DT-Transformer predicts disease trajectories using large-scale EHR data

Researchers have developed DT-Transformer, a foundation model for predicting disease trajectories using electronic health records. Trained on over 57 million structured EHR entries from 1.7 million patients across Mass General Brigham's network, the model demonstrates strong predictive capabilities. It achieved a median AUC of 0.871 across 896 disease categories in both held-out and prospective validation, indicating the potential of health system-scale data for clinical forecasting. AI

IMPACT Demonstrates the potential of large-scale, real-world health system data for training robust clinical forecasting models.

RANK_REASON The cluster describes a new research paper detailing a novel foundation model for disease prediction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Hugging Face Daily Papers →

• DT-Transformer
• Mass General Brigham

• paper
• model release
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →