Recognizing OCD and Referring for ERP or I-CBT: What Primary Care and Generalist Therapists Should Know
Last reviewed: 10/20/2026
Reviewed by: Dr. Kiesa Kelly
Recognizing OCD in primary care is harder than it looks, because the part you can see is usually the smaller part. A patient may come in for insomnia, a rash from handwashing, low mood or a stomach that will not settle, and mention none of the thoughts that are driving it. When the thoughts are about harm, sex or blasphemy, patients often hide them out of shame, and clinicians who do hear them can mistake them for a warning sign. This guide is for primary care clinicians, OB-GYNs and generalist therapists who want to refer for OCD treatment with confidence. If insomnia turns out to be the main problem rather than a sign of OCD, our guide on when to refer for CBT-I covers that referral.
The short version: Ask about intrusive thoughts and the private rituals that follow them, screen with the DOCS, and refer for exposure and response prevention (ERP) or inference-based CBT (I-CBT) with a clinician who treats OCD. At ScienceWorks, our team offers both in Tennessee, by telehealth or in person at our Nashville office.
In this article, you'll learn:
Three misconceptions that keep OCD hidden in general practice
Four questions that surface compulsions a patient will not volunteer
How to respond when a patient discloses a frightening harm or sexual thought, and when a fuller risk assessment is still the right call
Why we recommend the DOCS as a screen, and what a score can and cannot tell you
What to tell your patient the ERP or I-CBT referral is for, and how to coordinate with the prescriber
The short answer: ask about the hidden half of OCD, then refer for ERP or I-CBT
OCD has two halves. The obsession is the unwanted thought, image or urge. The compulsion is whatever the person does to neutralize it, and that can be an invisible mental act as easily as a visible one, such as silently repeating a phrase [1]. Patients with OCD are often ashamed of their symptoms and find them very hard to raise, which is why UK guidance asks clinicians to sensitively explore the hidden distress these disorders cause [2].
The cost of missing it is measurable. In a vignette survey of 208 primary care physicians, about half (50.5%) did not identify the OCD case, and the themes missed most often included obsessions about sexual orientation (84.6%), aggression (80.0%) and pedophilia (70.8%) [3]. In one outpatient sample, the self-reported mean gap between first symptoms and diagnosis was more than 12 years [4]. Our white paper on specialized OCD care walks through this research in more detail; this guide focuses on what to do in the room.
Three misconceptions do most of the hiding:
Misconception 1: "OCD means washing and checking." In reality, contamination and symmetry were the presentations physicians recognized most easily in that survey. The themes they missed most were taboo ones, such as aggression, pedophilia and sexual orientation [3].
Misconception 2: "If it were OCD, I would see the rituals." Many compulsions happen in the head. In one clinic sample of 80 people with OCD, just over half (53.75%) had mental compulsions, such as replacing a "bad" thought with a "good" one or silently reassuring themselves [5]. UK guidance names treatment for people with obsessive thoughts and no overt compulsions as its own category [2].
Misconception 3: "A patient who describes violent thoughts needs a risk workup first." UK guidance notes that intrusive sexual, aggressive and death-related thoughts are common in OCD and "are often misinterpreted as indicating risk" [2]. An unnecessary or lengthy risk assessment can make OCD worse [1]. Recognition comes first. A fuller risk assessment is warranted in the specific situations described below, and UK guidance asks clinicians to assess self-harm and suicide risk in everyone diagnosed with OCD [2].
🧩 Key takeaway: The visible ritual is optional and the shame is not. Ask about the thoughts and the private rituals directly, and expect taboo themes to be the ones a patient hides.
Four questions that surface compulsions you can't see
UK guidance lists six direct questions for patients at higher risk of OCD, including people with depression, anxiety or substance use: whether they wash or clean a lot, check things a lot, have a thought that keeps bothering them that they would like to get rid of but cannot, whether daily activities take a long time to finish, whether they are bothered by order or mess, and whether these problems trouble them [2]. They are a good start. They are also weighted toward visible behavior, so we add four questions aimed at the half a patient cannot be seen doing.
Mental reviewing: "When the thought shows up, do you go back over what happened in your head, or check your memory to be sure you didn't do something?" Reviewing past actions for evidence is a common covert compulsion in harm and sexual obsessions [1].
