Global Digi-Health Monitor: Calendar Week 9, 2025
Study Reveals Risks of Using AI for Health Information
A new Australian study warns about using ChatGPT for medical advice. About 10% of Australians asked AI health questions in early 2024, with higher rates among people with low health literacy or from non-English speaking backgrounds. Most users „somewhat“ trusted the AI’s answers. While AI offers instant, simple health information, it’s dangerous for questions needing clinical judgment. The study found 61% of users asked questions that should be directed to healthcare professionals instead. Researchers recommend building „AI health literacy“ and directing people to proper resources like HealthDirect, which offers nurse helplines and symptom-checking tools.
UF Team Develops AI to Bring Hospital-Level Care to Rural Areas
UF researchers are creating an AI system called Multi-Tags to power mobile healthcare units for rural communities. Led by Dr. Yonghui Wu, the project is part of the national PARADIGM program aimed at delivering hospital-level care where hospitals don’t exist. The system will guide basic lab procedures and provide real-time prompts to help generalist healthcare workers perform specialized tasks. Unlike traditional healthcare delivery, these high-tech vehicles will bring medical services directly to remote areas, reducing costs while improving outcomes. Dr. Wu brings valuable experience from developing GatorTron, a widely used clinical language model with over 1.8 million downloads.
AI Algorithm Decodes Immune System Data to Spot Disease
Stanford researchers created Mal-ID, a machine learning tool that diagnoses diseases by analyzing immune cell patterns. The system examines millions of B and T cell receptor sequences to identify conditions including lupus, diabetes, and COVID-19, outperforming traditional methods. Using techniques similar to those powering ChatGPT, it recognizes disease signatures even without knowing exactly what molecules the immune system targets. Beyond diagnosis, the tool could help categorize complex diseases into treatment-specific subtypes and uncover new therapeutic approaches by revealing hidden patterns in our body’s natural defense system.
How Big Data Is Transforming Pharmaceutical Research
Big data has revolutionized the pharmaceutical industry by accelerating drug development, optimizing clinical trials, enabling personalized medicine, and improving drug safety monitoring. Researchers now analyze vast genomic datasets to develop targeted therapies like CAR-T cell treatments and use computational models to simulate drug interactions before expensive lab testing. During clinical trials, big data helps select the ideal participants, monitor results in real-time, and sometimes even replace control groups with virtual ones. For patient care, it enables truly personalized treatment plans by combining genetic profiles with medical records and wearable device data. Despite these benefits, pharmaceutical companies face significant challenges including data standardization, ensuring accuracy, organizational silos, regulatory compliance, and finding skilled talent to manage these complex systems.
Teaching Electronic Health Records to Spot Hidden Liver Disease
Columbia University’s Dr. Julia Wattacheril has created software that finds undiagnosed fatty liver disease by sifting through electronic health records. Testing at Columbia identified 16,000 potential cases—two-thirds with no prior diagnosis in their records. The system mimics a liver specialist’s thinking, examining lab results, clinical notes, and radiology reports to spot patients with MASLD (metabolic dysfunction-associated steatotic liver disease), which affects up to 40% of Americans but often goes unrecognized until serious damage occurs. With new treatments available, including weight loss drugs and medications that reduce scarring, early detection has become more crucial. The team will soon validate the software by contacting doctors whose patients were flagged and offering specialized testing.