This Week in Family Medicine — Sep 13, 2026

Welcome to This Week in Family Medicine. This week we're covering 9 notable papers spanning prescribing safety and risk prediction in the older and chronically ill, acute presentations in the office — chest pain, skin infections and poisoning — and a set of papers on how we communicate, treat and immunise. Let's dive in.

We'll start with prescribing safety, where a large population study in The BMJ gives us something genuinely usable. Rochon and colleagues took 65 potentially inappropriate prescribing cascades identified by an international expert consensus and tested them against health administrative data covering roughly 2.3 million community-dwelling adults aged 66 and over in Ontario. A cascade, remember, is when a drug's side effect gets misread as a new condition and treated with a second drug. Twenty-four of the 65 cascades met all three prioritisation thresholds — common starting drug, common cascade, and a real temporal signal that the second drug tends to follow the first rather than precede it. Cardiovascular drugs were the most common cascade starters. The highest-incidence cascades were iron supplement leading to a laxative, affecting about one in eight of those users; a statin leading to a pain reliever, about one in nine; and a cholinesterase inhibitor leading to a sleep agent, about one in ten. The strongest temporal signals were corticosteroid to antipsychotic and laxative to antidiarrhoeal, each roughly two and a half times more likely to run in that order than the reverse. One sex difference stood out: the antidepressant to overactive bladder drug cascade was present in men but not women. The practical move here is simple and immediate — when a new symptom appears in an older patient, ask what you started in the preceding weeks before you reach for a new prescription. [1]

Staying with chronic disease, a second BMJ paper asks whether the risk equations we lean on actually travel. Liang and colleagues built five-year cardiovascular risk equations in parallel from primary prevention cohorts of people with diabetes in New Zealand and in China — around 48,000 and 47,000 patients respectively, with several thousand first cardiovascular events in each. Most of the sixteen predictors behaved similarly across the two countries, but age did not. In the Chinese cohort each additional decade of age carried a substantially steeper increase in risk than it did in New Zealand. Crucially, the usual fix — simple recalibration of the imported equation — did not correct the miscalibration, particularly in women. Replacing the age coefficient with the locally derived one did. The lesson for anyone using an imported risk calculator in a population it wasn't built for is that recalibration alone may not be enough, and age is the coefficient most likely to be wrong. [2] And rounding out the chronic disease theme, JAMA carries a comprehensive review of rheumatoid arthritis from Smolen and colleagues, which is worth reading mainly for the timelines: aim to diagnose within about six weeks of symptom onset, start methotrexate at seven and a half to ten milligrams weekly and escalate to twenty to twenty-five milligrams within four to eight weeks, with short-course glucocorticoids tapered off by three months. About 40 percent of newly diagnosed patients reach remission on that first-line approach; adding a biologic or a JAK inhibitor lifts remission or low disease activity to around 80 percent. [3]

Turning to acute presentations, the British Journal of General Practice reports the POB-HELP trial, which tackles one of the hardest calls in general practice — chest pain in the office. Van den Bulk and colleagues ran a cluster-randomised diagnostic trial in Dutch primary care, applying a decision rule combining the Marburg Heart Score with a point-of-care high-sensitivity troponin I test in 740 patients, against 87 receiving usual care. The rule missed one acute coronary syndrome, giving a negative predictive value of essentially 99.7 percent, and for myocardial infarction specifically it missed none — sensitivity and negative predictive value both 100 percent. But specificity was only about 49 percent, and, importantly, referral rates did not differ significantly between the intervention and usual care groups. So this is a safe rule-out tool that, in this trial, did not actually reduce referrals — a finding worth holding onto before investing in point-of-care troponin. [4] Alongside that, The BMJ offers a clinical review of skin and soft tissue infections from Long and colleagues, reinforcing that laboratory testing adds little in most patients, that point-of-care ultrasound is the tool of choice when you can't tell cellulitis from abscess, and that abscesses need incision and drainage with antibiotics reserved for selected patients. Non-purulent infection is usually streptococcal, so penicillin VK, dicloxacillin or a cephalosporin remains appropriate; decolonisation for recurrent disease stays controversial. [5] And from Nature Reviews Disease Primers, a primer on toxidromes — the opioid, sympathomimetic, serotonergic, antimuscarinic, muscarinic and nicotinic patterns — which is most valuable precisely when you cannot get a history from a drowsy or agitated patient, and which reminds us that co-ingestions can blur two toxidromes together. [6]

The remaining papers cover communication, vaccination and new therapeutics. In Annals of Family Medicine, Kube and Seewald tested something we all do daily: telling a patient the tests are normal. In an online experimental study of 349 people with moderate somatic symptom burden, participants watched a videotaped doctor delivering identical normal results, but with the doctor's warmth and competence manipulated independently. When the doctor displayed both high warmth and high competence, participants rated their likelihood of serious disease significantly lower, and were less inclined to seek a second opinion or further testing. The mediation analysis suggested this worked through how much patients cognitively valued the diagnostic information. The message is that normal results do not speak for themselves — how you deliver them determines whether reassurance actually lands. [7] In The Lancet, Doshi and colleagues report a phase 4 randomised trial in Uganda of fractional yellow fever vaccine doses in nearly 1,800 children aged 9 to 23 months. Both one-fifth and one-half doses of the 17DD vaccine met non-inferiority for seroconversion at four weeks against the full dose — evidence that supports dose-sparing during outbreak response and supply shortages, relevant to anyone counselling travellers or working in outbreak settings. [8] Finally, and furthest from the clinic, Nature Medicine reports the first-in-human trial of ARO-RAGE, an inhaled small interfering RNA targeting the receptor for advanced glycation end products in the lung. In 58 healthy volunteers and 19 patients with asthma, it was safe and well tolerated, with minimal systemic exposure and dose-responsive target engagement. This is a phase 1 safety and engagement signal only — no efficacy claims yet — but it's a platform worth watching for asthma and chronic obstructive pulmonary disease. [9]

If you only have time for one paper this week, make it the prescribing cascades study in The BMJ. It converts a familiar concept into a ranked, named list of the two dozen cascades you are most likely to see in your own older patients, and you can apply it at the next medication review without any new test or tool. [1]

Here are the key takeaways from this week in Family Medicine. First, when an older patient develops a new symptom, review what you prescribed in the weeks before — iron to laxative, statin to analgesic, and cholinesterase inhibitor to sedative are among the commonest cascades. Second, imported cardiovascular risk equations may misestimate risk in populations they weren't derived in, and simple recalibration does not necessarily fix it. Third, a Marburg Heart Score plus point-of-care troponin strategy safely ruled out myocardial infarction in primary care, but did not reduce referrals. Fourth, for skin and soft tissue infection, skip the bloods, use ultrasound when abscess is in question, and drain rather than reflexively prescribe. And fifth, warmth and perceived competence materially change whether a patient accepts a normal test result — reassurance is a skill, not an announcement.

That's your roundup for This Week in Family 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 a quick rating on Apple Podcasts or Spotify helps other physicians discover the show.

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