This Week in General Medicine — Aug 10, 2026

Welcome to This Week in General Medicine. This week we're covering seven notable papers spanning psychiatric drug trials with controlled substances, global health and adherence interventions, and population-level trends in stroke and infectious disease. Let's dive in.

We'll start with psychiatry, where two trials in Nature Medicine test compounds that have historically sat outside mainstream prescribing. In the first, a multicentre, double-blind, placebo-controlled trial randomised 171 adults with post-traumatic stress disorder and recurrent nightmares to dronabinol, synthetic delta-9-tetrahydrocannabinol, at doses between two-and-a-half and fifteen milligrams nightly, or to placebo, for ten weeks [2]. The primary endpoint was the nightmare item on the Clinician-Administered PTSD Scale, and dronabinol beat placebo with a between-group difference of one-and-a-half points and a moderate effect size of about 0.65. The safety picture deserves attention: adverse events were common in both arms, roughly ninety percent with dronabinol versus seventy-seven percent with placebo, discontinuation rates were similar, but seven patients on dronabinol had serious adverse events compared with none on placebo. The authors are explicit that long-term efficacy and safety remain unknown. Alongside that, a National Health Service site in England ran a two-arm, double-blind feasibility trial of a single twenty-five milligram dose of psilocybin versus placebo, with psychological preparation, dosing support and integration, in sixty patients with treatment-resistant major depression [4]. This was designed as a feasibility study, so the formal primary outcomes were recruitment, retention and variance estimation, and on those it succeeded, with fifty-nine of sixty participants completing all follow-up assessments. The depression signal was substantial, with roughly a ten-point advantage on the Montgomery-Åsberg scale at week three, sustained to week six. But read this for what it is: a sixty-patient feasibility trial in a single public healthcare setting, powered to inform a confirmatory trial, not to change practice. The common thread across both papers is that these agents are moving into conventional randomised evaluation within public health systems, and both research teams are appropriately cautious.

Turning to global health, two trials ask whether well-intentioned interventions actually deliver. In the BMJ, a pragmatic randomised trial across four public reference hospitals in Buenos Aires enrolled 555 adults with newly diagnosed drug-susceptible tuberculosis and added a patient-centred digital adherence tool to standard care: a mobile app with daily adherence reporting, two-way messaging with treatment supporters, educational content, and a weekly urine-based isoniazid test to verify adherence [3]. Treatment success rose from about seventy-four percent to eighty-two percent, an absolute gain of roughly seven percentage points, and loss to follow-up fell from about a quarter to about a sixth, roughly a thirty percent relative reduction. Benefit was greater in women and in those under thirty-five. Both results sat just at the edge of statistical significance, so this is a modest but real effect from a scalable, low-resource-friendly intervention. Contrast that with the IRMA trial in JAMA, which delivers a negative result against a World Health Organization recommendation [6]. More than two thousand six-month-old infants in rural, malaria-endemic southern Malawi were randomised four ways: iron syrup plus malaria chemoprevention, iron-containing micronutrient powders plus chemoprevention, chemoprevention alone, or placebo. After six months, neither iron intervention improved the Bayley cognitive composite score compared with chemoprevention alone, and chemoprevention alone did not significantly improve cognition compared with placebo. Iron raised ferritin but, notably, did not reduce anaemia prevalence. The reassuring finding is on safety: neither iron strategy increased malaria risk, and chemoprevention did reduce malaria incidence. So universal preventive iron in this setting is safe but, on these data, functionally ineffective for neurodevelopment, which is a genuine challenge to current guidance.

Our third theme is population-level disease burden. In PLOS Medicine, a record-linkage study across primary care, hospital admissions, death certificates and a national stroke audit captured nearly 644,000 first-ever strokes in England between 2018 and 2024 [5]. In adults under fifty-five, stroke incidence rose about seven percent during the pandemic period and continued climbing another four percent afterwards, while incidence in those fifty-five and over was flat during the pandemic and rose modestly after. The increase in younger people was most marked in ethnic minority groups, with roughly a thirteen percent rise in Black adults and a twenty-six percent rise in Asian adults. Crucially, the prevalence of modifiable risk factors barely changed, but undiagnosed pre-stroke hypertension rose by about seven percent and undiagnosed diabetes by about twenty-two percent in the under-fifty-fives, and that has not recovered. This is a primary care detection failure with a hard clinical endpoint attached. Rounding out the theme, Nature Reviews Disease Primers published a comprehensive review of influenza, reminding us that seasonal epidemics cause an estimated three to five million severe cases and between 290,000 and 650,000 deaths a year, that antigenic shift through reassortment with avian and swine viruses drives unpredictable pandemics, and that management still rests on annually updated vaccines and early neuraminidase or polymerase acidic protein inhibitor therapy [7]. And in oncology, the Lancet reported the PSMAddition trial, in which 1,144 men with PSMA-positive metastatic androgen pathway modulator-naive prostate cancer were randomised to lutetium-177 PSMA-617 added to androgen deprivation plus an androgen receptor pathway inhibitor, or to that backbone alone [1]. Adding the radioligand cut the relative risk of radiographic progression or death by twenty-eight percent, meeting the primary endpoint, though median progression-free survival was not reached in either arm. Grade three or worse adverse events were somewhat more frequent, fifty-one versus forty-three percent, and dry mouth affected nearly half of treated men versus four percent of controls.

If you only have time for one paper this week, make it the England stroke incidence study in PLOS Medicine [5]. It links a rising stroke burden in adults under fifty-five directly to undiagnosed hypertension and diabetes, and that is something you can act on in your next clinic session.

Here are the key takeaways from this week in General Medicine. First, stroke is rising in the under-fifty-fives, disproportionately in Black and Asian patients, alongside worsening detection of hypertension and diabetes, so check blood pressure and glucose opportunistically in younger adults. Second, dronabinol modestly reduced PTSD nightmares but with serious adverse events only in the active arm, and psilocybin for treatment-resistant depression remains at the feasibility stage. Third, a digital adherence app improved tuberculosis treatment success by about seven percentage points and cut loss to follow-up, a practical option in low-resource programmes. Fourth, universal iron supplementation in young Malawian children did not improve cognition and did not reduce anaemia, though it did not increase malaria risk. And fifth, lutetium-177 PSMA-617 added to standard hormonal therapy delayed radiographic progression in metastatic hormone-sensitive prostate cancer, at the cost of more toxicity, notably dry mouth.

That's your roundup for This Week in General Medicine. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if you have a paper you have been meaning to read, upload the PDF, or paste any link, at audioscholar dot C C. We will turn it into audio like this one, in any of thirty-one languages.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.