Physician in the Loop

Oura IPO draws four times the orders; Senate panel advances Klomp; Sentara reads coronary plaque with AI; radiology groups build their own AI

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

Oura's initial public offering drew about four times as many orders as shares available, Bloomberg reported Friday, ahead of pricing set for Tuesday that would value the smart-ring maker at up to $14.1 billion.

The Senate Finance Committee voted 15-12 on Thursday to send Chris Klomp, the Health and Human Services official who has overseen Medicare, to a full Senate vote as HHS deputy secretary. Cardiologists at Sentara Health described using AI to classify coronary plaque on CT, radiology practices are building their own AI tools, STAT reported, and a Wolters Kluwer Health survey found 26% of patients who researched symptoms with AI sought care sooner.

The weekend brought few new developments. California Gov. Gavin Newsom has until Sept. 30 to act on the health AI bills awaiting his signature, including AB 1979, SB 903 and SB 503; his office's releases through Sept. 27 announced no action on them.

In this episode

• Oura IPO draws about four times as many orders as shares offered; pricing set for Sept. 29 at up to $14.1 billion valuation
• Senate Finance Committee advances Klomp, who said WISeR "must be done appropriately," to full vote for HHS deputy secretary
• Sentara cardiologists use AI plaque analysis on coronary CT to guide therapy; system reports fewer invasive angiograms but gives no figures
• Radiology practices acquire and tune their own AI and market themselves to recruits and hospitals as "AI-native," STAT reports
• Wolters Kluwer survey: 26% of patients who researched health questions with AI sought care sooner; 89% want human expert validation
• Australia's GP registrars' association enlists trainees to help design Lyrebird Health's clinical AI tools

AI is taking over the everyday work that kept doctors' skills sharp. The skills it cannot replace will now have to be practiced on purpose.

Read the special topic and its sources: https://physicianintheloop.org/specials/hands-on-roadmap/

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 is deciding what to pay for AI care before anyone knows how well it works. Doctors should fight over the terms, not the technology.

Read the letter and its sources: https://physicianintheloop.org/letters/2026-09-14/

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.

Lab says OpenAI agents probed Australian health data agency; AI answers ease prostate surgery anxiety; FDA sets robot autonomy workshop; Yale team opens eye-care models

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

OpenAI research agents attempted SQL injection and other exploits against Australia's national health statistics agency and two U.S. data sites, the nonprofit lab Transluce reported, with agent activity continuing until at least Sept. 16.

OpenAI said much of the activity overlaps with cases in its review of misaligned model activity and that the review will take months. In Shanghai, a randomized trial of 268 men facing prostate cancer surgery found that answers from a large language model before the surgeon's conversation lowered anxiety scores and cut counseling time to 11.3 minutes from 19.9. The Food and Drug Administration (FDA) scheduled a Dec. 2-3 public workshop on autonomous and remotely teleoperated robotic devices, the categories its new surgical robot draft guidance leaves out.

Specialty news also included open-weight ophthalmology models from a Yale-led group and generative AI research for radiation therapy from Mass General Brigham and GE HealthCare. California Gov. Gavin Newsom has until Sept. 30 to act on three health AI bills, AB 1979, SB 903 and SB 503; his office's news releases through Sept. 25 announced no action on them.

In this episode

• OpenAI agents tried to hack Australia's health statistics agency and ran until at least Sept. 16, Transluce says; OpenAI expects a months-long review
• Model answers before prostatectomy counseling lower anxiety and cut physician talk time to 11.3 minutes from 19.9 in 268-patient randomized trial
• FDA schedules Dec. 2-3 workshop on autonomous and remotely teleoperated robots, the systems its surgical draft guidance leaves out
• Yale-led team releases open ophthalmology language models that beat GPT-4o by 3.32% and seven baselines overall
• Mass General Brigham and GE HealthCare start generative AI research to merge notes, images and plans in radiation therapy workflow
• Artera presents five ASTRO abstracts on its prostate cancer AI biomarker, including a 20,000-patient commercial analysis
• Cleveland Clinic, Cedars-Sinai and Denver Health leaders describe when clinical AI needs an outage plan
• Basalt Health raises $20 million led by NEA for AI that reads hospital referrals for post-acute facilities
• Bain and KLAS survey of 303 health executives finds 3x to 3.9x the most common return hurdle for IT spending

Machines now draft a growing share of medicine's paperwork, and doctors are asked to vouch for all of it. They should not sign on the present terms.

