorder. Behav Res Ther 1995;33:379-90. 2003:3:261-73.
30. van Balkom AJLM. De Haan E, van Oppen P. Spinhoven P, Hoogduin KAL, 39. Foa EB. Abramowitz JS, Franklin ME, Kozak MJ. Feared consequences, fixity
van Dyck R. Cognitive and behavioral therapies alone versus in combination of belief, and treatment outcome in patients with obsessive-eompulsive disorder.
with fluvoxamine in the treatment of obsessive compulsive disorder. J Nerv Behav Ther 1999;30:717-24.
MentDisord I998;186:492-9. 40. Abramowitz J, Franklin M, Street G, Kozak M, Foa E. Effects of eomorbid
31. Cottraux J, Note I, Yao SN, Lafont S, Note B, MoUard E, and others. depression on response to treatment for obsessive-eompulsive disorder. Behav
A randomized controlled trial of cognitive therapy versus intensive behavior Ther2000;31:5I7-28.
therapy in obsessive compulsive disorder. Psychothcr Psychosom 41. Beck AT, Ward CH, Medelsohn M, Mock J, Eribaugh J. An inventory for
2001;70:288-97. measuring depression. Arch Gen Psychiatry !961;4:561-71.
32. McLean PD, Whittal ML, Thordarson DS, Taylor S, Soehting 1, Koch WJ, and
42. Abramowitz J, Foa E. Does comorbid major depressive disorder influence
others. Cognitive versus behavior therapy in the group treatment of
outcome of exposure and response prevention for OCD? Behav Ther
obsessive-eompulsive disorder. J Consult Clin Psychol 2001;69:205-14.
2000;31:795-800.
33. Bmmelkamp PMG, Bcens H. Cognitive therapy with obsessive-compulsive
43. Chambless DL, Steketee. Expressed emotion and behavior therapy outcome:
disorder: a comparative evaluation. Bchav Res Ther 1991;29:293-300.
a prospective study with obsessive-compulsive and agoraphobic outpatients.
34. Abramowitz JS, Franklin ME, Foa EB. Empirical status of cognitive-behavioral
J Consult Clin Psychol 2000;67:658-65.
therapy for obsessive-compulsive disorder: a meta-analytic review. Romanian
Journal of Cognitive and Behavioral Psychothcrapies 2002;2:89-104.
35. Vogel PA, Stiles TC, Gotestam KG. Adding cognitive therapy elements to
exposure therapy for obsessive compulsive disorder: a controlled study.
Manuscript received and accepted March 2006.
Behavioural and Cognitive Psychotherapy 2004;32:275-90.
'Associate Professor and Director, OCD/Atixiety Disorders Program,
36. Kozak MJ, Foa EB. Mastery of obsessive-compulsive disorder: therapist manual.
San Antonio (TX): The Psychologieal Corp; 1997. Department of Psychiatry and Psychology, Mayo Clinic, Rochester,
37. Franklin ME, Abramowitz JS, Kozak MJ, Levitt J, Foa EB. Effectiveness of Minnesota.
exposure and ritual prevention for obsessive-compulsive disorder: randomized Address for correspondence: Dr J Abramowitz, OCD/Anxiety Disorders
compared with nonrandomized samples. J Consult Clin Psychol Program, Department of Psychiatry and Psychology, Mayo Clinic, 200 First
2000;68:594-602. St, SW, Rochester, MN, 55905; [email protected].

Resume : Le traitement psychologique du trouble obsessionnel-compulsif

Le traitement psychologique du trouble obsessionnel-compulsif (TOC) a l'aide de methodes
d'exposition avec prevention de la reponse (EPR) est une des grandes reussites du domaine de la
sante mentale. En l'espace de quelque 20 ans, le pronostic des personnes souffrant de TOC est
passe de mediocre a tres bon, par suite de la mise au point de l'EPR. Nonobstant ce succes, les
interventions sont loin d'etre parfaites, car une minorite substantielle de patients soit refusent le
traitement, soit abandonnent prematurement, soit n'en retirent aucun benefice. Cet article
commence par un sommaire de Telaboration de l'EPR, depuis les premieres recherches
zootechniques sur l'apprentissage de l'evitement, menees dans les annees 1950. Ensuite, je presente
les mecanismes de l'EPR. La majorite de l'article passe en revue la documentation des resultats du
traitement par l'EPR pour le TOC et comprend des comparaisons avec la therapie cognitive, la
nouvelle venue au sein des traitements psychologiques du TOC.

416 Can J Psychiatry, Vol 51, No 7, June 2006