Daily Briefing for October 20, 2026

Show Notes

Doximity and OpenEvidence settle dueling lawsuits over physician AI tools

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

Doximity and OpenEvidence dismissed all claims in their federal lawsuit over physician AI tools, ending a case in which OpenEvidence accused Doximity staff of posing as physicians, Fierce Healthcare reported Oct. 19.

The claims and counterclaims were dismissed with prejudice, and each company bears its own costs and fees; the joint notice disclosed no other terms. Separately, Oxford researchers reported a labor-onset model that pairs clinical risk factors with computerized fetal heart tracing, and a BMJ review found that trials of large language model tools often leave out model versions, code and harms. The day brought few other developments.

In this episode

• Doximity and OpenEvidence dismiss all claims in lawsuit that accused Doximity staff of posing as physicians to probe its AI
• Model pairing clinical risk factors with 20 minutes of computerized fetal heart tracing sorts labor risk across 147,350 births, Lancet Digital Health study finds
• Trials of large language model tools often omit model versions, code and harms reporting, BMJ review of 461 trial records finds
• XRlabs raises $8.4 million for operating-room software that merges live surgical video, navigation and imaging; it has no FDA clearance
• Siemens Healthineers and Paradromics build an imaging workflow for implanting brain-computer interfaces, starting with speech

What is Physician in the Loop?

How technological innovation is changing medicine for patients and the people who practice on the front lines.

Voice 1: This is Physician in the Loop, the daily briefing on AI in medicine. It's Tuesday, October 20th, 2026.

Voice 1: Doximity and OpenEvidence dismissed every claim in their lawsuit over physician AI tools, with each company covering its own costs and fees.

Voice 2: Remind me what OpenEvidence actually accused Doximity of doing.

Voice 1: OpenEvidence sued in June of 2025, claiming Doximity executives and employees posed as physicians, used National Provider Identifier credentials to get onto its platform, and used prompt-injection techniques to try to pull confidential information like system prompts.

Voice 2: Doximity didn't just deny that, it swung back.

Voice 1: It counterclaimed for false advertising, unfair and deceptive practices, and defamation. Now both sides have dropped everything, with prejudice, meaning neither can refile these claims.

Voice 2: Any dollar figure attached to settling all that?

Voice 1: None reported. The joint notice doesn't disclose any other terms beyond each side paying its own way.

Voice 2: Sarah Tishler, a partner at Beck Reed Riden who wasn't involved in the case, said this litigation could have helped define the legal limits around AI competitive intelligence and prompt-injection attacks.

Voice 1: Instead, that question goes unanswered, at least for now.

Voice 2: A different kind of AI question turned up in obstetrics this week.

Voice 1: Oxford researchers built a model called Fit4Labour that pairs seven clinical risk factors, things like maternal age, pre-eclampsia, diabetes, and thick meconium, with features pulled from the first 20 minutes of fetal heart tracing. They validated it across more than 147,000 births in Oxford, Cambridge, and Sweden's Stockholm-Gotland region.

Voice 2: Did it actually outperform how clinicians already screen for this?

Voice 1: Not clearly. At a 15% screen-positive rate, its sensitivity for severe fetal compromise was about 33%, against about 40% for the clinical benchmark. The authors said that difference wasn't statistically significant.

Voice 2: What about telling high risk from low risk within its own results?

Voice 1: That held up better. In the Oxford validation set, severe compromise occurred in about 1% of the highest-risk group, against about 0.1% in the lowest.

Voice 2: Worth remembering this was a retrospective study.

Voice 1: It was. The authors say the model could still produce interpretable output in real time, which is what they're building toward next.

Voice 2: A BMJ review raised a harder question, about whether anyone can actually check these trials at all.

Voice 1: Researchers led by Dr. Danielle Bitterman at Mass General Brigham reviewed 68 trial reports, 15 protocols, and 378 registrations for large language model trials dating back to 2017. Reporting completeness averaged about 40% against one checklist, about 44% against another, and about 59% against a third.

Voice 2: What's actually missing from that reporting?

Voice 1: Model versioning, inference settings, access to protocols, data, and code, and plans for tracking harms were all reported inconsistently. The authors say that limits anyone's ability to judge safety before these tools reach clinical use.

Voice 2: What kind of trials are we even talking about here?

Voice 1: Mostly small, single-site randomized studies published in 2024 and 2025, a lot of them testing patient education or management recommendations.

Voice 2: So the tools are already being tested in people, just without enough of a paper trail to check the work.

Voice 1: Money moved in a different direction too, toward the operating room.

Voice 2: XRlabs raised $8.4 million for operating-room software called SurgicalOS, built to merge live surgical video, navigation, and imaging into one feed. It has no FDA clearance.

Voice 1: What's the evidence it's actually worked anywhere?

Voice 2: The company says it's been used in more than 100 surgical cases across three continents, with live-procedure evaluations run under research governance while standard, cleared systems still guided the actual clinical decisions.

Voice 1: So the approved systems were doing the real work the whole time.

Voice 2: That's how the company frames it. The seed round was led by Playground Global, and former Medtronic chief executive Omar Ishrak is now board chair. XRlabs plans to seek regulatory authorization in the United Kingdom, Europe, and the United States, but none of that has happened yet.

Voice 1: A platform built for live surgery that still needs every one of those approvals.

Voice 2: Before we go, something to watch.

Voice 1: Noridian holds an open meeting on October 22nd, a teleconference on a proposed coverage decision, DL40454, that would stop covering AI brain MRI quantification across 13 Western states and three territories.

Voice 2: Is anyone else weighing the same question?

Voice 1: Yes. Palmetto GBA holds its own open meeting on October 26th, on a matching proposal, DL40464, covering Alabama, Georgia, North Carolina, South Carolina, Tennessee, Virginia, and West Virginia.

Voice 2: Two contractors, two meetings four days apart, both deciding whether to keep paying for the same kind of scan.

Voice 1: Sources and corrections are at physicianintheloop.org, where you can sign up for the free email briefing. This briefing was read by AI voices.