The AI Psych Brief

Glucagon-Like Peptide 1 Receptor Agonists, Mortality, and Cardiovascular Outcomes in Serious Mental Illness

McIntyre RS, Zhang-James Y, Kwan ATH · JAMA Psychiatry · August 2026

This target trial emulation in the TriNetX multinational EHR network compared adults starting a GLP-1 receptor agonist with propensity-matched adults starting an SGLT2 inhibitor — an active comparator that also has proven cardiovascular benefit — in separate cohorts with and without serious mental illness (bipolar disorder, major depression, schizophrenia). The primary analysis included 195,184 matched pairs with SMI.

Among people with SMI, 4-year all-cause mortality was 4.91% with GLP-1 RA initiation vs 6.45% with SGLT2 inhibitor initiation (HR 0.76; absolute difference about 1.5 percentage points), and the absolute reduction was larger than in people without SMI. In those with type 2 diabetes, semaglutide was associated with lower 3-point MACE (HR 0.77), and the associations were driven by semaglutide and tirzepatide and held across sensitivity analyses.

Read the paper →  doi:10.1001/jamapsychiatry.2026.2574

Hearing Intervention for Older Adults With Mild Cognitive Impairment: The CHOICE Randomized Clinical Trial

Chen Y, Shi H, Xiang M, Li X, Wang Z, et al.; CHOICE Collaborative Research Group · JAMA Neurology · September 2026

In this open-label multicenter RCT in Shanghai, 703 adults aged 60 or older with both mild cognitive impairment (CDR 0.5) and moderate-to-severe hearing loss were randomized to hearing aids or hearing-care education and followed for 24 months.

The primary outcome — conversion to dementia-level impairment (CDR 1 or more) — was not significantly different: 3.09% vs 4.74% (RR 0.59, 95% CI 0.25–1.38). However, a prespecified secondary outcome favored hearing aids strongly: 15.3% vs 2.4% improved to a CDR of 0. There were no study-related adverse events. Conversion rates were low in both arms, which limited power for the primary outcome.

Read the paper →  doi:10.1001/jamaneurol.2026.3395

Gabapentin's absorption saturates — pushing the dose buys less than you think

Gabapentin is absorbed in the small intestine by the L-amino acid transporter, a saturable carrier, so its bioavailability falls as the dose rises. The product label gives about 60% at 900 mg/day, 47% at 1,200, 34% at 2,400 and only 27% at 4,800 mg/day. Going from 1,800 to 3,600 mg/day therefore roughly doubles the pill burden and sedation-prone peaks without doubling exposure. Splitting the same daily total into three or four doses absorbs better than larger, less frequent doses.

Pregabalin is the contrast case: it is absorbed by a non-saturable mechanism, roughly 90% bioavailable with linear kinetics, so a dose increase gives a predictable increase in exposure. Both drugs are eliminated unchanged by the kidney, so the variable that actually matters for toxicity is creatinine clearance — the older patient with declining renal function becomes confused or ataxic on a dose that was fine for years. Check renal function before blaming new confusion on dementia or delirium of another cause.

In MBT, move the patient to the pole they're stuck at the opposite end of

MBT describes mentalizing along four polarities: automatic–controlled, self–other, internal–external, and cognitive–affective. Healthy mentalizing moves flexibly along each; patients with borderline organization get stuck at one pole, and the stuck pole predicts the failure mode. Someone fixed at cognitive mentalizing offers fluent, plausible accounts with no felt affect — close to pretend mode — while someone fixed at affective is flooded and certain, close to psychic equivalence. MBT's technical answer is the contrary move: the therapist deliberately pulls toward the neglected pole rather than following the patient further along the one they are on. Following the patient's lead feels empathic but deepens the imbalance. The move is made gently, often with a not-knowing question, and it is a temporary corrective, not a correction of content.