How Long an SSRI Trial in OCD Must Run Before Calling It a Failure
Six weeks of fluoxetine and no clear improvement looked like a failed trial to the covering resident — but OCD's own antiobsessional response window runs nearly twice that long.
A.K., a 27-year-old woman, is in the third year of a biology PhD program and shares a house with two other graduate students; her days are largely dictated by a research schedule she describes as "already unforgiving before any of this." She has no significant medical history and takes no other medications.
She was diagnosed with OCD eight months ago after her advisor noticed she was repeatedly re-running the same completed lab protocol steps, convinced she had contaminated a reagent despite no objective indication of it. Fluoxetine was started six weeks ago at 20 mg and titrated to 60 mg over the first three weeks, and her Y-BOCS today is 24 — down only two points from her pre-treatment score of 26. The resident covering her case this week, unfamiliar with her longer history, documented this as "minimal response to an adequate trial" and drafted a plan to switch her to sertraline starting next visit.
The trial isn't actually finished. Depression trials typically show most of their treatment effect by four to six weeks, which is where the six-week convention that guides so much of general psychiatric prescribing comes from — but OCD's antiobsessional response follows a genuinely slower and later curve. Randomized trials in OCD consistently show continued, sometimes accelerating improvement between weeks six and twelve that simply isn't present yet in her chart, and most OCD treatment guidelines specify eight to twelve weeks at a therapeutic dose, not six, before a trial can be called inadequate. A two-point Y-BOCS drop at six weeks on a fully titrated dose is not evidence the drug isn't working — it is roughly what the trajectory looks like at this point in a trial that eventually succeeds. Switching her now would restart a clock that hasn't finished running, on a drug she has tolerated without any reported side effects.
Is this trial actually finished
This isn't a failed trial, it's an unfinished one. OCD's antiobsessional response consistently lags behind the general antidepressant response curve — a two-point Y-BOCS improvement at six weeks on a fully titrated dose, with no side effects reported, is close to what the literature shows at this point in trials that go on to succeed by week twelve. Switching her now discards data that hasn't finished accumulating.
I agree with continuing, and I'd add the practical reason this matters beyond her individually: switching prematurely doesn't just risk losing a drug that might have worked, it restarts an entire new trial-length clock on a different agent, in a patient whose functioning is already under real strain from her research schedule. The six-week convention is a depression-trial habit, not an OCD one, and it's an easy one for a covering resident unfamiliar with the case to reach for by default.
I'd also flag the ERP waitlist directly in the plan. She's tolerating the medication well, so there's no reason not to continue it through the full twelve-week window, but if ERP doesn't actually start until close to that reassessment date, we should be honest that we're evaluating medication alone for most of this trial, not the combined treatment we'd ultimately want her on.
Agreed: fluoxetine continued unchanged at 60 mg, with a formal Y-BOCS reassessment scheduled at week 12 rather than today, and the chart note corrected to reflect an ongoing, not failed, trial. Her ERP referral status will be tracked explicitly so the eventual week-12 reassessment can note whether she received medication alone or combined treatment.
The group agreed the switch plan should not have been drafted at six weeks in the first place, and flagged the six-week convention as worth naming explicitly in her chart so it isn't repeated by another covering provider before week twelve arrives.