Monthly update - March 2025

Top Picks of the Month

A group of German researchers, demonstrate that LLMs can achieve highly accurate de-identification of clinical notes. This result holds significant potential for accelerating medical research by reducing the time and effort typically spent on the de-identification process.

Deidentifying Medical Documents with Local, Privacy-Preserving Large Language Models: The LLM-Anonymizer (Paper, Code)

​Another group of researchers from Germany investigated the use of LLMs as encoders for Electronic Health Records (EHRs), and compared their performance against other emebedding methods (BERT models) and the latest EHR foundation models (CLIMBR-T-Base). They found that LLM-based embeddings often matched or outperformed the other methods.

Large Language Models are Powerful EHR Encoders (Pre-print)

---

Other Notable Mentions

• Blood-based biomarkers of Alzheimer’s disease and incident dementia in the community (Paper)

• Comparison of Frontier Open-Source and Proprietary Large Language Models for Complex Diagnoses (Research letter)

• Demographic bias of expert-level vision-language foundation models in medical imaging (Paper)

• Empowering Medical Multi-Agents with Clinical Consultation Flow for Dynamic Diagnosis (Pre-print)

• Family-based genome-wide association study designs for increased power and robustness (Paper)

• Leveraging large-scale biobank EHRs to enhance pharmacogenetics of cardiometabolic disease medications (Paper)

• Making large language models into reliable physician assistants (Opinion paper)

• Medical Large Language Model Benchmarks Should Prioritize Construct Validity (Pre-print)

• Meta-analysis of genome-wide associations and polygenic risk prediction for atrial fibrillation in more than 180,000 cases (Paper)

• Multimodal AI Predicts Clinical Outcomes of Drug Combinations from Preclinical Data (Paper)

• Multimodal generative AI for medical image interpretation (Opinion paper)

• Neuroprotective mechanisms of exercise and the importance of fitness for healthy brain ageing (Paper)

• Optimal dietary patterns for healthy aging (Paper)

• Performance of Large Language Models on the Internal Medicine Mock Exam (Research letter)

• Transformers and genome language models (Review paper)

• LLMs in medicine. evaluations, advances, and the future (Blog post)

Final Notes

This month I scanned a total of 18 publications spanning 8 different countries.

My top picks are based on my topics of interest—definitely a bit biased! If you came across any cool papers published in March 2025 (related to health data science topics) that you think deserved more recognition, please send them my way and I’ll try to include them in next month’s update!