Translated from the Portuguese original. Read the original (Português)
Exposure and Response Prevention (ERP) is the behavioral procedure at the center of treatment for obsessive-compulsive disorder (OCD). It has two halves that only work together: approaching, in a planned and agreed-upon way, whatever triggers the discomfort, and not performing the ritual that usually relieves it. The idea is not to “endure suffering for its own sake”: it is to give the person the chance to discover firsthand that the feared catastrophe does not happen — and that they can ride out the wave without the ritual.
Exposure and Response Prevention (ERP, also called Exposure and Ritual Prevention, EX/RP) is a behavioral therapeutic procedure in which the person comes into contact, in a systematic, planned, and graded way, with the stimuli that evoke anxiety or discomfort — objects, situations, images, intrusive thoughts — while deliberately refraining from the avoidance responses and compulsive rituals that normally reduce that discomfort. By blocking the relief produced by the ritual, ERP undoes the contingency that maintained the behavior and allows the person to learn, through direct contact with reality, that the feared consequence does not occur. It is the central component of cognitive-behavioral therapy for OCD.
Where ERP comes from
In 1966, Victor Meyer, of Middlesex Hospital in London, published a paper that would change the treatment of OCD. He reported the case of two patients with a long history of handwashing rituals and described an approach that was unusual at the time: instead of soothing them, he placed them in contact with what they feared and prevented the rituals from being carried out.
Meyer described the procedure as a “modification of expectations” and pointed to two ingredients he considered decisive. The first: the patients noticed that not performing the ritual did not produce the overwhelming anxiety they had anticipated. The second: the disastrous consequences they expected simply did not happen. The procedure would later be given the name apotrepic therapy — from the Greek, “to dissuade,” “to turn away.”
These are exactly the two discoveries, made through experience rather than argument, that modern ERP still tries to produce. No one convinces a fear with an explanation; what convinces it is repeated contact with a world in which the catastrophe does not come.
Why the ritual does not go away on its own
First, it's worth separating the two words that give the disorder its name. Under the DSM-5-TR, obsessions are recurrent and persistent thoughts, urges, or images, experienced as intrusive and unwanted, that cause marked anxiety or distress. Compulsions are repetitive behaviors or mental acts that the person feels driven to perform in response to an obsession, or according to rules that must be applied rigidly, aimed at reducing distress or preventing some dreaded event — with no realistic connection to what they are meant to neutralize, or clearly excessive. For a diagnosis to apply, this set needs to consume time (for example, more than one hour a day) or cause clinically significant distress or impairment.
With that distinction made, the practical question remains: why are rituals so resistant? They persist because they work — they just work too fast, and in the wrong way.
The classic explanation combines two learning processes. A neutral stimulus comes to evoke fear through respondent conditioning: the doorknob, the stove, an aggressive thought. Then, the ritual that removes this discomfort is strengthened by negative reinforcement — washing your hands, checking the door, praying mentally, asking someone for reassurance. The relief arrives in seconds, and it is that speed that seals the problem.
The side effect is cruel: because the person never stays in the situation long enough without the ritual, they never find out what would happen without it. Every ritual performed retroactively “proves” that it was necessary. That is why the cycle closes in on itself instead of unraveling on its own.
The two halves of the procedure
The name of the technique is not decorative: it has two parts, and the literature is insistent that neither one works well on its own. Tap the tabs:
Approaching what triggers the discomfort
Exposure is graded and agreed upon in advance. Therapist and patient build a hierarchy of situations together, from what causes little discomfort to what causes a lot, using a subjective scale from 0 to 100 (the SUDS, short for “subjective units of distress”). The person chooses where to start and moves up at their own pace. The contact can be in vivo (touching the doorknob, leaving without checking the stove) or imaginal (vividly evoking the feared thought or scene — an essential resource when the fear is of something that cannot be staged, such as a catastrophe or an aggressive thought). None of this is a surprise: the person knows in advance what is going to happen and consents to each step.
