๐ Public releases ๐๐ต๐ฟ๐ฒ๐ฒ ๐ฑ๐ฒ-๐ถ๐ฑ๐ฒ๐ป๐๐ถ๐ณ๐ถ๐ฒ๐ฑ ๐น๐ผ๐ป๐ด๐ถ๐๐๐ฑ๐ถ๐ป๐ฎ๐น ๐๐๐ฅ ๐ฑ๐ฎ๐๐ฎ๐๐ฒ๐๐ (EHRSHOT, INSPECT, MedAlign) with 25,991 patients & 295M clinical events (1997-2023)โnow freely available for ๐ฬฒ๐ฬฒ๐ฬฒโฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒโฬฒ๐ฐฬฒ๐ธฬฒโฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ๐ฬฒ in healthcare!
These datasets support ๐บ๐๐น๐๐ถ๐บ๐ผ๐ฑ๐ฎ๐น ๐น๐ฒ๐ฎ๐ฟ๐ป๐ถ๐ป๐ด, ๐๐ถ๐บ๐ฒ-๐๐ผ-๐ฒ๐๐ฒ๐ป๐ ๐บ๐ผ๐ฑ๐ฒ๐น๐ถ๐ป๐ด, and ๐ณ๐ฒ๐-๐๐ต๐ผ๐ ๐๐ฎ๐๐ธ๐ to advance responsible AI in chronic disease and cancer care.
These datasets support ๐บ๐๐น๐๐ถ๐บ๐ผ๐ฑ๐ฎ๐น ๐น๐ฒ๐ฎ๐ฟ๐ป๐ถ๐ป๐ด, ๐๐ถ๐บ๐ฒ-๐๐ผ-๐ฒ๐๐ฒ๐ป๐ ๐บ๐ผ๐ฑ๐ฒ๐น๐ถ๐ป๐ด, and ๐ณ๐ฒ๐-๐๐ต๐ผ๐ ๐๐ฎ๐๐ธ๐ to advance responsible AI in chronic disease and cancer care.
๐ We're thrilled to announce the general release of three de-identified, longitudinal EHR datasets from Stanford Medicineโnow freely available for non-commercial research-use worldwide! ๐View quoted post on X
Read our HAI blog post for more details: hai.stanford.edu/news/advancingโฆ
๐๐ฎ๐๐ฎ๐๐ฒ๐ ๐ฆ๐๐บ๐บ๐ฎ๐ฟ๐ถ๐ฒ๐
๐ 3 longitudinal EHR datasets
๐ฅ Scale: 25,991 patients, 441,680 visits, and 295M clinical events (median: 4,882 events per patient)
โณ Timeframe: Patient trajectories from 1997 to 2023 (median: 10 years per patient)
๐ฆ๐๐ฎ๐ป๐ฑ๐ฎ๐ฟ๐ฑ๐ถ๐๐ฒ๐ฑ ๐๐ฒ๐ป๐ฐ๐ต๐บ๐ฎ๐ฟ๐ธ ๐ง๐ฎ๐๐ธ๐
๐ฏ Few-shot Learning
๐ค Multimodal Learning & Time-to-Event Modeling
โ Long Context Instruction Following & Temporal Reasoning
Thanks to @MichaelWornow, Ethan Steinberg, @Zepeng_Huo, @HennyJieCC , @BediSuhana42170 , @AlyssaUnell, @drnigam @StanfordMed