CMS launches 'CRUSH' fraud crackdown; USPSTF misses third straight meeting; Stanford AI model reads CT scans — Morning Medical Update Weekly Recap

Welcome to your Morning Medical Update Weekly Recap for Friday, March 6, 2026.

CMS launches three-pronged health care fraud crackdown

The Trump administration came out swinging on health care fraud last week. At a White House press conference, Vice President J.D. Vance, Health and Human Services (HHS) Secretary Robert F. Kennedy Jr., and Centers for Medicare & Medicaid Services (CMS) Administrator Mehmet Oz, M.D., MBA, announced three simultaneous actions: a six-month nationwide freeze on new Medicare enrollments for seven categories of durable medical equipment (DME) suppliers, a $259.5 million hold on federal Medicaid matching funds to Minnesota over potentially fraudulent claims, and a public request for input on what could become the most significant anti-fraud rulemaking in years.

That last piece is what CMS is formally calling the CRUSH initiative — that’s Comprehensive Regulations to Uncover Suspicious Healthcare. Comments are open through March 30th.

More on CMS’ announcement: ”CMS launches three-pronged plan to CRUSH health care fraud”

USPSTF March meeting ‘postponed for rescheduling’

The U.S. Preventive Services Task Force (USPSTF) was supposed to meet this month. It didn't, and HHS hasn't said why.

The 42-year-old panel, which determines which preventive services insurers are required to cover at no cost, last met in March of 2025. Its July 2025 meeting was abruptly canceled by HHS, November 2025 was lost to a government shutdown, and now March is "postponed for rescheduling."

Concern is growing that HHS Secretary Robert F. Kennedy Jr. could reshape or disband the panel entirely, following his overhaul of the CDC's vaccine advisory committee last June. Reuters has more.

AI-powered CT scans could fast-track diagnoses

Artificial intelligence (AI) may soon help clinicians get more from routine CT scans. Supported by funding from the National Institutes of Health (NIH), researchers developed a machine learning model, called “Merlin,” that can analyze 3D abdominal CT scans and perform a wide range of tasks, from identifying anatomy to predicting disease years before symptoms appear.

The system was trained on more than 15,000 CT scans linked to radiology reports and nearly one million diagnostic codes from Stanford University School of Medicine. In testing across more than 50,000 scans from multiple hospitals, Merlin predicted diagnostic codes correctly more than 81% of the time and identified patients at higher risk of developing chronic diseases within five years about 75% of the time. Researchers say tools like Merlin could eventually speed clinical assessments and uncover disease markers that might be missed by the human eye, potentially allowing clinicians to move more quickly from imaging to diagnosis. The research is published in Nature.