Not performing the ritual
Response prevention means deliberately refraining from the behavior that would bring relief — and that goes well beyond what can be seen from the outside. What counts as a response to be prevented includes the visible ritual (washing, checking, arranging), the mental ritual (repeating a phrase, praying, “undoing” the thought, mentally reviewing the scene), avoidance (not going, not touching, postponing), and reassurance-seeking (asking again, “are you sure everything is okay?”). When one of these channels stays open, it works as an escape valve: the exposure happens on paper, but the learning does not.
One without the other teaches nothing new
Exposure without response prevention is the routine a person with OCD already lives every day: they encounter the trigger, perform the ritual, feel relief — and the cycle closes intact. Response prevention without exposure is just holding back the urge while carefully avoiding any trigger, which keeps the fear protected from any contradiction. The two together produce an unprecedented situation: the trigger present, the ritual absent, and time passing. It is in that window that new information gets in.
None of this is tracked in the dark. The reference instrument for measuring the severity of OCD is the Y-BOCS (Yale-Brown Obsessive Compulsive Scale), published by Goodman and colleagues in 1989: it has 10 items administered by a clinician in an interview, each scored from 0 to 4, with a total of 0 to 40 and separate subtotals for obsessions and compulsions. The scale was built precisely so it would not depend on the content of the obsessions — it asks, for both obsessions and compulsions, how much time they take up, how much they interfere with life, how much distress they cause, how much resistance the person can put up, and how much control they feel they have. That's five questions on each side: the ten items. The result is a number that can be compared week to week, showing whether the treatment is working.
What counts as “response”
This is the point where ERP is most often applied halfway. Flip the cards to see the four forms “response” takes:
The last card deserves a paragraph of its own. Family accommodation is the name given to the set of changes a family makes to reduce the suffering of a person with OCD: giving the requested reassurance, taking part in the ritual, facilitating avoidance — going somewhere in the person's place, taking the other street — and changing the household routine. It is a gesture of care, and that is exactly why it is hard to let go of. A meta-analysis of 41 studies (Wu et al., 2016) found a moderate-magnitude association between more family accommodation and greater symptom severity (r = 0.42). That is why ERP usually involves the family explicitly — not to turn anyone into an enforcer, but to agree, with everyone together, on how to respond to requests for reassurance without hostility and without feeding the cycle.
Why ERP works: two explanations
There are two readings of what happens during a well-done exposure, and they are not mutually exclusive.
- Habituation. The oldest and most intuitive explanation: by maintaining contact with the trigger for long enough, the emotional response decreases on its own, both within each exposure and across repetitions. It's the curve from the graph above. In clinical practice, this translates into “stay in the situation until the discomfort drops.”
- Inhibitory learning. The more recent reading, associated with Michelle Craske and colleagues (2008, 2014): the original association between trigger and danger is not erased — a new association is learned on top of it, one that competes with the old one. The critical ingredient is not the discomfort dropping, but the violation of expectation: the person explicitly predicts what they fear will happen, and then observes that it did not. The greater the surprise, the stronger the learning.
The difference is not academic; it changes the instructions given to the patient. In the habituation model, the success criterion is “the discomfort went down.” In the inhibitory learning model, the criterion is “what I predicted did not happen — and I handled it,” even if anxiety remains high by the end of the exercise. From this second reading came practical recommendations that are now widely used: varying the exercises instead of always repeating the same one, practicing in different contexts, removing safety signals (the hand sanitizer in the pocket, the friend who tags along), and naming the emotion in words during the exposure.
Seen through the lens of behavior analysis, ERP combines two processes of extinction: exposure without the feared consequence weakens the conditioned emotional response, and preventing the ritual removes the negative reinforcement that sustained the compulsion. And, as with any new learning, the next challenge is generalization and maintenance — which is why training is carried out of the office and into the home, work, and the street.
