Thursday, October 1, 2026

The AI Psych Brief

🔬 Section 1

Psychiatric Research Article

Prior Sexual Violence, PTSD Symptoms, and Recovery Trajectories After New Trauma

Mellen EJ, Poplin T, Jovanovic T, et al.; Harnett NG, Ressler KJ · JAMA Psychiatry · July 2026

The investigators drew on AURORA, a large US cohort of adults recruited in emergency departments within 72 hours of a traumatic injury (mostly motor vehicle crashes). They followed 2,423 people for a year with the PCL-5 at 2 weeks, 8 weeks, 3, 6, and 12 months, and compared those who reported an earlier history of sexual violence with those who did not.

More than half of the sample (52%) disclosed prior sexual violence. At every time point, that group had higher PTSD symptoms after the new, nonsexual trauma (about 4.6 PCL-5 points higher), even after adjusting for other earlier nonsexual trauma. Hypervigilance (criterion E) stayed the most elevated symptom cluster, while negative cognitions and mood (criterion D) recovered most slowly.

💡 Why it matters

When a patient arrives after a car crash or assault, ask about earlier sexual violence. It marks a slower, more symptomatic recovery, so it is a reason to follow up early and to watch for the lingering criterion D picture of guilt, detachment and a foreshortened sense of self, rather than just the startle response.

Read the paper → doi:10.1001/jamapsychiatry.2026.2219

🩺 Section 2

General Medicine Article

QTc changes and early major adverse cardiovascular events associated with antidepressant treatment for major depressive disorder in adults: individual participant data network meta-regression of double blind randomised trials

Ostinelli EG, Capocci S, Li Z, Friederich C, Mugnai G, Pottegård A, Ribichini FL, Cipriani A · BMJ · September 2026

This Oxford–Verona group pooled individual patient ECG data from 35 double-blind trials (8,679 adults with major depression) covering placebo and 10 antidepressants. They modelled the change in QTcF over the first eight weeks, adjusting for baseline QTc, age, sex, BMI, potassium and eGFR. A separate aggregate analysis of 139 trials (52,398 participants) looked at early major adverse cardiovascular events.

Compared with placebo, escitalopram (+8.7 ms) and amitriptyline (+5.3 ms) were the only drugs that meaningfully lengthened QTc. Venlafaxine and vortioxetine were most often associated with the smallest intervals. The effect varied a lot between individuals, especially for escitalopram, fluoxetine and mirtazapine. Across the larger trial pool, antidepressants were not associated with more early MACE or non-suicidal sudden death than placebo (risk difference 0.01%).

💡 Why it matters

This is internal-medicine-grade evidence for a question psychiatrists field from cardiology. For most patients the absolute cardiac risk of starting an antidepressant is very low. The QTc effect is still drug-specific and depends on the patient, so a patient with a long baseline QTc, low potassium or reduced renal function should push you away from escitalopram and tricyclics, not away from antidepressants.

Read the paper → doi:10.1136/bmj-2026-100600

💊 Section 3

Psychiatric Fact

Fluoxetine's washout is measured in norfluoxetine half-lives, not fluoxetine's

Stopping fluoxetine does not end its effects on the day of the last dose. Fluoxetine's own half-life is 1 to 4 days acutely (it lengthens with chronic dosing because the drug inhibits its own clearance). Its active metabolite norfluoxetine has a half-life of about 1 to 2 weeks, so it takes roughly five weeks to clear. That is why the label requires at least 5 weeks off fluoxetine before starting an MAOI, compared with 2 weeks for most other SSRIs. The same tail affects any cross-taper. Both fluoxetine and norfluoxetine are potent CYP2D6 inhibitors, so a patient can stay a functional poor metabolizer for weeks after stopping. If you switch to a 2D6 substrate such as aripiprazole, vortioxetine, atomoxetine, a tricyclic or metoprolol, hold back on dosing until the inhibition fades. Otherwise the patient gets excess exposure early, and the new drug seems to 'stop working' a month later when levels fall. The practical upside is that the long tail acts as a self-taper, which is why fluoxetine rarely causes discontinuation syndrome.

🛋️ Section 4

Psychotherapy Teaching Pearl

Transference work helps most where it looks riskiest

Many trainees believe here-and-now transference interpretation should be saved for high-functioning patients, and that fragile patients need support instead. Høglend's First Experimental Study of Transference Work (FEST) tested this directly. Patients were randomized to dynamic therapy with or without moderate transference interpretation. The benefit was concentrated in patients with low quality of object relations, those whose relationships had been few, unstable or hostile. Well-functioning patients did about as well without it. One likely reason is that a patient with poor object relations cannot yet think about their relationships outside the room, while the relationship with the therapist is live, shared and observable. What matters is the dose and tone: a few interpretations a session that are tentative and linked to the patient's outside life, not a steady stream of 'you're doing it with me.'

🗒️ Vignette

A 34-year-old man with a string of broken friendships says he doesn't know why people 'always end up flaking.' Asking about the friendships gets nowhere. In the fourth session he is late, remarks that 'this probably isn't going anywhere either,' and waits. Instead of reassuring him or going back to the friends, the resident says tentatively: 'I wonder if you're already getting ready for me to flake too, and it feels safer to decide that first.' He goes quiet, then says he usually leaves before he can be left. That abstract pattern is now happening in the room, where both of them can look at it. Over the following weeks it gives the resident a way back into the friendships the patient could not describe.