Inference-Based CBT: The OCD Therapy Without Exposure

You washed your hands. You saw the water run over them, you felt the soap, you dried them. And then a small voice says: but maybe there’s still something on them.

Most OCD treatment focuses on what you do after that voice speaks: the extra wash, the check, the question to a loved one. Inference-based CBT (I-CBT) goes one step earlier. It asks a different question: why did “maybe” win against what your own eyes and hands just told you?

This guide explains what inference-based CBT is, how it differs from ERP, what the research shows, and who might want to ask a therapist about it.

What is inference-based CBT?

Inference-based cognitive behavioural therapy is a form of CBT for OCD developed in Canada by Kieron O’Connor, Frederick Aardema and colleagues (Malandain et al., Revue Médicale Suisse, 2026). Earlier papers call it the inference-based approach (IBA) or inference-based therapy (IBT). It’s all the same family.

Its central idea is simple: in OCD, the obsession starts with a doubt, and that doubt is built from imagination rather than from evidence. “Maybe the door is unlocked.” “Maybe my hands are dirty.” I-CBT treats those first “maybes” as the real beginning of the obsession, not as harmless thoughts that only become a problem once you react to them (O’Connor et al., Cognitive Behaviour Therapy, 2005).

The goal is not to argue you out of your fear or to make you face it on purpose. It’s to help you trust your senses and common sense again, so the doubt loses its grip before the compulsion ever gets going (Aardema et al., Psychotherapy and Psychosomatics, 2022).

Inferential confusion: when “maybe” beats “I can see it”

I-CBT is built on one concept: inferential confusion. It means trusting an imagined possibility more than what you can directly see, hear or touch, until the possibility starts to feel like a likely fact (Aardema et al., Behaviour Research and Therapy, 2005).

The researchers who first described the approach noticed patterns in how people with OCD reason their way into doubt (O’Connor & Robillard, Behaviour Research and Therapy, 1995; Julien et al., Journal of Affective Disorders, 2016). Here’s what that can sound like:

• Borrowing a fact from somewhere else. “Germs can survive on surfaces for days, so there could be germs on my hand right now,” even though your hand looks and feels clean.
• Distrusting your senses. “I saw the stove knob in the off position, but maybe I didn’t really see it.”
• Reasoning backwards from a story. “If I were a bad person, I’d have thoughts like this, so these thoughts suggest I might be.” Researchers call this inverse reasoning: starting from an imagined cause and drawing conclusions about reality.
• Building a story that feels real. Small links of “what if” stack up into a vivid scene, and the scene starts to feel like a memory or a prediction.

This isn’t just a theory on paper. In a study of 183 people, inferential confusion was linked to OCD symptoms even after taking obsessive beliefs and mood into account, and scores were higher in people with OCD than in people with anxiety disorders or no diagnosis (Aardema et al., 2005). People with OCD also used inverse reasoning more than both healthy volunteers and people with anxiety or mood disorders when the scenarios touched on OCD-type concerns (Wong et al., Journal of Anxiety Disorders, 2019).

If this sounds close to intolerance of uncertainty, it is a near cousin. The difference: intolerance of uncertainty is about how hard it is to sit with doubt. Inferential confusion is about how the doubt gets built in the first place.

Inference-based CBT vs ERP: what’s actually different?

Exposure and response prevention is the best-established psychological treatment for OCD, and NICE recommends CBT that includes ERP for adults (NICE CG31). In ERP, you deliberately face feared situations and practise not doing the compulsion.

Inference-based CBT works differently:

	ERP	Inference-based CBT
Main target	The compulsion and the anxiety	The doubt that starts the obsession
Planned exposure exercises	Yes, the core of treatment	No deliberate, prolonged exposure
Key question	“Can I face this without the ritual?”	“Is there any real-world reason for this doubt?”
What you practise	Staying with anxiety, dropping rituals	Spotting the story, trusting your senses

Sources: Aardema et al., 2022; Wolf et al., Psychotherapy and Psychosomatics, 2024.

