The first-line treatments for OCD are specific forms of cognitive behavioral therapy and a particular class of antidepressants, the selective serotonin reuptake inhibitors (or SSRIs). As a practitioner, I am equally comfortable with recommending either (or both) in particular cases. But the choice between them is a complex one.
Some practitioners, and many patients and families, end up having very strong feelings on this issue.
For some, thinking of OCD as a brain illness and not 'all in my head' implies that it should be treated with a medication. Accepting that a treatment based just on talking and behavioral exercises could be effective can seem to undercut the reality or the seriousness of the illness, which to some may make it seem more like a moral failing or a form of weakness, rather than a disease. I have talked to many patients who express this view in one way or another and are not willing to countenance psychotherapy for their symptoms.
Needless to say, I disagree with this perspective. There is absolutely no contradiction between OCD (or any other mental disorder) having biological roots, with clear heritability and demonstrable correlates in the brain, and the fact cognitive-behavioral treatments can be highly efficacious. We know that experience has measurable effects on brain structure and function. Effective CBT can lead to normalization of perturbed brain circuits in patients with depression or OCD. And so being wedded to a biological model of disease in no way argues against CBT as a primary therapeutic modality.
Others are dead set against medication. This may be grounded in concern about side effects or a general aversion to drugs. Some patients I have spoken to are wedded to the idea that they should be able to overcome their symptoms through effort and that relying on medication as a 'crutch' is in some way a sign of weakness. Others have a more general sense that medications are 'unnatural'. In some cases such a concern may synergize with their OCD symptoms, with medication being seen as an intolerable form of contamination. Such concerns are understandable, and can be very powerful. My own view, however, is that anything that helps reduce symptoms and alleviate suffering can be a good thing. Medication can help patients with OCD, and in a great many cases, both anecdotally and in studies, the benefits very clearly outweigh any side effects.
Which works better? This is a more complicated question than you might think. It has been most directly asked by a group of studies from Edna Foa and colleagues at the University of Pennsylvania (and collaborators other sites) directly comparing intensive, expert CBT and standardized pharmacotherapy, in both adults and children. In adults, the results are quite clear: medication is of benefit, but patients receiving twice-weekly, expert CBT do better. However, this superiority of CBT is not so clear in children. Furthermore, the comparisons done in the most rigorous studies do not necessarily reflect typical circumstances in the real world. Typically psychotherapy in these studies is done by experts, under highly controlled conditions, at moderate to high intensity - ideal circumstances that are rarely attainable in clinical practice. Psychotherapy outcomes can be quite variable, even between expert sites. Pharmacotherapy in these studies, on the other hand, is pretty rigid, with single agents and limited options for clinicians to change medications, optimize treatment, or attempt pharmacological augmentation. Thus, while it is reasonably clear that optimal psychotherapy is superior to rigid pharmacotherapy, it is not clear how this comparison plays out in clinical practice more generally.
Things are also somewhat murky regarding combination treatment. In the large trial of adults from the Penn group, combining medication with CBT provided no additional benefit; in the study in children, on the other hand, combination therapy was clearly superior. Again, this is with fairly inflexible pharmacotherapy and optimal psychotherapy, so it's possible that the real-world benefits of medications are understated by these studies.
One issue on which there is substantial agreement in the literature (and in my clinical experience) is on the benefit of adding CBT when medications alone aren't doing the job. Again, the best studies come from the Penn group (and their collaborators at Duke, Columbia, and elsewhere). Both in adults and in children, when pharmacotherapy has led to only modest improvements (or none at all), the addition of expert CBT is of clear benefit.
My view is that, in the real world, looking across these and other studies and combining them with my own clinical experience and with that of colleagues, expert CBT and competently managed pharmacotherapy are about equally effective. Both can be of substantial benefit to 50-60% of patients, and of some moderate help to more than that. About 25-30% of patients don't receive much benefit from either.
This brings us back to where I started - with the reality that the choice of CBT or medication, or their combination, is a highly individualized one. Patient preference plays a huge role; there's no point in pushing someone towards a treatment that they don't believe in. Medication side effects can prevent individual patients from ever having a good trial of an SSRI (or other meds). An inability or unwillingness to tolerate the symptom provocation and anxiety that necessarily accompany CBT may limit an individual's ability to benefit from it, even under otherwise ideal circumstances. Speaking of ideal circumstances (and the lack thereof), the limited availability of skilled CBT therapists can be a huge factor. The fact is, medication is easier to administer with fidelity (it's not that hard for a doctor to write a prescription, or for a patient to take a pill as directed) than therapy, which needs to be done properly by a well-trained practitioner to be maximally effective. Past history can be a valuable guide - certainly, if a patient has had a good experience with CBT in the past, I'm inclined to refer them back to it, and if they've had a good response and few side effects with a particular medication, then that's a good place to start.
The last thing I want to comment on in this post is the problem of not knowing when to stop when a treatment isn't helping.
Most CBT treatment trials last from 8-16 weeks. It's clear that, in some cases, there is ongoing benefit for longer than this - there's nothing magic about an 8, 12, or 16-week trial; these are just the somewhat arbitrary durations that have been picked for studies. But at some point, response plateaus. Ideally this is after substantial improvement; but sometimes it becomes clear that a patient just isn't getting better. Continuing therapy beyond this point has drawbacks - it may be inconvenient, consume patient and health-care resources, expose the patient to ongoing exposure-induced anxiety without benefit. A much bigger problem is the foregone opportunity to consider something new - a switch of therapist or technique; addition of medication; referral to an intensive program.
(I don't mean to disparage the utility of ongoing supportive therapy in sick patients; this can be really valuable and help them manage their symptoms and their lives in important ways. I'm speaking here about treatment aimed at symptom reduction.)
The risk of continuing a futile treatment is greater with medications, because of the possibility of ongoing or cumulative side effects. It can be much easier, for a psychiatrist, to start a medication than to stop one. Stopping medications in sick patients can be a very anxiety-provoking thing to do - what if the meds are in fact helping, and I destabilize someone by stopping them? It's easier to add something new. Sometimes this can help; there is clear evidence that adding additional medications on top of standard SSRI pharmacotherapy can help some patients, and the identification of additional strategies for such pharmacological 'augmentation therapy' is a very active area of research. But all too often medication piles on top of medication, with new ones being added, or doses raised, to manage crises or side effects. This can lead to astonishing cocktails and to side effects that are every bit as bad as the original symptoms. Medications can help, but they can also harm, and they need to be managed with care. It is hard, for both psychiatrists and patients, to admit that the available options just aren't working. All too often, however, this is a fact that we must acknowledge.
We are fortunate to have treatments for OCD that often work. That wasn't true 30 years ago. But choosing among them in individual cases is often more art than science, requiring the integration of factors more complex and personal than can be captured in any formal study. I dream of the day that I can use some objective test - a brain scan, a blood test, whatever - to identify the treatment that will work best for a new patient sitting in front of me, without guesswork or trial-and-error. That day may come; it's a focus of research in my Clinic, and at other centers.
For now, the best my patients and I can do is to work together to figure out which of the alternatives is the best fit for them. We know that many will improve, but some will not. It's frustrating. But it's a start.