Getting better

What works for OCD: every treatment worth knowing

In the office, when someone asks what actually works, this is the map I draw. There are more treatments on offer for OCD than anyone has years to try, so it helps to know the two first-line treatments, what's behind them, and what "working" actually means.

What actually works for OCD?

Two treatments carry first-line status for obsessive-compulsive disorder in the major guidelines. One is exposure and response prevention (ERP), the form of cognitive behavioral therapy built for this disorder. The other is SSRI medication. Everything else on the shelf is either a supplement to that pair or an alternative whose evidence is still maturing.

Most people don't arrive at this question fresh. By the time you're comparing treatments, you've usually paid tuition somewhere. Perhaps a therapy you liked that never went near the compulsions, or a medication graded and abandoned in week three, or a piece of advice that dissolved on contact with one bad night. So set the bar where it belongs, which is evidence, and hold whatever is newest or most natural-sounding to that same bar.

The American Psychiatric Association's practice guideline puts the whole answer in one sentence: "CBT and serotonin reuptake inhibitors (SRIs) are recommended as safe and effective first-line treatments for OCD." Unpack the acronyms and that's two treatments. Exposure and response prevention, recommended in the same guideline because it has the best evidentiary support. And the SSRI family of medications, along with an older cousin, clomipramine.

The International OCD Foundation (IOCDF) treatment guide says the same thing. It calls ERP "the first-line psychological treatment due to its very strong evidence base," and its medication guide names SSRIs the first-line medications for OCD. Two major guidelines, and the same short list from both.

One more thing belongs up front, and that's what "works" means, because every treatment claim you'll read tonight leans on that word. When researchers call a treatment successful for a given person, they mean something specific. Symptoms down by at least 35% on the standard severity scale, with the person clearly improved overall. In plain terms, the field starts counting success when the disorder loses a third or more of its grip, and many people get well past that line.

The National Institute of Mental Health (NIMH) is just as plain about the ceiling. There's no cure for OCD, and the available treatments help people manage symptoms and take part in day-to-day life and improve quality of life. Read those two facts together and the real promise is better than it may first sound. The thoughts may still visit. What treatment changes is who runs the day. What recovery actually looks like is its own page, and early is the right time to read it.

The first-line pair

The therapy half of the pair, exposure and response prevention, asks two things of you. You come into contact with whatever sets the fear off, on purpose, at a pace you helped set. And then you don't do the compulsion, the ritual OCD is insisting the moment requires.

That second part is the treatment. Every compulsion so far has taught your brain the alarm was worth sounding. Skip it, and your brain gets to watch what actually happens instead. The anxiety rises, crests, and often settles without help. And even on the days it settles slowly, the learning runs either way, because your brain is recording that the feared outcome didn't need the compulsion. Enough repetitions, and the alarm starts recalibrating.

In the IOCDF's research summary, patients average a 60% reduction in OCD symptoms. Average means average, some do better and some do worse, and the same guide says plainly that ERP isn't effective for everyone. Numbers like that are still why the guidelines put it first.

The next question is usually about the dose of the therapy itself. How much is a real course?

So the therapy runs on a scale of months, and knowing that going in is part of starting well.

The other half of the pair is medication. Per the IOCDF's medication guide, SSRIs are the first-line medications for OCD, and they and clomipramine are the only medications shown to work for it on their own. In a given SSRI trial, about 40 to 60% of people get a clinically significant improvement, and among the responders, symptoms drop by 40 to 50% on average. Symptoms get quieter without going away, and for many people quieter is what makes everything else possible.

Medication also runs on its own clock and its own doses, and both run longer and higher than depression care may have taught you to expect. An adequate trial takes eight to 12 weeks, at OCD-level doses. A verdict reached in week three, at a depression dose, isn't a verdict. What to actually expect from an SSRI has its own page.

And the two treatments aren't competing schools. The research studies them alone and in combination, and in practice they're often used together, and they cover for each other. Per the IOCDF, doing ERP can lower the relapse risk that follows stopping an SSRI. One retrains the loop. The other turns the alarm's volume down. That's the pair.

What if first-line isn't enough?

A first treatment that falls short still leaves obsessive-compulsive disorder with evidence-based next steps. They include checking the dose and length of the trial you already ran, adding a low-dose augmenting medication, and stepping up to intensive daily formats or to newer therapies with maturing evidence. One round falling short narrows the plan, it doesn't end it.

Somewhere after a treatment that fizzled, a darker question tends to show up, usually at night. Some version of what if I'm the one nothing works for. That's OCD's kind of question. It demands a verdict on your whole future, tonight, ahead of the evidence, and nobody honest can issue one. What you can have instead of a verdict is a next step.

Audit the round that fell short. On the medication side the questions are concrete. Was the dose an OCD dose, and did the trial run the full eight to 12 weeks? On the therapy side, did the work ever include actual exposure, with a ranked list and practice between sessions? A round that had none of those wasn't first-line treatment failing. It was first-line treatment missing.