What the evidence says
ERP is a psychological intervention with a robust empirical basis, and it is the recommended psychological treatment at every step of the British NICE guideline for OCD.
What the research showsThe reference trial is Foa and colleagues (2005): a randomized, placebo-controlled study with 122 adults in outpatient treatment and double masking in the medication conditions, which compared intensive ERP, clomipramine (up to 250 mg/day), the combination of the two, and placebo over 12 weeks. All active treatments outperformed placebo, and ERP alone was superior to clomipramine alone. Response rates, assessed by the global clinical improvement scale, were 62% (intention-to-treat) and 86% (among completers) for ERP; 70% and 79% for the combination; 42% and 48% for clomipramine; and 8% and 10% for placebo. A systematic review with meta-analysis of 37 randomized trials published between 1993 and 2014 (Öst et al., 2015) found large effects for cognitive-behavioral therapy for OCD — around d = 1.3 compared with both waitlist and placebo conditions. The NICE guideline (United Kingdom, CG31) organizes care in steps: for mild functional impairment, low-intensity psychological treatments including ERP (up to 10 therapist hours); for moderate impairment, a choice between an SSRI antidepressant or more intensive CBT including ERP (more than 10 hours); and, for severe impairment, combined treatment. For children and adolescents with moderate to severe impairment, the guideline indicates CBT including ERP, involving the family and age-appropriate, as the treatment of choice.
It's worth reading these numbers honestly. The difference between 62% and 86% in Foa's study is not a statistical detail: it is the distance between “everyone who entered the study” and “those who made it to the end.” ERP asks a lot of the people who do it, some refuse the treatment and others drop out — and that is part of the picture. Beyond that, a good result on average does not guarantee a good result for every person: some of those who complete treatment improve without becoming free of symptoms. None of this disqualifies the technique; it simply places what is reasonable to expect, in the spirit of evidence-based practice. One last point of vocabulary: “ERP” is the name of the package designed for OCD. What runs across several conditions is exposure — its first half — present in CBT for the anxiety disorders, which since the DSM-5 form a diagnostic group separate from OCD.
ERP is not “facing fear just any way”Exposure done by surprise, without prior agreement, without a hierarchy, without consent, or against the person's will is not ERP — it is an aversive experience that tends to increase fear and destroy trust in the treatment. ERP is collaborative by definition: the person understands the rationale, chooses the steps, and can stop. It is also not an exercise to improvise alone from a video: without mapping out mental rituals, avoidance, and reassurance-seeking, the exercise turns into suffering without learning. Look for a professional with specific training in CBT for OCD.
ERP and autism: a distinction that needs to be made first
OCD can occur in autistic people, and in those cases ERP is a legitimate treatment option. But there is a prior step, and it is decisive: distinguishing a compulsion from a repetitive autistic behavior.
The difference is not in appearance — it is in function and lived experience. The compulsion is unwanted: it arises from a distressing obsession, is performed to avoid something feared, and tends to be felt as something that invades life. Stereotypy (stimming) and restricted interests, on the other hand, generally regulate, organize, and bring pleasure — many autistic people describe them as part of who they are. In a qualitative study with 15 autistic adults living with OCD symptoms (Long, Cooper & Russell, 2024), this boundary shows up in the participants' own descriptions: while the repetitive behaviors were described as central to their own identity, OCD was described as an intruder — in the metaphor that gives the article its title, “autism is the arena and OCD is the lion.”
The ethical consequence is direct: ERP is not performed on stimming, restricted interests, or sensory needs. That would not be treating OCD; it would be taking away regulation resources the person depends on. A review of the field (Bedford, Hunsche & Kerns, 2020) points to this differential assessment as a central clinical challenge at this intersection.
