Medication Response in Dissociative Identity Disorder: What Does "Not Working" Actually Mean?
A woman with dissociative identity disorder isn't responding to an adequate sertraline trial for a new depressive episode. The team is divided on what "not responding" can even mean when a single symptom score may not reflect her whole system.
Eight weeks at 100mg of sertraline is, by any ordinary reading, an adequate antidepressant trial, and S.M. reports today that she feels essentially the same as she did at the start of it. In most patients that settles the question. The reason it doesn't settle it here is that the team cannot be confident whose report it is holding.
S.M. is 38 and was diagnosed with dissociative identity disorder in her late twenties, after years of fragmented memory and unexplained gaps in her sense of time that she had spent most of her twenties trying to manage quietly on her own, until a crisis brought her into sustained care. For the past ten years she has been in phase-oriented trauma therapy with the same psychologist — still in the first, stabilization-focused phase that the field's own guidelines describe, before any deeper processing work. By most functional measures she has done well. She has lived in the same one-bedroom apartment for twelve years, three blocks from the coffee shop where she does most of her freelance copyediting most weekday mornings: the same block of hours, the same table when she can get it, the same order of tasks. She has kept steady client relationships and has not been psychiatrically hospitalized in over eight years.
Two to three months ago something new appeared on top of that baseline — poor sleep, a real drop in appetite, loss of interest in editing work she had previously found satisfying, low energy most days. Both she and her therapist read it as a distinct major depressive episode rather than a fluctuation in her usual state, and her psychiatrist started sertraline eight weeks ago. The complication is that her system does not report symptoms through one consistent channel. Different self-states have, at points earlier in her treatment, described meaningfully different levels of distress about the same events — a documented phenomenon in this population, not an inconsistency in her account — and today's report came from the state that presents most consistently at appointments, which is not necessarily the state carrying the depressive episode. Declaring this trial a failure and moving on would be the textbook next step. It would also mean acting on a measurement the team has specific reason to think is partial.
Eight weeks in, deciding what "no response" means
Eight weeks at an adequate dose with no meaningful change is the standard threshold for calling this trial unsuccessful, and I don't see a dissociation-specific reason to treat it differently. No medication acts on the dissociative process itself — the depressive episode is a distinct, separately treatable target, and the ordinary next step is to optimize the dose or move to a different agent, the same as I would for any patient at this point.
I agree the depressive episode itself is a distinct, treatable target — I'm not arguing to leave it alone. What I want named directly is that the report we're working from came from one self-state, and I've documented real differences in how distress gets reported across her states before. The ISSTD's own guidelines emphasize whole-system stabilization for exactly this reason — a single visit's symptom picture may not represent what's actually happening across her whole system, and treating today's report as equivalent to a non-dissociative patient's single, consistent voice may be assuming something we don't actually know.
I don't think the answer is to wait for a fuller cross-system read before touching her medication — that risks leaving a real, measurable depressive episode undertreated while we chase a level of certainty this format may never fully give us. I'd rather act on the best available proxy now than let the genuine complexity become a reason not to act at all.
Concretely: use her observable functioning — sleep, appetite, work output, the concrete things her therapist and her editing clients can both independently track — as the working measure, increase the sertraline dose per the standard algorithm, and document the state-dependent uncertainty explicitly in her chart rather than pretending today's single report settles the question completely.
Agreed: increase the sertraline dose per the standard algorithm, using observable functioning (sleep, appetite, work output) as the working outcome measure, with an SNRI switch held in explicit reserve if the higher dose shows no improvement.
Not agreed, and left open rather than resolved: whether "observable functioning" is actually an adequate proxy for whether the depressive episode is improving across her whole system, or whether it simply substitutes one incomplete measure for another. The psychologist's concern about state-dependent reporting wasn't answered by the plan that was adopted — it was set aside as unresolvable within today's decision, with an explicit note that future visits should actively ask whether other states are describing the same trajectory before the team treats a future "response" or "non-response" as settled.