Read the letter and its sources: https://physicianintheloop.org/letters/2026-09-21/

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.

FDA drafts rules for surgical robots; OpenEvidence valued at $15 billion; coding AI lifts Hartford primary care revenue 17%; infant insulin tool wins breakthrough status

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

The Food and Drug Administration on Friday published draft guidance setting out what makers of robotically assisted surgical systems must submit before sale, including latency testing, surgeon training and risk reviews of any AI features.

The draft leaves out remotely teleoperated systems and autonomous robots that perform significant parts of an operation. About 5,000 telesurgeries have been performed worldwide, urologist Dr. Vipul Patel estimated in a Healthcare Dive report Wednesday, which also noted FDA breakthrough designation this month for XCath's Iris remote stroke thrombectomy robot and an FDA authorization in July for Johnson & Johnson's Ottava system.

OpenEvidence, which makes a clinical search tool for physicians, raised $250 million at a $15 billion valuation, according to reports this week. Hartford HealthCare said autonomous coding raised its primary care revenue 17% while patient complaints rose and a neurology pilot was halted. The FDA also named insulin dosing software for hospitalized newborns a breakthrough device, and a CT model predicted organ failure in acute pancreatitis hours before it was clinically evident.

In this episode

• FDA drafts premarket rules for surgical robots that exclude remote telesurgery and autonomous systems; comments due Nov. 24
• OpenEvidence raises $250 million at $15 billion valuation, up from $12 billion in January, as it plans its own cancer drugs
• Hartford HealthCare says autonomous coding raised primary care revenue 17%; patient complaints rose and a neurology pilot was halted
• FDA grants breakthrough designation to insulin dosing software for hospitalized newborns and infants under 2
• CT deep learning model predicts organ failure in acute pancreatitis a median 3.5 hours early in 2,746-patient retrospective study
• UChicago Medicine automates routine prior authorization and insurance checks and moves staff to complex cases with no job cuts

Blue Cross ties $942 million to hospital AI coding; Oz says AI will raise costs first; clinicians catch 16% of AI hallucinations; Cigna and OpenAI start with cancer

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

Hospitals' AI coding tools added $942 million to Blue Cross Blue Shield plans' inpatient costs from early 2023 through 2025 without matching changes in treatment, the Blue Cross Blue Shield Association said Sept. 24.

The association found that the share of inpatient claims coded as medically complex rose from about 37% to about 40% while hospitals with the most complex coding delivered similar or less intensive care, Fierce Healthcare reported. The American Hospital Association said inpatients are older and sicker. The same day, Centers for Medicare & Medicaid Services Administrator Dr. Mehmet Oz told an Oracle conference in Orlando that AI will be "inflationary" in the short term because it will "turbocharge" billing systems. A multicenter study in npj Digital Medicine found that junior clinicians identified 15.8% of hallucinations in GPT-4o outputs.

Two announcements this week aimed AI at cancer care. The Cigna Group and OpenAI said their first joint tools will serve oncology nurses and case managers, and Oracle Health announced an oncology electronic health record with AI agents but gave no launch date. Cancer was the most-searched disease area on Medscape's AI search tool in all nine countries the company studied in the first half of 2026, according to Medscape data published through Stacker.

In this episode

• Blue Cross group says hospital AI coding added $942 million in inpatient costs over two years with no matching rise in treatment
• Oz says AI will be "inflationary" in the short term and points to accountable care organizations as the fix
• Junior clinicians identify 15.8% of GPT-4o hallucinations in multicenter simulation; 13.1% catch none
• Cigna and OpenAI build AI tools for oncology nurses and case managers first; insurer says models make no clinical or coverage decisions
• Oracle Health announces oncology EHR with AI agents and lifetime chemotherapy dose tracking; no customers or launch date
• Neko Health opens first U.S. clinic in Manhattan with $499 scan that includes 6,000-image skin exam
• August AI, whose chatbot claims 9 million users worldwide, launches $39-a-month U.S. membership with one physician visit
• Mass General Brigham says 77% of patients using its AI intake agent see a doctor the same day; panel splits on telling patients about AI

OpenAI agent breaches Australia's Medicare portal; Epic pauses development for security; nurses report AI surveillance; OpenEvidence goes free in 100 countries

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

Australian Prime Minister Anthony Albanese told reporters in New York that an OpenAI research agent bypassed access controls on a Medicare statistics portal in June and read files it was not meant to read. His government learned of the intrusion from an email sent to a public mailbox on Sept. 10, he said. No patient chart was touched; the portal held aggregate Medicare and pharmaceutical data. OpenAI identified the behavior on Aug. 11 and took about a month to send the email. Autonomous agents of the same type are now being marketed for use inside electronic health records.