When OCD is in fact present, the evidence exists and is encouraging, though modest. A randomized clinical trial (Russell et al., 2013) followed 46 autistic adolescents and adults with comorbid OCD, comparing CBT for OCD (average of 17.4 sessions) with an active anxiety-management control (14.4 sessions). Both groups improved, with within-group effects of 1.01 and 0.6 and a numerical difference favoring CBT in response rates — 45% versus 20%, considering response a drop of more than 25% in severity. Read this carefully: the control was active and plausible, and the study does not demonstrate CBT's superiority over it. Furthermore, most participants did not meet the response criterion in either group — a sign of how much adaptation is still needed.
Adapting, in fact, is the rule and not the exception. A systematic review of 12 studies of modified CBT for autistic youth (Walters, Loades & Russell, 2016) found reductions in anxiety, OCD, and depression — and noted that, in practice, more adaptations are applied than the NICE guideline itself recommends. The modifications described fall into three categories: adding something to the protocol to accommodate the person (for example, waiving eye contact), leaving out parts of the protocol (such as work on core beliefs), and changing how the usual practice is conducted, diversifying the forms of communication.
Test your understanding
Quick quiz
Three questions to reinforce the concept. You'll see the explanation right away.
Further reading
- Meyer, V. (1966). “Modification of expectations in cases with obsessional rituals”. Behaviour Research and Therapy, 4(4), 273–280.
- Foa, E. B., Liebowitz, M. R., Kozak, M. J. et al. (2005). “Randomized, placebo-controlled trial of exposure and ritual prevention, clomipramine, and their combination in the treatment of obsessive-compulsive disorder”. American Journal of Psychiatry, 162(1), 151–161.
- Öst, L.-G., Havnen, A., Hansen, B., & Kvale, G. (2015). “Cognitive behavioral treatments of obsessive-compulsive disorder. A systematic review and meta-analysis of studies published 1993–2014”. Clinical Psychology Review, 40, 156–169.
- Craske, M. G. et al. (2008). “Optimizing inhibitory learning during exposure therapy”. Behaviour Research and Therapy, 46(1), 5–27.
- Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. (2014). “Maximizing exposure therapy: an inhibitory learning approach”. Behaviour Research and Therapy, 58, 10–23.
- National Institute for Health and Care Excellence (NICE). Obsessive-compulsive disorder and body dysmorphic disorder: treatment (CG31), 2005.
- Goodman, W. K., Price, L. H., Rasmussen, S. A. et al. (1989). “The Yale-Brown Obsessive Compulsive Scale. I. Development, use, and reliability”. Archives of General Psychiatry, 46(11), 1006–1011.
- Wu, M. S., McGuire, J. F., Martino, C., Phares, V., Selles, R. R., & Storch, E. A. (2016). “A meta-analysis of family accommodation and OCD symptom severity”. Clinical Psychology Review, 45, 34–44.
- Russell, A. J., Jassi, A., Fullana, M. A. et al. (2013). “Cognitive behavior therapy for comorbid obsessive-compulsive disorder in high functioning autism spectrum disorders: a randomized controlled trial”. Depression and Anxiety, 30(8), 697–708.
- Long, H., Cooper, K., & Russell, A. (2024). “‘Autism is the Arena and OCD is the Lion’: autistic adults’ experiences of co-occurring obsessive-compulsive disorder and repetitive restricted behaviours and interests”. Autism, 28(11), 2897–2908.
- Walters, S., Loades, M., & Russell, A. (2016). “A systematic review of effective modifications to cognitive behavioural therapy for young people with autism spectrum disorders”. Review Journal of Autism and Developmental Disorders, 3(2), 137–153.
- Bedford, S. A., Hunsche, M. C., & Kerns, C. M. (2020). “Co-occurrence, assessment and treatment of obsessive compulsive disorder in children and adults with autism spectrum disorder”. Current Psychiatry Reports, 22(10), 53.
- American Psychiatric Association. DSM-5-TR: Diagnostic and Statistical Manual of Mental Disorders. 5th ed., text revision, 2022.
This entry is educational and does not replace assessment, diagnosis or treatment by a qualified health professional. Every person is unique — for a specific case, seek individual guidance.