The NIH's AIM-AHEAD initiative has launched training programs designed to equip researchers and clinicians with skills they need to address health care challenges using AI and ML. The NIH is accepting applications for four eight-month AI/ML training programs that include a comprehensive curriculum in data science and hands-on training with NIH data resources, writes Susan Gregurick, the NIH's associate director for data science and director of the Office of Data Science Strategy.

Health system leaders are integrating artificial intelligence into existing strategic initiatives, focusing on tangible outcomes like reducing documentation burdens and improving access, according to interviews with AI and informatics leaders at multiple health systems. "AI should be embedded into whatever your strategic projects are or whatever problems are being solved at the health system level," said Deepti Pandita, chief medical information officer and vice president of clinical informatics at UCI Health.

The Mayo Clinic has reduced unnecessary thyroid function tests by 15% and concurrent orders by 19% after implementing a guideline-based cascade approach, starting with thyroid-stimulating hormone tests. Since researchers added a cascade algorithm and prompts to the EHR system, the clinic has saved nearly $50,000 a year, researchers reported at a meeting of the College of American Pathologists.

A decision-support tool predicted two-year hip fracture risk with higher sensitivity but lower specificity than the standard fracture liaison service approach, according to a study presented at the American Society for Bone and Mineral Research's annual meeting. The tool, which analyzed data from more than 3.5 million patient records in Sweden, revealed that number of prescriptions, alcohol-related disorders and the use of anti-dementia drugs were among the risk factors.

An AI algorithm predicted keratoconus progression with greater than 90% accuracy after a second visit, according to a presentation at a meeting of the European Society of Cataract and Refractive Surgeons. The model uses clinical and imaging data to stratify patients into high- and low-risk categories, potentially improving treatment and monitoring.

A deep learning toolkit that integrates multi-omics and medical imaging data may improve precision oncology, researchers reported in Nature Communications. Flexynesis analyzes diverse data types simultaneously to identify biomarkers and potentially effective treatments and is available through various platforms such as PyPI and Docker.

The University of South Florida's College of Public Health is offering a new curriculum in health care simulation operations. Students are trained to use simulation technology using lifelike dummies and computer programs.

The US has made the least progress among high-income Western countries in reducing deaths from chronic diseases, according to a study in The Lancet. From 2010 to 2019, chronic disease mortality declined in most age groups in the US but rose for adults ages 20 to 45. Between 2010 and 2019, chronic disease deaths decreased in some 80% of countries worldwide.

The Trump administration is planning to award a no-bid contract to Rensselaer Polytechnic Institute to research a potential link between vaccines and autism, according to a government procurement notice. The decision, based on RPI's "unique ability" to connect data on children and mothers, has raised questions about the company's access to data and the necessity of the study, as extensive research over two decades has found no credible association between vaccines and autism.

Data in the CDC's Morbidity and Mortality Weekly Report found that influenza hospitalizations in the US reached record levels during the 2024-25 season, with rates 1.8 times to 2.8 times greater when compared with median cumulative rates since the 2010 to 2011 season. Hospitalization rates were lowest for children and teens ages 5 to 17 and highest for adults 75 and older.