Reassurance seeking: "Do you ask people, or search online, to make sure something is okay, or that you are not a bad person?" Reassuring and rationalizing work as covert compulsions that keep the obsession going [1], and in one lab study, reasons for choosing online reassurance typically involved interpersonal concerns, with participants with OCD more likely than controls to voice concern about the consequences of revealing their worries to others [6].
Avoidance: "Are there places, people or objects you now steer around because of a thought?" A man who has moved every knife out of his kitchen may be describing avoidance rather than danger; avoidance of knives or sharp implements is one of the features that suggest OCD in thoughts of violence [1].
"Just right": "Do some things have to feel exactly right before you can stop, even when you know they're fine?" Symmetry, completeness and exactness form one of the four symptom dimensions the DOCS measures [7].
A composite example from primary care: a 38-year-old man comes in for tension headaches and poor sleep. He mentions, almost in passing, that he has stopped holding his infant niece. When you ask the reviewing question, he goes quiet, then says he spends evenings searching online for whether people who have thoughts like his are dangerous, and replays every family visit to be sure nothing happened. Nothing about his visit looked like OCD until the second half came into view.
A composite example from a generalist therapy caseload: a 27-year-old client has spent months in weekly therapy for "relationship anxiety." Each session circles the same question of whether she really loves her partner, and each time she leaves relieved for a day or two. When you ask about reassurance, she describes texting friends the same question several times a week and keeping a running mental list of moments that "prove" her feelings. The pattern of doubt, then neutralizing, then brief relief is the signal to consider OCD and to screen with the DOCS.
🔍 Key takeaway: Ask about reviewing, reassurance, avoidance and "just right" by name. Patients rarely volunteer these, and they answer far more readily when the question shows you already know the pattern.
When a patient tells you an intrusive thought scares them
How you respond in the first minute shapes whether the patient ever raises it again. People with OCD are usually deeply ashamed of these thoughts and often hold back the details [1]. UK guidance advises telling people who are distressed by obsessive thoughts that such thoughts are occasionally experienced by almost everybody, and that, when frequent and distressing, they are a typical feature of OCD [2].
The research behind that advice is reassuring. Intrusive thoughts are close to universal, and their content is indistinguishable from clinical obsessions; what differs in OCD is the meaning the person attaches to them [1]. Among 388 postpartum women in one Canadian study, 44.4% reported unwanted intrusive thoughts of intentionally harming their infant, and those women were not more likely to report aggression toward their baby than women without such thoughts [8].
What tends to backfire is hedging. Patients have described clinicians who said they were "unlikely to act" on the thoughts but would notify authorities anyway, who suggested avoiding work with children "to be on the safe side," or who implied an unconscious wish [1]. Each of these confirms the doubt the patient came in with. A related trap is reassurance seeking on either side of the desk, such as "Are you 100% confident that these are obsessional thoughts rather than something else?" Questions like this increase doubt and anxiety in both patient and clinician, a transmission of obsessional worry from the patient to the clinician [1].
If a patient struggles to say what the thought is, it can help to name a range of common themes and ask whether any apply, so they can answer yes or no rather than describe the thought in detail [1]. Then name what you are hearing, and explain the referral.
What ego-dystonic obsessions sound like
"Ego-dystonic" means the thought is at odds with the person's values. When the content is repugnant to them, it frightens them precisely because it is the opposite of what they want [1]. In a review written for clinicians assessing risk in OCD, the features that point toward OCD include the thought being experienced as repugnant, no past behavior consistent with it, active avoidance of triggers (for example, knives), efforts to suppress it, frequent occurrence, marked distress and guilt, a strong wish for help, and other obsessive-compulsive symptoms elsewhere [1]. The same review stresses that no single feature settles it; the overall pattern is what counts [1].
A composite example from an OB-GYN visit: at six weeks postpartum, a patient asks whether it is normal to picture dropping her baby down the stairs. She has stopped carrying him past the landing, asks her partner to give every bath, and checks repeatedly at night that she has not hurt him. She is tearful and ashamed, and she asks directly whether she should be kept away from him. This is the shape of an intrusive-thought presentation: unwanted, avoided, checked and confessed. A plan that keeps her from her child can make things worse: placing an adult with the mother at all times can increase her doubts, distress and avoidance [1], and in one Canadian study, mothers with intrusive thoughts of harming their infant were not more likely to report aggression toward the baby [8]. In our clinical judgment, reasonable next steps are a calm explanation, a check for the features in the next section (especially psychotic symptoms, depression and suicidal thoughts), a DOCS [10] and a referral.