Extend the medication road. Per the IOCDF's guide, about one-third of people who don't respond to an SSRI alone respond when a low-dose antipsychotic medication is added, used off-label for this purpose, and aripiprazole and risperidone carry the most evidence for that role. Clomipramine, the oldest effective OCD medication, is another established option. This branch is psychiatrist territory, and asking for that referral is progress in its own right.

Turn up the intensity. Treatment comes in doses too. The Cleveland Clinic's ERP page puts it plainly: "Most people attend weekly sessions for at least a few months. Sometimes, intensive daily programs work, too." When a weekly hour isn't enough, more hours per week is a real answer.

Consider the therapies behind the first-line one. Inference-based CBT, I-CBT for short, is one of the two that show up most in the research, and it works on the reasoning that produces the obsessional doubt rather than using exposure exercises. It's promising, and its evidence is still maturing. In a 2024 head-to-head trial against CBT with exposure, people improved with both, but the trial left open whether I-CBT matches it. People did find it easier to tolerate.

Acceptance and commitment therapy, ACT, is the other, and it has randomized trials of its own. In a 2010 trial, adults who got eight sessions of ACT, with no exposure exercises built in, did better on OCD severity than adults who got relaxation training, and the researchers' own verdict was that ACT is "worth exploring further as a treatment for OCD." In a 2025 trial of group treatment, symptom scores after group ACT were about as good as after group CBT built on exposure, at the end of treatment and a year later.

Both results deserve their asterisks. These are single trials, in specific formats and against specific comparisons, set next to a first-line treatment whose evidence base the guidelines call very strong. The three-way comparison, and how people actually choose among them, lives at ERP, I-CBT, and ACT.

And the options keep going. In 2018 the FDA permitted marketing of deep transcranial magnetic stimulation for OCD, and its announcement framed the device as another option for patients who have not responded to traditional treatments. That territory belongs to a psychiatrist who knows the whole field, which is one more reason to make that referral early.

A first round that falls short is a fork, not a wall.

What doesn't work for OCD?

Most of what fails in obsessive-compulsive disorder fails for one reason. It never targets the obsession-compulsion loop. That includes open-ended talk therapy that explores the thoughts but leaves the compulsions alone, and it includes reassurance from any source. On the medication side, the IOCDF is direct that antipsychotics on their own are not effective for OCD.

Almost everything on this list works for something, which is exactly what makes it expensive. Each entry arrives with a true story from a different condition.

Start with open-ended talk therapy, the kind built on exploring the thoughts, where they came from and what they might mean about you. For depression or grief or a hard year, that exploration is often the medicine itself.

Run it on OCD and something different happens. Every hour spent examining the doubt treats the doubt as worth examining. The content gets analyzed while the compulsions keep running, and the cycle keeps getting fed. It can be kind and skilled and real therapy, and the disorder doesn't even notice it's in the room.

This detour is what usually happens. In a nationally representative US survey of the disorder, 93% of adults with severe OCD had received some mental health treatment in the past year. The share who'd received treatment specifically for OCD was 30.9%. Most of the people in that gap were in care, just care pointed somewhere else. Why regular talk therapy didn't work tells that story from the inside.

Reassurance is the other dead end, and this one you've probably tested personally. You ask, or you search, and the answer helps. Real relief, for a while. Then it fades, the doubt rebuilds a notch louder, and it asks again. Every round teaches your brain that the question deserved an answer, and that lesson is what keeps the cycle running. That holds for a 2 AM search the same as for a therapy hour spent being talked out of the fear each week.

The IOCDF's guide is direct that antipsychotics on their own, without an SSRI, are not effective for OCD. Their evidence-based place is as low-dose additions alongside one. And quietly stopping an SSRI once you feel better has a documented cost, because relapse after discontinuing is described as high. Stopping is a planned decision made with a prescriber.

None of this makes the years you may have spent on that list your fault. You walked through the doors that were open. Just make the next door one that targets the loop.

How to choose where to start

Start with what you can reach, and with what you'll actually do. If a therapist trained in exposure work is reachable, that's the front door. Vet them before you commit, the questions to ask are on their own page, next to the timeline.

If medication is the reachable first step, a prescriber who knows OCD's doses and timelines can start that work now, and the therapy door stays open the whole time. And if neither is reachable this year, structured self-guided work exists, with real limits. Doing ERP on your own lays out where its edges are.

That reach-first logic isn't improvisation. The UK's national guideline builds OCD care the same way. For milder impairment, or when someone prefers to start there, the guideline offers low-intensity ERP built on structured self-help materials, and care steps up to intensive therapy or medication when the first step isn't enough.

One thing is missing from that list on purpose, and that's waiting for certainty about which door is perfect. OCD is a certainty-demanding disorder, and it will happily turn the choosing itself into one more compulsion, another comparison and another tab and another night of research, while the appointment never quite gets made. At some point the most clinical thing you can do is close the tabs and make the reachable appointment, and for a brain that runs on doubt, that's no small task.

For what it's worth, this is what I'd want in the hands of someone I love. Two proven places to start, with real options behind them and a short list of things to stop paying for. The rest is miles, and miles are walkable.