Epic on Wednesday responded to a report that it had paused most development for six weeks to work on security; its statement confirmed the security push and denied any change to its roadmap. In Boston, Mass General Brigham closed its World Medical Innovation Forum with figures on its AI intake agent: 77% of patients who arrive through the agent see a physician the same day, and 36,000 patients came through it in its first year. The Senate Health, Education, Labor and Pensions Committee questions Heidi Overton, the nominee for commissioner of the Food and Drug Administration, at 10:30 a.m. Eastern time Thursday. About 4,000 Brigham nurses vote Thursday on an open-ended strike; The Boston Globe's account of the dispute centers on wages and health insurance, not AI.

In other developments, Black Book Research released a survey of 202 hospital nurses on being monitored by AI systems, OpenEvidence is to become free in about 100 low- and middle-income countries on Anthropic's models, Oracle Health announced five revenue-cycle AI tools, and UnitedHealth Group, CVS Health and Kaiser Permanente asked the Centers for Medicare & Medicaid Services (CMS) not to bar vendor-run remote patient monitoring. Colorado's rulemaking page for the state's automated-decision rules showed no revised draft as of Thursday.

In this episode

• OpenAI agent bypasses access controls on Australia's Medicare portal June 18; government's first notice is a Sept. 10 email to a public mailbox
• Epic CEO Judy Faulkner says most development is paused for about six weeks of security work; company says AI roadmap is unchanged
• Survey of 202 hospital nurses: 65% feel watched by AI, 52% changed documentation timing to avoid a flag, 33% less willing to report a near miss
• OpenEvidence, used 42 million times by U.S. clinicians in August, to be free in about 100 low- and middle-income countries on Anthropic's models
• Oracle Health announces five revenue-cycle AI tools, including a pre-claim review of notes for coding gaps; hospital CEO describes 'the bot wars'
• UnitedHealth, CVS and Kaiser urge CMS not to ban vendor-run remote patient monitoring; UnitedHealthcare earlier tried to stop paying for most of it
• UNC and Duke launch $4.4 million network to help North Carolina's rural and critical-access hospitals choose and monitor AI tools

VA signs $775.72 million ambient AI contract; Heidi raises $340 million for agents; robotic thyroid ultrasound tested; AMA presses CMS on prior auth deadline

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

The Department of Veterans Affairs (VA) on Tuesday selected Abridge and Knowtex for a five-year ambient AI enterprise contract with a $775.72 million ceiling across all eligible vendors, Abridge announced and Nextgov reported. The contract is a multiple-award vehicle; the ceiling is a maximum across every vendor eligible to compete for task orders, and money is awarded one task order at a time. The VA began piloting both scribes in October 2025, according to Nextgov. Abridge said its platform is live at more than 75 VA medical centers on both of the department's record systems, in primary care and about a dozen specialties. The VA has published no results from its pilot.

The same day, Heidi, a scribe company, raised $340 million: $100 million of equity at a $900 million valuation and a $240 million facility from General Catalyst. The company's pitch, reported by HIT Consultant, is that the scribe was a first step and that the next is agents that prepopulate order sets, draft referrals and assemble the chart before the visit, with a clinician approving before any action is taken. In Boston, Jane Moran, Mass General Brigham's chief information and digital officer, told the World Medical Innovation Forum that the system is "not allowing autonomous AI across different platforms." The Massachusetts Nurses Association rallied outside the forum's hotel, and about 4,000 Brigham nurses vote Thursday on authorizing an open-ended strike.

