Saturday, October 24, 2026

The AI Psych Brief

🔬 Section 1

Psychiatric Research Article

Caring Letters and Suicide Mortality Among US Veterans Who Contacted the Veterans Crisis Line: A Prespecified Secondary Analysis

Reger MA, Garrido MM, Lauver M, Manchester C, Palani S, Legler A, Bourgeois JE, Auty SG, et al. · JAMA Psychiatry · October 2026

Caring letters (brief, non-demanding notes of concern mailed after a crisis) are one of the few suicide-prevention interventions with a randomized trial showing reduced suicide deaths, in Motto's original study, but that result has never been cleanly replicated. The VA sent veterans who called the Veterans Crisis Line nine letters over a year, randomly signed by either a VA clinician or a peer veteran. This prespecified secondary analysis added suicide mortality data that weren't available when the parent trial was first reported. It compared 87,926 letter recipients with a historical control group of 139,576 callers from before the program started.

Across 227,502 veterans there were 625 suicides in the 12 months after the index call. Suicide mortality was 0.29% with letters versus 0.27% without (P = .31). Peer-signed and clinician-signed letters performed the same. The parent trial had previously found that letter recipients used more outpatient and inpatient mental health care.

💡 Why it matters

At the scale of a national crisis line, caring letters did not move suicide deaths, and who signs them doesn't matter. They are cheap and they increase engagement, so treat them as a bridge to care, not a stand-alone suicide-prevention intervention. They don't replace safety planning, means counseling, and a scheduled follow-up visit.

Read the paper → doi:10.1001/jamapsychiatry.2026.2866

📖 Section 2

Psychotherapy Research Article

Long-Term Psychodynamic Psychotherapy for Depression: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Moser M, Barrett A, Woll-Weber CFJ, Fonagy P, Luyten P, Campbell C · Depression and Anxiety · September 2026

This preregistered meta-analysis pooled 16 randomized trials (1,992 patients) of long-term psychodynamic psychotherapy (LTPP) for depression. Only trials of at least 9 months and at least 30 sessions were included, and the evidence was graded with GRADE and updated empirically-supported-treatment criteria. At the end of treatment, LTPP reduced depressive symptoms more than all comparators pooled (SMD 0.34, 95% CI 0.20–0.47) and more than active treatments specifically (SMD 0.27, 95% CI 0.10–0.45). Gains were generally sustained at follow-up.

The authors rated the certainty of evidence as high for the all-comparator result and moderate against active treatments, which supports a 'Strong' recommendation. Risk of bias was a concern in most trials. There was little data on functioning, cost, or long-term maintenance, and too few studies to say which patients benefit most.

💡 Why it matters

When a depressed patient has already done a course of short-term therapy and wants something longer and more exploratory, there is now graded RCT evidence for a modest but durable advantage of long-term psychodynamic therapy. Whether it's the right choice for a given patient is still a clinical judgment, because moderator data don't exist yet.

Read the paper → doi:10.1155/da/6172409

💊 Section 3

Psychiatric Fact

Modafinil's contraceptive interaction outlasts the prescription by a month

Modafinil and armodafinil are concentration-dependent CYP3A4 inducers. Ethinyl estradiol and progestins are CYP3A4 substrates, so the label warns that steroidal contraceptives may fail during treatment and for one month after discontinuation. That month is the part that gets missed: induction reverses only as new enzyme turns over. Advise a non-hormonal or non-systemic backup for the whole course plus four weeks. This matters more now that modafinil is used off-label for antipsychotic sedation, depressive fatigue, and shift-work schedules in young women. Implants and pills are both systemically exposed, so neither is immune.

At the same time, modafinil inhibits CYP2C19, which prolongs the elimination of diazepam, phenytoin, omeprazole, and clomipramine. The label specifically flags tricyclics in CYP2D6 poor metabolizers, who lean on 2C19 as the backup pathway. Their TCA levels can climb. So one drug both lowers some co-medications and raises others. Check both directions before you add it.

🛋️ Section 4

Psychotherapy Teaching Pearl

The patient's response to an interpretation is the next piece of transference

Residents usually judge an interpretation by whether the patient agrees. Agreement tells you little. Compliant patients agree with everything, and defended ones reject what's accurate. The more informative data is how the interpretation was received. Was it taken in as food, as an attack, as a gift to be admired, or as proof that the therapist finally understands and will now take over? Betty Joseph and Merton Gill each argued that the patient's reaction to an intervention re-enacts the very object relation under discussion. So the session doesn't end with the interpretation. It moves into the patient's experience of having been interpreted. Casement's 'trial identification' is the practical check: before you speak, imagine hearing the words from the patient's position. After you speak, track what they do with them.

🗒️ Vignette

A man in twice-weekly therapy describes his father's cutting corrections. The resident links this to his flinch whenever she comments on his lateness. He says, 'That's very insightful,' then goes quiet and talks about work. The resident notices the polish of the agreement and the sudden change of subject. She doesn't repeat the content. She says: 'I think what I just said may have landed like one of your father's corrections. Accurate, maybe, but something you had to take without a word.' He laughs, then admits he felt caught and wanted to hand her a good-patient answer to make it stop. The second intervention was about the experience of the first, and that is where the work moved.