The short answer: a standard course of exposure and response prevention runs 12 to 20 sessions, typically one hour each, which is about three to five months at a weekly pace. Most people notice the first genuine shift between weeks 3 and 6, after the assessment and hierarchy-building phase ends and real exposures begin. The average outcome is roughly a 60% reduction in OCD symptoms, not zero symptoms.

The number that matters more than the calendar is exposures completed. Twenty sessions of talking about OCD does nothing. Twelve sessions of doing exposures does a great deal.

What the standard course looks like

The International OCD Foundation describes a typical course as 12 to 20 sessions, adjustable to individual need, with sessions generally lasting about an hour. On average, patients achieve a 60% reduction in OCD symptoms, along with reductions in general anxiety, depression, interpersonal stress, and functional interference (IOCDF, Exposure and Response Prevention).

Session frequency changes the calendar substantially:

  • Weekly: 3 to 5 months
  • Twice weekly: 6 to 10 weeks
  • Intensive outpatient: several hours per day for weeks
  • Concentrated (Bergen 4-day): four consecutive days

Faster is not automatically better, but the evidence does not support the intuition that slower is safer either. More frequent sessions early tend to build momentum, because the gap between exposures is when avoidance creeps back.

A realistic week-by-week timeline

Sessions 1 to 3: nothing feels different. This is assessment, psychoeducation, and building the exposure hierarchy, an ordered list of situations, thoughts, and triggers that provoke obsessional fear. You will also map your compulsions, including the mental ones you may not have counted as compulsions. Many people get discouraged here because they came for treatment and are making lists instead. That is the work.

Sessions 3 to 6: first exposures, worst feelings. Anxiety typically goes up before it comes down. You are deliberately approaching something you have spent years avoiding while not performing the ritual that normally ends the discomfort. Expect this phase to feel like the treatment is making things worse. It is the phase with the highest dropout.

Weeks 3 to 6: the first shift. Something specific happens: an exposure that terrified you two weeks ago becomes merely unpleasant. This is the learning ERP is built to produce, that the feared outcome does not arrive and that anxiety falls on its own.

Sessions 8 to 15: climbing the hierarchy. Harder exposures, longer without rituals, and increasing amounts done independently between sessions. Most of the total gain accumulates here.

Sessions 15 to 20: generalizing and relapse prevention. Applying the same approach to new triggers without a therapist scripting it, and building a plan for the flare-ups that will happen later.

What the outcome data actually shows

The reference trial is still Foa and colleagues, published in the American Journal of Psychiatry in 2005. It randomized adults with OCD to intensive exposure and ritual prevention, clomipramine, both, or placebo over 12 weeks. Response rates among all treated patients and among completers were 62% and 86% for exposure and ritual prevention, 42% and 48% for clomipramine, 70% and 79% for the combination, and 8% and 10% for placebo (Foa et al., 2005).

Two things in those numbers are worth naming. The gap between 62% and 86% is the cost of dropping out; ERP works considerably better for people who finish it. And the placebo response of 8% to 10% is unusually low, which tells you OCD does not tend to improve on its own with attention and time.

The fast version: concentrated ERP

The Bergen 4-day treatment delivers ERP over four consecutive days, individually tailored but run in a group setting with roughly one therapist per patient.

Reported outcomes are strong. Studies have found around 86% of patients showing a clinically significant response at post-treatment with roughly 68% in remission six months later, and one four-year follow-up reported 73% meeting strict consensus remission criteria at post-treatment and 69% at four years (Frontiers in Psychology, 12-month follow-up). Dropout was 1.3%, against roughly 19.7% across psychotherapy trials generally.

The catch is access. Programs delivering this format are concentrated in a handful of countries and centers, and it is not what most United States clinics mean when they say "intensive."

What makes ERP take longer

  • Homework not done. The between-session exposures are where most of the learning happens. Session-only ERP moves at a fraction of the pace.
  • Reassurance-seeking continuing outside sessions. Asking a partner or a search engine whether it is fine undoes the exposure you just completed. See why reassurance makes OCD worse.
  • Hidden mental compulsions. Reviewing, counting, praying, or arguing internally during an exposure converts it into a ritual. Common in Pure O presentations and easy for an untrained therapist to miss.
  • Staying at the bottom of the hierarchy. Comfortable, and ineffective past the first few weeks.
  • Untreated depression. It reduces the capacity to do effortful, unpleasant homework.
  • A therapist without ERP training. Supportive talk therapy for OCD is not a slower version of ERP. It is a different thing that does not work for this.

Where medication fits in the timeline

SSRIs for OCD work on a slower clock than they do for depression, commonly 8 to 12 weeks at an adequate dose before the benefit is clear, and the doses required are typically higher. If you are doing both, expect ERP to show its effect first.

Combination treatment beat clomipramine alone in the Foa trial and was roughly comparable to ERP alone on response rate. For many people the practical answer is medication to make ERP possible, and ERP to produce the durable change. Our overview of how ERP works covers the mechanics in more depth.

The bottom line

Plan for 12 to 20 sessions, expect the first three to feel like nothing is happening, expect weeks 3 to 6 to feel worse before better, and expect roughly a 60% reduction rather than the disappearance of OCD. If you have completed 20 real exposure-based sessions with homework and nothing has moved, the question to ask is not whether ERP works. It is what specifically was delivered, and whether mental compulsions or reassurance-seeking quietly neutralized it.

Last updated: August 2026. This article is for informational purposes only and does not constitute medical advice. If you are in crisis, call or text 988.