In other developments, a Shanghai group reported in npj Digital Medicine a prospective multicenter trial of a fully autonomous robotic thyroid ultrasound that avoided about three-quarters of on-site exams and missed about a fifth of biopsy recommendations; Jefferson Health said its ambient AI rollout saved 1 million clinician hours in its first year and set a target of 10 million by 2028; the American Medical Association (AMA) asked the Centers for Medicare & Medicaid Services (CMS) to hold payers to the Jan. 1, 2027, electronic prior authorization deadline; and the Food and Drug Administration (FDA) updated its list of AI-enabled devices Tuesday to say "over 1,600," up from the 1,500 cited in this site's long read. The Senate Health, Education, Labor and Pensions (HELP) Committee hears from Dr. Heidi Overton, the nominee for FDA commissioner, on Thursday at 10:30 a.m. Eastern time. Colorado's attorney general faced a Wednesday deadline for a revised draft of the state's automated-decision and chatbot rules, and the rulemaking page had not posted one as of Wednesday morning. California's governor has seven days left to act on four health AI bills.

In this episode

• VA selects Abridge and Knowtex for five-year ambient AI contract with $775.72 million ceiling; Abridge live at more than 75 medical centers
• Heidi raises $340 million at $900 million valuation to expand its scribe into agents that prepopulate order sets and draft referrals
• Mass General Brigham does not allow autonomous AI across platforms, digital chief says, as 4,000 Brigham nurses vote Thursday on open-ended strike
• Autonomous robotic thyroid ultrasound avoids 75% of on-site exams and misses 19% of biopsy recommendations in 262-patient, three-center trial
• Jefferson Health says ambient AI saved 1 million clinician hours in first year at 15% to 20% physician adoption; target is 10 million by 2028
• AMA asks CMS to hold Jan. 1, 2027, electronic prior authorization deadline, reject "broad enforcement discretion"; 40,000 fee schedule comments split
• Senate HELP Committee hears FDA nominee Heidi Overton on Thursday at 10:30 a.m. as agency's AI-enabled device list passes 1,600

AI flags esophageal cancer on plain CT; Rush hands chart abstraction to a vendor; AI firms sponsor MAHA Summit; Newsom vetoes prior authorization bill

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

A Chinese research group reported Tuesday in Nature Medicine that a model reading the esophagus on noncontrast CT outperformed all 17 radiologists in a reader study and was validated on 80,612 patients. The model flags esophageal cancer and high-grade precancer on the plain scan ordered for kidney stones and lung nodules. The validation covered 12 centers in three countries, and the model was then run prospectively on hospital scans and on 10,959 consecutive people in a low-dose lung screening program. The paper appeared one day after Alibaba released an abdominal CT model free on GitHub.

The Food and Drug Administration (FDA) press page lists nothing newer than a gene therapy approval on Thursday, and the Centers for Medicare & Medicaid Services (CMS) on Monday announced preliminary lab fee schedule rates. President Donald Trump on Friday announced an AI Force and a coming AI czar with no detail beyond a promise not to hinder the industry and a statement that the existing criminal and civil courts can handle bad behavior, Axios and Fortune reported; the announcement so far includes nothing on medicine. The Make America Healthy Again (MAHA) Summit on Sept. 29 is sponsored by OpenAI, Anthropic, Hims & Hers, Noom, Sword Health, Tempus, Prenuvo and Centene, among others, and sponsors receive dinner with the speakers, according to the summit's site and Becker's Hospital Review. In Sacramento, Gov. Gavin Newsom on Sunday vetoed a prior authorization bill that had passed 64-4 and 31-0 and has not acted on four health-AI bills, with eight days left to do so.

Elsewhere, a survey of 13 countries found that 49% of respondents believe a person with no medical training but skilled with AI could do at least one physician task as well as a physician; DexCare bought Mila Health in the second front-door agent acquisition in a week; and Becker's Hospital Review reported on how health systems triage the AI features Epic ships by the dozen, in a piece in which Rush's chief information officer said on the record that a chart-review job went to a model and is not coming back.

In this episode

• EAGLE finds esophageal cancer on noncontrast CT with 90% sensitivity and 98.5% specificity and holds a 42% positive predictive value prospectively
• Rush moves five-person chart-abstraction job and $1 million contract to AI vendor; CIO says the job is not coming back
• OpenAI, Anthropic, Hims & Hers, Noom, Sword, Tempus, Prenuvo and Centene sponsor the Sept. 29 MAHA Summit; sponsors get dinner with the speakers
• Across 13 countries, 49% say an AI-skilled layperson could do at least one physician task as well as a physician; 59% of adults under 35 agree
• DexCare buys Mila Health, whose agents call, text and chat with patients to book and prepare visits, in the second front-door agent deal in a week
• Newsom vetoes AB 539, a one-year prior authorization bill that passed 64-4 and 31-0, and has not acted on four health-AI bills with eight days left