Physician in the Loop

Most of next year's Medicare cut to doctors is written into law. Congress should tie the yearly update to what practices pay, and the new same-day cut will move appointments more than money.

Read the special topic and its sources: https://physicianintheloop.org/specials/medicare-2027-pay-rule/

The voice on this episode is generated by AI. It reads a special topic from Physician in the Loop as written; every fact in it links to its source.

Stopping statins after 70 meets noninferiority but adds 1 event per 40 patients

The voices on this episode are generated, and so is the script, from the day's post at Physician in the Loop. Every fact in the post links to its source.

Read the post and its sources

Get the free email briefing

Adults 70 and older who stopped primary-prevention statins had more heart attacks, strokes and deaths, but within a preset noninferiority margin, a 1,880-patient trial reported Nov. 7 in the New England Journal of Medicine.

The STREAM trial enrolled people 70 and older with at least two chronic conditions and no known atherosclerotic disease, and the investigators estimated one additional major cardiovascular event or death over two years for every 40 people who stopped, according to HCPLive. Study author Dr. Rita Redberg called for shared decision-making rather than a blanket recommendation to stop, the report said.

Also at the American Heart Association's Scientific Sessions, Cleerly reported that its AI analysis of coronary CT beat stress-test-guided care at 180 days in a randomized trial whose results have not been peer reviewed. The Food and Drug Administration asked for feedback on robotic devices that act on their own or are run from a distance, and a 1,706-patient trial in Colombia and Jamaica found single-pill combination therapy with team-based care cut systolic pressure 10.1 mm Hg more than usual care.

In this episode

• Stopping statins in adults 70 and older without heart disease met noninferiority in a 1,880-patient trial, though events rose to 12.7% from 9.2%
• AI plaque analysis on coronary CT beat stress-test-guided care at 180 days in a 530-patient trial, the company reports
• FDA asks how to weigh the benefits and risks of robots that operate on their own or are run from afar
• Single-pill combinations delivered through team-based care cut systolic pressure 10.1 mm Hg more than usual care in a 1,706-patient trial
• FDA grants accelerated approval to subcutaneous amivantamab for head and neck cancer after platinum and immunotherapy, with a 42% response rate

Stents or bypass for stable heart disease add no survival at 11 years

The voices on this episode are generated, and so is the script, from the day's post at Physician in the Loop. Every fact in the post links to its source.

Read the post and its sources

Get the free email briefing

Routine angiography and revascularization for stable coronary disease with moderate or severe ischemia did not extend survival over 11 years compared with medicines alone, according to ISCHEMIA follow-up data JAMA published Saturday.

The same day at the American Heart Association's Scientific Sessions in Chicago, full results showed that pelacarsen, which cut lipoprotein(a) by 78%, and the interleukin-6 antibody ziltivekimab, which lowered CRP, did not reduce major cardiovascular events. A randomized trial of AI-screened ECGs found a statistically fragile gain in atrial fibrillation detection, and a Stanford study using a language model found lipid therapy was left unchanged at 89% of visits where LDL was above goal.

Weill Cornell Medicine disclosed that patient imaging details were entered into AI tools without authorization for a radiology research study in June.

In this episode

• Routine stents or bypass for stable coronary disease add no survival over medicines alone at 11 years in 5,179-patient ISCHEMIA follow-up
• Two heart drugs fail in full results: pelacarsen cut lipoprotein(a) 78% and ziltivekimab cut CRP, but neither prevented heart attacks or strokes
• An AI-flagged ECG raised detection of atrial fibrillation or its precursors by a fragile margin in a 14,665-adult randomized trial
• A language model reading clinic notes found lipid therapy was not intensified at 89% of visits where LDL sat above goal
• Starting statin plus ezetimibe up front did not cut cardiovascular events at 3 years versus a statin alone in a 3,048-patient trial
• Weill Cornell says patient imaging details were fed into AI tools without authorization for a radiology research study

U.S. measles cases pass 5,000, the most since 1991

The voices on this episode are generated, and so is the script, from the day's post at Physician in the Loop. Every fact in the post links to its source.

Read the post and its sources

Get the free email briefing

The United States has recorded 5,182 confirmed measles cases in 2026, the most since 1991, federal data released Friday showed, days before a regional review that could end the country's measles elimination status.

The count is more than double the 2,289 cases of 2025, and children and teenagers make up about 59% of them, according to Centers for Disease Control and Prevention figures reported by Bloomberg and Al Jazeera. The Pan American Health Organization said in March that it would review U.S. elimination status at its November meeting; a country loses the status after 12 months of uninterrupted spread of one viral lineage.

The Food and Drug Administration also opened an investigation of tumors in patients treated with the spinal muscular atrophy gene therapy Zolgensma, and HCA Healthcare said about 10,000 of its physicians will be documenting with ambient AI by year's end.

In this episode

• U.S. measles cases reach 5,182, the most since 1991, as elimination status goes up for review next week
• FDA investigates tumors after Zolgensma gene therapy, with viral DNA found inserted near a cancer gene in two patients
• HCA Healthcare puts ambient AI in the hands of 10,000 physicians by year's end and an AI nurse handoff tool in 100 hospitals
• An AI flag on positive ANA tests cut time to autoimmune diagnosis from 138 to 43 days in a Vanderbilt randomized trial
• Medicare cuts 340B drug payments to sales price minus 28% and pays office rates for hospital imaging off campus from 2027
• Ochsner plans primary care panels of 3,000 to 5,000 patients built around AI agents, pharmacists and ambient listening
• Starting overactive bladder patients on beta-3 agonists instead of anticholinergics did not lower dementia risk in an 830,393-patient cohort

In 2022 one emergency department in thirteen lacked an attending physician around the clock. Every rural one should have an emergency physician on video at all hours, paid for by Medicare rather than grants.

