ATHENA's Adverse Event Hypotheses Validated Across 5.4M Patient EHRs
marinkazitnik · x · 2026-07-07
A research team conducted a large-scale validation of adverse event association hypotheses generated by ATHENA using longitudinal electronic health records from 5.4 million patients under Israel's Clalit Health. The adjusted odds ratios for predicted associations ranged from 1.48 to 1.84, while the negative control group remained near zero. This robustly demonstrates the effectiveness of ATHENA's predictions in population-scale real-world data, serving as a crucial supplement to previous expert blind evaluations.
More from Research
- Infini-AI Lab open-sources Hermit, a framework for agent harness evolution and continual learning — BeidiChen · 2026-10-27
- Meta FAIR's Self-Direction turns agent failures into reusable LoRA skills, +13.8pp on Terminal-Bench — karen_ullrich · 2026-10-27
- DONIC hits 86.79% on bimanual dexterous tracking, crushing prior SOTA's 9.43% — liuziwei7 · 2026-10-27
- Why don't more LLMs ship 10M-token context? A researcher asks about attention tradeoffs — agarwl_ · 2026-10-27
- PSNR below 15 dB in papers likely means the metric was misused, researcher warns — kwangmoo_yi · 2026-10-27
- Ortet's first health eval: 34 open-weight models, 50 tasks, generic rankings fall short — thegautamkamath · 2026-10-27