A useful way to picture it: ERP teaches you to live with “maybe.” I-CBT asks where “maybe” came from. In one study, people who got better at resolving obsessional doubt during inference-based therapy also had better treatment outcomes (Aardema & O’Connor, Journal of Behavior Therapy and Experimental Psychiatry, 2012). In I-CBT, compulsions are seen as steps further away from reality: each extra check feeds the imagined story instead of settling it (O’Connor & Robillard, 1995).

What does the research say about I-CBT?

The evidence base is smaller than for ERP, but it now includes several randomized trials. Here are the main ones.

Early trial: three treatments compared (2005)

Fifty-four people with OCD were randomly assigned to inference-based therapy, cognitive therapy based on appraisals, or ERP, for 20 weeks. All three groups improved. People with strong conviction in their obsessions did better with the inference-based approach than with appraisal-based cognitive therapy (O’Connor et al., 2005).

OCD with poor insight (2015)

Some people with OCD feel fairly sure their fears are realistic. This is called poor insight, and it often makes treatment harder. In a Dutch trial, 90 people with poor insight received 24 sessions of either the inference-based approach or standard CBT. Both groups improved, with no overall difference. In a small after-the-fact analysis of the 23 people with the poorest insight, the inference-based group improved more (Visser et al., Psychotherapy and Psychosomatics, 2015). That is a hint worth following up, not a firm finding.

Across OCD subtypes (2017)

In an open trial (no random assignment), 125 people across five major symptom subtypes started 24 sessions. Among the 102 who finished, 59.8% showed clinically significant improvement, including people with treatment-resistant OCD and strongly held beliefs. A small wait-list group did not improve (Aardema et al., Clinical Psychology & Psychotherapy, 2017).

I-CBT vs cognitive therapy vs mindfulness (2022)

In a two-site trial, 111 people were randomized to I-CBT, appraisal-based CBT, or an adapted mindfulness program. All three reduced OCD severity, with no significant difference between them. I-CBT led to greater improvement in overvalued ideas, and higher remission rates than mindfulness at the mid-point (Aardema et al., 2022).

The big head-to-head: I-CBT vs CBT with ERP (2024)

This is the largest trial so far. In the Netherlands, 197 people were randomized to 20 sessions of either standard CBT (with ERP) or I-CBT, and followed for a year. Both worked, and there was no significant difference in OCD symptoms at any point. But the study was designed to prove I-CBT was “not worse” than CBT, and the statistics couldn’t quite confirm that, so the authors call that question inconclusive. The clear difference: people rated I-CBT as more tolerable (Wolf et al., 2024).

Self-help version (2015)

A self-help version of inference-based therapy was tested online with 50 people with OCD. After four weeks, the self-help group had modest but significant drops in obsessions and washing symptoms compared with a wait-list (Moritz et al., Journal of Cognitive Psychotherapy, 2015). Short and small, but encouraging.

Bottom line: inference-based CBT has performed about as well as standard CBT in the trials so far, and people find it easier to stick with. It is not yet proven equal to ERP, and it is not named in major guidelines like NICE CG31.

Why might it work? Clues from 2026

Two new studies looked under the hood.

Reasoning change seems to matter most. Researchers went back to the 2022 trial and tracked three possible mechanisms: inferential confusion, obsessive beliefs (like inflated responsibility), and mindfulness. All of them were linked to getting better. But when tested side by side, only the drop in inferential confusion explained symptom improvement on its own, and this held in all three treatments, not just I-CBT (Ouellet-Courtois et al., Behaviour Research and Therapy, 2026). In other words, untangling imagination from reality may be part of how OCD treatment works in general.

The brain responds differently. In 55 people from the Dutch trial who had brain scans before and after 20 sessions, symptom improvement went with decreased activity after CBT with ERP, but with increased activity in some areas after I-CBT (Vriend et al., Journal of Affective Disorders, 2026). It’s early, but it suggests the two therapies may help through different routes.

Who might inference-based CBT suit?

