Flu activity climbs on West Coast; Boston physician sentenced in telemedicine fraud; AI chatbot interviews primary care patients — Morning Medical Update

Flu activity climbs on the West Coast

H1N1 accounts for nearly every subtyped virus as last season closes at 198 pediatric deaths.

Nationally, 3.6% of clinical lab specimens tested positive for influenza in the week ending Oct. 3, and flu hospital admissions rose to 1.0 per 100,000 people, according to the latest FluView report from the U.S. Centers for Disease Control and Prevention (CDC). Positivity reached 9.5% in the region covering California, Arizona, Nevada and Hawaii, and the region covering Washington, Oregon, Idaho and Alaska had the highest admission rate, at 1.4 per 100,000.

Influenza A(H1N1)pdm09 made up 125 of the 127 subtyped influenza A viruses that week. Outpatient visits for influenza-like illness stayed below baseline nationally, at 1.8%. The report also closed out the 2025-26 season with 198 pediatric flu deaths reported to CDC, including three added that week.

Boston physician sentenced in telemedicine fraud schemes

Medicare paid more than $27 million on claims tied to orders she signed for patients she never treated.

Le Thu, 71, of Boston, was sentenced Oct. 6 to 36 months in prison and two years of supervised release for her role in telemedicine health care fraud and kickback schemes, the U.S. Attorney's Office for the District of New Jersey said. She was also ordered to pay more than $27 million in restitution. Thu previously pleaded guilty to conspiracy to violate the Anti-Kickback Statute and to making false statements related to health care matters.

Prosecutors said that from 2014 to 2017, Thu took kickbacks of about $35 per approved prescription from a telemedicine company for durable medical equipment and compounded medications, some billed to TRICARE. From 2017 to 2020, she signed, and had unqualified people sign, thousands of false orders for genetic tests and medical equipment for Medicare beneficiaries she had not treated. Those orders led to more than $40 million in Medicare claims, and Thu was paid more than $200,000 for signing them.

AI chatbot interviews primary care patients before visits

No conversation needed a safety stop, but the study was not built to measure outcomes.

A conversational artificial intelligence (AI) system interviewed patients ahead of urgent primary care visits at Beth Israel Deaconess Medical Center (BIDMC) in Boston, and none of the 98 completed conversations required a safety-stop intervention, according to a feasibility study published in The Lancet. Patients used the text-based Articulate Medical Intelligence Explorer, developed with Google researchers, from home while a board-certified internist monitored every exchange in real time.

Supervising physicians identified one hallucination and added clinical clarification in five cases. In 44 visits where physicians reviewed the AI transcript or summary beforehand, they said it helped them prepare 75% of the time. Patients raised concerns about confidentiality and the chatbot's trustworthiness. The study assessed feasibility, not whether AI improves health outcomes. BIDMC has the news release.