A Semi-Automated System For Detecting Treatment Associated Adverse Events From Clinical Notes In Electronic Health Records Systems

Value Proposition

Adverse events (AE) are harmful and negative outcomes that happen when a patient has been provided with medical care and they frequently occur in any medical system, with at least one in ten patients affected. Medical treatment may include a procedure, surgery, or medication, and AEs may include side effects, injury, psychological harm or trauma, or death. Any patient who undergoes treatment may experience a negative outcome as a result of that treatment.

Technology Description

UCSF investigators have developed a deep learning model for identifying treatment-related adverse evets using electronic health record (EHR) data. The model was adapted to the task of mining for SAE to steroid-sparing immunosuppressants in outpatient clinic notes for patients with IBD.

Related Materials

• Silverman, A. L., Madhumita Sushil, Balu Bhasuran, Ludwig, D., Buchanan, J., Racz, R., Mahalakshmi Parakala, Samer El-Kamary, Ohenewaa Ahima, Belov, A., Choi, L., Billings, M., Li, Y., Habal, N., Liu, Q., Tiwari, J., Butte, A. J., & Rudrapatna, V. A. (2024). Algorithmic Identification of Treatment-Emergent Adverse Events From Clinical Notes Using Large Language Models: A Pilot Study in Inflammatory Bowel Disease. Clinical Pharmacology & Therapeutics, 115(6), 1391–1399.

Patent Status

Patent Pending