Medical reference company MDCalc launched CrossCover by MDCalc on Wednesday, Nov. 11, giving clinicians an independent safety and privacy layer that sits between them and the artificial intelligence chatbots they use at work. The tool automatically strips identifying patient and provider details out of doctors’ prompts before sending them to an AI model, then checks the returned answers for medical evidence and potential accuracy issues.
Physicians using generative AI to look up clinical questions often risk exposing protected health details or relying on unverified chatbot outputs. CrossCover acts as a filter in both directions. When a clinician types a question, the platform removes personal identifiers to reduce privacy risks, according to STAT’s Wednesday report. When the AI model replies, CrossCover flags underlying assumptions, highlights areas of uncertainty, surfaces supporting medical sources, and connects relevant MDCalc clinical calculators.
“We can take away some of the challenges and risks and problems with the technology so that physicians can feel more confident in using it,” Dr. Graham Walker, MDCalc’s co-founder and Head of Artificial Intelligence, told STAT, as cited by Superpower Daily.
How the CrossCover workflow and partnerships work
Clinical decision-support platforms AMBOSS and Vera Health joined CrossCover as initial partners at launch, according to MDCalc’s announcement on Business Wire. Rather than locking doctors into a single proprietary model, CrossCover is structured around a three-step workflow listed on its product page: clinicians first “Use the clinical AI you choose” by asking a question through partner models, then “Review the answer and its evidence” as CrossCover checks the sources and assumptions, and finally make the clinical call themselves using their own medical judgment.
In its Wednesday release, Walker said the company is not building CrossCover to get physicians to use less AI, adding that MDCalc expects doctors to use much more of it. Fellow MDCalc co-founder Dr. Joseph Habboushe said the name CrossCover refers to a hospital practice that signals trust, shared responsibility, and continuity of care among physicians, which is how the company hopes clinicians will view the new platform.
Pushback against clinical AI data collection
MDCalc’s decision to strip information out of AI prompts runs counter to how major health-tech competitors are building their tools. As STAT health tech reporter Brittany Trang noted on Wednesday, competing clinical AI companies are actively expanding the patient and administrative data they collect.
Electronic health record vendor Epic, traditionally a system of record, now wants to display clinical recommendations directly to doctors. Abridge, which started as an ambient medical scribe for recording patient visits, now asks for patient insurance information. Meanwhile, clinical question-answering platform OpenEvidence wants doctors to enter patient records and visit details so its models can generate personalized recommendations.
Walker’s return to MDCalc and early access testing
As background, emergency physicians Walker and Habboushe co-founded New York-based MDCalc in 2005. The free clinical decision-support platform hosts more than 800 clinical calculators and tools and is used every month by more than 80% of U.S. physicians.
Walker practiced emergency medicine for 18 years, including at Kaiser Permanente San Francisco Medical Center, and co-founded the physician platform Offcall in 2024 before returning to MDCalc on Sept. 14, 2026, in the newly created role of Head of Artificial Intelligence. At the time of his September appointment, Walker said that while AI helps doctors spend more time on patients, it also raises ethical and practical concerns around bias, de-skilling, liability, and clinical risk.
MDCalc is currently recruiting doctors into a founding group of “Charter Physicians” to test CrossCover in early access and weigh in on product and safety decisions before a wider rollout. That group is led by Founding Charter Physicians Graham Walker, MD, Joe Habboushe, MD, and Shazia Siddique, MD. Broader industry debate over clinical AI and data access will continue next week at the HLTH conference, where STAT will interview executives from Epic, Novartis, and Johnson & Johnson.