When a risk assessment is still the right call
Recognizing an obsession does not end the safety conversation. The risks that matter in OCD usually come from somewhere other than the obsession itself, and some situations call for a fuller assessment [1,2]:
The thought is not ego-dystonic. It feels attractive, wanted or pleasurable at times, the patient is ambivalent about having it, or it is being elaborated in detailed fantasy [1]. Keep this separate from the obsessional doubt itself: people with OCD can become confused about whether they wanted a thought, and self-reported sexual arousal is an unreliable indicator of sexual interest [1]. In children and adolescents, the absence of ego-dystonicity does not rule out OCD [1].
There is a relevant history, such as violence, exploitation or abuse of others, or sexual offending [1]. This is different from the minor past incidents people with OCD often volunteer as evidence against themselves; overreporting can itself be a form of reassurance seeking [1].
There are psychotic features, such as command hallucinations, or, after childbirth, possible psychotic symptoms, especially with no history of OCD, where the differential is postpartum (puerperal) psychosis [1]. UK perinatal guidance advises referring a woman with sudden onset of symptoms suggesting postpartum psychosis for immediate assessment [13].
The patient is depressed, hopeless or thinking about suicide. People with OCD are at greater risk of suicide than the general population, and UK guidance asks clinicians to assess self-harm and suicide risk in people with OCD, especially when depression is present [1,2]. Ask directly even when the thoughts look clearly obsessional: hopelessness, or a patient's belief that they pose a real danger to others, can motivate suicide, and in rare cases a person terrified of harming loved ones has taken their own life to protect them [1].
The compulsions themselves are causing harm, such as self-injury done to neutralize a thought, aggression when someone blocks a ritual, or children affected by a parent's rituals or preoccupation [1,2].
If you are uncertain about the risk, UK guidance advises consulting a clinician with specific expertise in OCD [2]. Follow your own professional and legal reporting obligations; nothing in this guide changes them. If a patient is in immediate danger, call 911, and for a suicidal crisis the 988 Suicide and Crisis Lifeline is available by call or text.
For the patient side of this, our article on harm OCD and taboo intrusive thoughts is written to be shared with patients who need to hear that the thought is not the person.
⚖️ Key takeaway: A repugnant, avoided, distressing thought points to OCD and calls for recognition and referral. Wanted content, a relevant history, psychotic features, suicidality or harm from the compulsions are the reasons to assess further, and suicide risk is worth asking about in every case.
A screen to use: the DOCS
We recommend the Dimensional Obsessive-Compulsive Scale (DOCS) as the screen. It is a 20-item self-report that takes about five minutes [7]. It measures four symptom dimensions: contamination; responsibility for harm and mistakes; unacceptable or taboo thoughts about topics such as sex, violence and blasphemy; and symmetry, completeness and exactness [7]. Each dimension has five items rated from 0 to 4 for time, avoidance, distress, interference and difficulty disregarding the thoughts, and the items rate the severity of the patient's own symptoms rather than a fixed list [7]. In its development studies, scores showed good reliability, validity and sensitivity to treatment across clinical and nonclinical samples [9].
For OB-GYNs, the DOCS has been tested as a perinatal screen. In a Canadian study of 574 pregnant people followed into the postpartum period, the DOCS total score was the most accurate of the measures compared for detecting perinatal OCD, and neither the full Edinburgh Postnatal Depression Scale nor its anxiety subscale met the authors' criteria for an accurate OCD screen [10].
The developers describe a total of 18 as a reasonable reference point for distinguishing people with OCD from people without a disorder, and 21 for distinguishing OCD from other anxiety disorders [7]. They have not established mild, moderate or severe ranges [7]. In a general primary care population, where OCD is less common, an elevated score is a reason for a diagnostic interview, not a diagnosis. Our guide to reading DOCS scores explains what the total and the four dimension scores can and cannot tell you, and patients can take the DOCS on our site and bring the result to you.