Research hasn’t yet shown who does best with which therapy. Treat these as starting points for a conversation with a clinician, not rules:

• You turned down ERP, or dropped out. Fear of exposure is one reason people refuse or leave CBT (Wolf et al., 2024). I-CBT, rated more tolerable in that same trial, may be a way in.
• Your fears feel realistic to you. If you’re fairly convinced the danger is real, early studies suggest the inference-based approach may help (O’Connor et al., 2005; Visser et al., 2015).
• You “know” it’s irrational but the doubt keeps winning. I-CBT works directly on why the doubt feels convincing, rather than on how much it scares you.
• ERP helped, but not enough. The 2017 open trial included people who hadn’t responded to earlier treatment (Aardema et al., 2017).

ERP still has the deepest evidence, and it’s the standard first choice. I-CBT is a credible alternative to discuss, not a reason to stop a treatment that’s working.

3 I-CBT-style exercises you can try today

These are simplified exercises in the spirit of the approach, not a replacement for therapy.

1. Catch the first “maybe”. Next time an obsession starts, write down the very first doubt in one line: “Maybe I didn’t lock the door.” Then ask: what did I directly see, hear or feel that started this doubt? Often the honest answer is: nothing. The doubt arrived from a story, not from the situation.

2. Spot the borrowed fact. OCD likes to bring in true facts from somewhere else. “Doors do get left unlocked sometimes.” True in general, but is it true here, now? Write the general fact on one side and what your senses tell you about this door on the other.

3. Let your senses have the last word, once. Look at the thing (your hands, the stove, the lock) one time, with full attention, and name what you see out loud: “The knob is pointing to off.” Then walk away. The aim isn’t to check better. It’s to give reality one clear vote and stop letting the imagined story re-run the count. (If one look tends to turn into many, skip this one and work on it with a therapist instead.)

Practice · exercises from the ocd.app app

Practise spotting the OCD story

Face the thought 2 min Spot the OCD thought 2 min Ride the wave 1 min

Exposure and response prevention (ERP) means facing the thoughts OCD makes you avoid, without doing the ritual. Pick a thought from the app that makes you uneasy, then stay with it.

• There is no joy without control
• I have to avoid uncertainty
• I must have complete control over everything
• Uncertainty is my enemy
• The worst feeling is uncertainty
• I need absolute certainty
• Without control - I'm useless
• I have to know everything
• I must have absolute control
• Not knowing is unbearable

Educational practice, not therapy or a diagnosis. More exercises from the app · Practise daily in the ocd.app app →

If you’d like a structured way to practise between sessions, the ocd.app app has short daily CBT-based exercises for noticing OCD thoughts and stepping out of the loop.

How to find an I-CBT therapist

Inference-based CBT is less widely available than ERP. A few ways to look:

• Ask your current OCD therapist whether they use I-CBT or would consider it.
• The official I-CBT website has a self-reported provider list (the site notes the list is not a certification).
• Whatever the label, look for a clinician with specific experience treating OCD.

You can also start with our OCD test to get a sense of where you are before you talk to someone.

Frequently asked questions

What is inference-based CBT for OCD? Inference-based CBT (I-CBT) is a form of CBT that targets the obsessional doubt itself. It helps people see that the doubt comes from imagined possibilities rather than evidence, and teaches them to trust their senses again, without deliberate exposure exercises (Aardema et al., 2022).

Is I-CBT as effective as ERP? In the largest trial, 197 people did about equally well with I-CBT and standard CBT with ERP, but the result was formally inconclusive on whether I-CBT is “not worse.” People rated I-CBT as more tolerable (Wolf et al., 2024).

Does I-CBT use exposure? No deliberate, prolonged exposure. Instead, it focuses on reasoning: spotting how the obsessional doubt was built and reconnecting with what your senses tell you (Aardema et al., 2022).

What is inferential confusion? Inferential confusion means trusting an imagined possibility more than direct evidence, so the possibility starts to feel like a likely fact. It is linked to OCD symptoms independently of other obsessive beliefs (Aardema et al., 2005).

How long does I-CBT take? Trials have used 20 to 24 sessions, roughly weekly (Wolf et al., 2024; Aardema et al., 2017).

---

This article is for educational purposes and is not a substitute for professional diagnosis or treatment. If you’re in crisis, contact your local emergency services or a crisis line (in the US, call or text 988; in the UK, call Samaritans on 116 123).