A reproductive psychiatrist argues the Clancy deadlock has put her specialty in the dock

Buried in the essay is an accidental control, and it is the most interesting thing in it. The author describes being called at 3 a.m. to evaluate a middle-aged woman with depression and suicidal thoughts, who told her she had been in that same emergency room weeks earlier and that no one had helped her ^[9]. The author had been the psychiatrist on the earlier shift. By her account, that visit included a discussion of medication options, collateral gathered from family through social work, a conversation about alcohol use as possible self-medication, an offer of voluntary hospitalization that the patient declined, and safety planning before discharge ^[10].

So one set of hours has two records: a documented care process and a patient's recollection of the same night. They diverge, which is itself the point: two honest accounts of the same hours can look like different nights entirely. One patient at one hour of one morning is too small a sample to carry a claim about a field. What it does illustrate is the measurement problem sitting under the courtroom question the author objects to, which is whether the health care system failed a particular person ^[5]. Patient recollection and clinical record are different instruments, each an interpretation of the encounter rather than a verbatim transcript of it.

The forecast itself, that the fallout will cost the patients who need help most ^[11], is a prediction about clinician behavior, and the piece does not measure it. By my count the essay offers one first-person clinical narrative and no quantitative estimate of prescribing, documentation, or disclosure ^[1]. That is what an opinion column is for. It is also why the prediction should be held rather than repeated.

Measuring it is tractable on the prescribing side. You would want monthly psychotropic fills for perinatal patients in Massachusetts, with two comparison series: non-perinatal psychiatric prescribing in the same clinics, which controls for practice-wide caution and staffing changes, and perinatal prescribing in states where the coverage was thinner. A dip with no comparison series tells you nothing, because perinatal prescribing moves for reasons that have nothing to do with any trial.

The disclosure side is harder, and this is where I would not expect data soon. There is no registry of the sentence a woman decides not to say in a screening room. You can proxy it with screening-positive rates or with time from symptom onset to first visit, but heavy coverage of a case like this plausibly pushes in both directions at once: more people have a name for what is happening to them, and more people suspect that saying the name aloud invites consequences beyond the clinic. Those two effects can cancel in an aggregate number while landing very differently on the patient who was already least likely to come in.

What the essay does establish, on its own authority, is the shape of the work being judged. The author's account is that clinicians can do everything right with the tools available and still have things go catastrophically wrong ^[6], and that reproductive psychiatry involves tracking hormones, sleep, trauma, medical conditions, relationships and the demands of daily life as they interact and change ^[7]. She also says the defense portrayed psychiatrists as spending too little time with patients and trying too many drugs ^[8]. That is her characterization of an adversarial proceeding, not a neutral summary of it.

My view, with its condition attached: the documentation half of her prediction is the credible half, because defensive charting responds to narrative rather than to evidence and needs no verdict to begin. The prescribing half I would treat as unmeasured until someone runs it against a control series.