📋 Key takeaway: The DOCS is short, covers taboo thoughts, and has perinatal evidence. Use the score to decide on a diagnostic interview, not to make the diagnosis.
ERP or I-CBT: what to tell your patient the referral is for
Patients engage better when they know what the referral is for. Tell them ERP means gradually approaching what they avoid while dropping the ritual, with practice between sessions, and that UK guidance recommends CBT that includes ERP for adults with OCD [2]. Tell them I-CBT works on the reasoning that makes an obsessional doubt feel credible and does not use deliberate exposure; a multisite trial was designed to test whether I-CBT is non-inferior to standard CBT with ERP, could not establish that (the result was inconclusive), and found I-CBT was rated more tolerable [11]. The choice depends on the presentation and what the patient can sustain, and a side-by-side comparison of ERP and I-CBT covers the details.
For harm themes specifically, our page on I-CBT for harm OCD describes how that work is approached.
Medication and therapy together: coordinating with the prescriber
Many patients reach a specialist already taking an SSRI. UK guidance recommends low-intensity ERP-based treatment for mild impairment, a choice between an SSRI and more intensive CBT including ERP for moderate impairment, and both together for severe impairment [2]. When an SSRI is working, it advises continuing it for at least 12 months and, when reducing or stopping, tapering gradually to minimize discontinuation symptoms [2], a process the prescriber manages. In a randomized trial of adults with OCD, intensive ERP alone and ERP combined with clomipramine performed similarly at 12 weeks, and both outperformed clomipramine alone [12]. The trial tested clomipramine rather than an SSRI, and it does not tell you whether a particular patient should continue a current SSRI during ERP; that stays a prescriber decision. We do not prescribe, and this guide gives no dosing advice; medication decisions stay with you or the patient's prescriber.
What makes the combination work is coordination. UK guidance recommends clear agreement among professionals about who is responsible for monitoring and treatment [2]. In practice, that means telling the patient not to change or stop medication on their own because therapy is starting, agreeing on who watches side effects and mood, and sharing progress. With the patient's written permission, we share measures and updates with the referring clinician and the prescriber. Our information for referring clinicians explains how that works.
💊 Key takeaway: Therapy and medication are partners, not alternatives. Agree on who monitors what, and make sure the patient knows not to stop medication on their own when ERP begins.
If you take one thing from this guide, make it the order of operations: ask about the hidden half, respond to a frightening thought with calm recognition rather than suspicion, screen, and refer to someone who treats OCD every week. A patient who has carried these thoughts alone for years often needs only one clinician who asks the right question.
How to refer: Your patient can book a free phone consultation (or text 931-223-1095) and mention you. No referral or prior authorization is needed.
Please don't send health information by email or web form; once your patient signs a release, we arrange a secure exchange. How to refer a patient to ScienceWorks →
Frequently Asked Questions
What screening measure should I use if I suspect OCD?
We recommend the Dimensional Obsessive-Compulsive Scale (DOCS), a 20-item self-report that rates how severe the patient's own symptoms are across four themes, including taboo thoughts. Its developers describe a total of 18 as a reference point against people without a disorder and 21 against other anxiety disorders. Treat an elevated score as a reason for a diagnostic interview, never as a diagnosis.
Are intrusive thoughts about harming someone a safety risk?
On their own, usually not. Unwanted harm thoughts are common, and when a patient finds them repugnant and works to avoid or cancel them, that pattern points to OCD rather than intent. A fuller assessment is warranted when the thoughts feel wanted or pleasurable, there is a history of violence or offending, psychotic features are present, the patient is depressed, hopeless or suicidal, or the compulsions cause harm. Ask about suicide and self-harm either way.
Should my patient keep taking an SSRI during ERP?
That decision belongs to the prescriber, and we coordinate with prescribers rather than prescribe. UK guidance recommends continuing an SSRI that is working for at least 12 months, and combining an SSRI with CBT that includes ERP when OCD causes severe impairment, so in our view starting ERP is not, on its own, a reason to stop. If the patient or prescriber does plan to stop, the same guidance advises tapering gradually, a process the prescriber manages.
