Drastic Changes Coming to HHS with 10,000 Positions Eliminated
The Department of Health and Human Services (HHS) announced a sweeping reorganization Thursday, March 27, 2025, that will cut 10,000 federal jobs, consolidate nearly half of the agency’s divisions, and introduce a renewed focus on chronic disease prevention. The move, which HHS Secretary Robert F. Kennedy, Jr. described as a fundamental realignment of the agency’s mission, is projected to save taxpayers $1.8 billion annually.
The restructuring is being carried out under President Trump’s Executive Order on government efficiency and marks one of the most significant federal health shakeups in decades.
At the heart of the plan is the creation of the Administration for a Healthy America, or AHA, a new agency that will merge several existing offices, including the Health Resources and Services Administration, the Substance Abuse and Mental Health Services Administration, and the Office of the Assistant Secretary for Health.
Kennedy framed the reorganization as a shift from treating disease to preventing it, placing new emphasis on access to clean water, nutritious food and the elimination of environmental toxins.
“We aren’t just reducing bureaucratic sprawl,” he said in a statement. “We are realigning the organization with its core mission and our new priorities in reversing the chronic disease epidemic.”
Google Launches New AI Model for Drug Discovery
Google DeepMind released a new open-source AI tool this week aimed at speeding up the development of new therapies and reducing the costs of bringing drugs to market.
The tool, called TxGemma, is trained to understand how potential medicines behave—predicting things like toxicity, effectiveness and how a drug might perform in a clinical trial.
TxGemma can analyze therapeutic data, explain its reasoning and even generate predictions across different stages of drug research. Unlike traditional AI tools that are trained for one narrow task, the AI model can tackle many at once and be customized with a company’s own data.
Researchers can use the model in two ways: one version offers straight predictions, while another supports more conversational, exploratory analysis. Google is also offering a companion tool called Agentic-Tx, which combines TxGemma with online research tools to solve more complex scientific problems.
The models are available for free on platforms including Hugging Face and Google’s Vertex AI, with guides to help researchers adapt them to their own projects.
Company Raises $25 Million to Build Diverse Genomic Biobank
Galatea Bio has raised $25 million to build a biobank aimed at sequencing the genomes of 10 million people, with a focus on individuals of non-European ancestry – groups that are typically underrepresented in genetic research.
The funding will support the Miami-based company’s AI-powered tools for predicting disease risk, identifying drug targets and improving clinical trial design. Galatea Bio’s approach blends genetic data with real-world health records to create more accurate models for conditions like heart disease, cancer and neurodegenerative disorders.
At the center of its efforts is the Galatea Global Biobank, which aims to close gaps in existing datasets that often lead to less effective treatments for diverse populations. The company recently launched a new tool, StrataRisk, to improve genetic risk predictions for these groups.
Study Suggests Mammograms Can Also Identify Early Heart Disease
A new study being presented at the American College of Cardiology conference this weekend will show that mammograms, typically used for breast cancer screenings, could also help detect early signs of heart disease, with the help of artificial intelligence.
Researchers from Emory University and Mayo Clinic trained an AI model to analyze calcium buildup in breast arteries, a marker linked to cardiovascular risk. Using data from more than 56,000 mammograms, the model accurately predicted the likelihood of heart attacks, strokes or death within five years, especially in women under 60.
While mammograms already capture this calcification, it’s rarely reported or quantified in clinical practice. The AI tool could change that by automatically calculating a cardiovascular risk score from routine imaging.
While the technology is not yet FDA approved, researchers say it could give clinicians a new way to identify heart disease risk early, particularly in younger women who are often underdiagnosed and may not receive heart disease screenings.