Read the special topic and its sources: https://physicianintheloop.org/specials/rural-ed-tele-emergency/

The voice on this episode is generated by AI. It reads a special topic from Physician in the Loop as written; every fact in it links to its source.

AI was sold to medicine as a way to save time and money. This week's evidence suggests it often spends both, and it should be paid for only when it saves them.

Read the letter and its sources: https://physicianintheloop.org/letters/2026-11-02/

The voice on this episode is generated by AI. It reads the week's Friday letter at Physician in the Loop as written; every fact in the letter links to its source.

Smartwatch AI spots more hypertension than an office cuff check

The voices on this episode are generated, and so is the script, from the day's post at Physician in the Loop. Every fact in the post links to its source.

Read the post and its sources

Get the free email briefing

A Google artificial intelligence model reading consumer smartwatch sensors detected 65.8% of hypertension cases over seven days, versus 55.3% for an initial office blood pressure check, in a 196-person study published Thursday in Nature.

Both approaches had 90.0% specificity against 24-hour ambulatory monitoring, and the watch model needed no cuff calibration, the authors reported. Separately, a JAMA study of 5.2 million newborns found that the share not given vitamin K at birth more than doubled from 2017 to 2025, Wisconsin reported its first measles death since 1990, and the Centers for Medicare & Medicaid Services proposed dropping obstetric and emergency readiness training rules for hospitals.

In AI evidence, a 400-person randomized trial in JAMA Network Open found a generative AI maternal health chatbot did not improve anxiety or depression at 12 months, and Mayo Clinic Platform opened a clinical AI chatbot covering more than 35 specialties to Mayo clinicians and outside organizations in its programs.

In this episode

• Google's smartwatch AI catches 65.8% of hypertension cases in a week, beating an initial office cuff reading at 55.3%
• Newborns who skip the vitamin K shot have more than twice the risk of intracranial bleeding, a 5.2 million-infant study finds
• An infant dies of measles in Wisconsin, the state's first measles death since 1990, as cases reach 431
• A Washington dairy farmworker dies of bird flu, the first U.S. avian influenza death of 2026
• Medicare proposes scrapping hospital obstetric and emergency readiness training rules and asks about loosening CRNA supervision
• A generative AI chatbot for pregnant and postpartum patients fails to improve anxiety or depression in a 400-person trial
• Mayo Clinic opens an AI chatbot covering more than 35 specialties to its clinicians and partner organizations
• FDA clears PulseAI's model that reads atrial fibrillation on smartwatch and other wearable ECGs
• Electronic record alerts fail to cut routine testing before cataract surgery in a UCLA randomized trial

Measles cases this year are more than twice last year's. Texas is investigating the pediatricians who vaccinate children, and the children who depend on those doctors will pay.

Read the special topic and its sources: https://physicianintheloop.org/specials/measles-and-the-investigations/

The voice on this episode is generated by AI. It reads a special topic from Physician in the Loop as written; every fact in it links to its source.

Medicare cuts physician pay 2.31% for 2027, deeper than it proposed

The voices on this episode are generated, and so is the script, from the day's post at Physician in the Loop. Every fact in the post links to its source.

Read the post and its sources

Get the free email briefing

The Centers for Medicare & Medicaid Services cut the physician fee schedule conversion factor 2.31% for 2027 on Wednesday, to $32.63, a deeper reduction than the 1.68% it proposed in July.

Clinicians in qualifying alternative payment models take a 1.83% cut. Most of the decline reflects the expiration of a one-year 2.5% increase for 2026, partly offset by statutory updates of 0.25% and 0.75%, Medical Economics reported. The same rule pays 20% less for a visit billed the same day as a minor procedure, replaces the G2211 add-on with modifiers worth 16% or 32% of a visit, and tightens remote monitoring billing. A companion outpatient rule raises hospital rates 2.3% and creates Medicare's first payment category for AI and other software services.

"Physicians will once again begin the year with lower rates as practice costs continue to rise," the American Academy of Family Physicians said, according to Chief Healthcare Executive. CMS Administrator Dr. Mehmet Oz said fee-for-service payment "has rewarded volume over value."

In this episode

• Medicare cuts the physician conversion factor 2.31% for 2027 and pays 20% less for visits billed the same day as minor procedures
• Medicare creates its first payment category for AI and other software services and moves 14 algorithm-only lab tests out of the lab fee schedule
• Medicare limits remote monitoring to established patients and delays its ban on outside monitoring vendors to 2029
• Federal regulators push out two surprise-billing arbitrators, one of which sided with clinicians in 89% of 2023 cases
• A South Dakota health system blocks OpenEvidence after finding a physician storing more than 100 patient records in it
• Patients of athenahealth practices can now send their records into Claude, and the company says HIPAA does not apply
• Philips sets aside $322 million to settle Justice Department criminal and civil probes of its 15 million-device CPAP recall
• Pennsylvania confirms a sixth measles death as its outbreak reaches 1,467 cases and 283 hospitalizations

Theranos is remembered for defrauding investors. For doctors the lesson is what happened to patients. Products that change care still reach them on their makers' word, before the evidence.

Read the special topic and its sources: https://physicianintheloop.org/specials/theranos-lessons/

The voice on this episode is generated by AI. It reads a special topic from Physician in the Loop as written; every fact in it links to its source.