About the Author
Dr. Kiesa Kelly is a licensed clinical psychologist and the founder of ScienceWorks Behavioral Healthcare, with more than 20 years of experience in psychological assessment and evidence-based treatment. Her clinical training includes the University of Chicago, Vanderbilt University and the University of Wisconsin, and NIH-funded research training earlier in her career.
Much of her clinical work concerns presentations where the surface content misleads, OCD's taboo themes among them, where the thought that frightens a patient most is often the clearest evidence of what the condition is doing. She reviews every clinical article published here before it goes live.
References
1. Veale D, Freeston M, Krebs G, Heyman I, Salkovskis P. Risk assessment and management in obsessive–compulsive disorder. Adv Psychiatr Treat. 2009;15(5):332-343. https://doi.org/10.1192/apt.bp.107.004705
2. National Institute for Health and Care Excellence. Obsessive-compulsive disorder and body dysmorphic disorder: treatment (CG31). 2005, updated 2024. Accessed October 20, 2026. https://www.nice.org.uk/guidance/cg31
3. Glazier K, Swing M, McGinn LK. Half of obsessive-compulsive disorder cases misdiagnosed: vignette-based survey of primary care physicians. J Clin Psychiatry. 2015;76(6):e761-e767. https://doi.org/10.4088/JCP.14m09110
4. Ziegler S, Bednasch K, Baldofski S, Rummel-Kluge C. Long durations from symptom onset to diagnosis and from diagnosis to treatment in obsessive-compulsive disorder: a retrospective self-report study. PLoS One. 2021;16(12):e0261169. https://doi.org/10.1371/journal.pone.0261169
5. Pal V, Ramdurg S, Chaukimath S. Assessment of the prevalence and types of mental compulsions in patients with obsessive-compulsive disorder in North Karnataka: a cross-sectional study. Cureus. 2024;16(10):e71960. https://doi.org/10.7759/cureus.71960
6. Parsons CA, Kim HJ, Singh S, Lkhagva T, Wang J, Alden LE. Covert or connected: motivations for online and interpersonal reassurance-seeking in OCD. J Anxiety Disord. 2025;115:103057. https://doi.org/10.1016/j.janxdis.2025.103057
7. University of North Carolina at Chapel Hill. Dimensional Obsessive-Compulsive Scale (DOCS): home, methods of development and scoring pages. Accessed October 20, 2026. https://docs.web.unc.edu/
8. Fairbrother N, Collardeau F, Woody SR, Wolfe DA, Fawcett JM. Postpartum thoughts of infant-related harm and obsessive-compulsive disorder: relation to maternal physical aggression toward the infant. J Clin Psychiatry. 2022;83(2):21m14006. https://doi.org/10.4088/JCP.21m14006
9. Abramowitz JS, Deacon BJ, Olatunji BO, et al. Assessment of obsessive-compulsive symptom dimensions: development and evaluation of the Dimensional Obsessive-Compulsive Scale. Psychol Assess. 2010;22(1):180-198. https://doi.org/10.1037/a0018260
10. Fairbrother N, Albert A, Keeney C, Tchir D, Cameron RB. Screening for perinatal OCD: a comparison of the DOCS and the EPDS. Assessment. 2023;30(4):1028-1039. https://doi.org/10.1177/10731911211063223
11. Wolf N, van Oppen P, Hoogendoorn AW, et al. Inference-based cognitive behavioral therapy versus cognitive behavioral therapy for obsessive-compulsive disorder: a multisite randomized controlled non-inferiority trial. Psychother Psychosom. 2024;93(6):397-411. https://doi.org/10.1159/000541508
12. Foa EB, Liebowitz MR, Kozak MJ, et al. Randomized, placebo-controlled trial of exposure and ritual prevention, clomipramine, and their combination in the treatment of obsessive-compulsive disorder. Am J Psychiatry. 2005;162(1):151-161. https://doi.org/10.1176/appi.ajp.162.1.151
13. National Institute for Health and Care Excellence. Antenatal and postnatal mental health: clinical management and service guidance (CG192). 2014, updated 2020. Accessed October 20, 2026. https://www.nice.org.uk/guidance/cg192
Disclaimer
This article is for informational and educational purposes for clinicians and is not a substitute for individual clinical judgment, assessment or legal advice. Reading it does not create a clinician-patient relationship. If a patient is in immediate danger, call 911. For a mental health crisis, call or text 988.