Trainee Rounds: Ghadir Ali and David Pellow
DATE: May 28, 2025 (Wed.)
TIME: 12pm to 1pm
PRESENTERS: Ghadir Ali and David Pellow
Learn more about the presenters
Ghadir Ali
Title of Talk
Balancing Automation and Expert Oversight: Evaluating LLM-Assisted Discharge Summaries in Clinical Setting
Description
Clinical documentation is an essential yet time-intensive task for hospitalists, contributing to cognitive burden and burnout. Large Language Models (LLMs) offer a potential solution by summarizing daily patient notes into discharge summaries. However, their integration into clinical workflows raises concerns about accuracy, safety, and usability. This study explores how LLMs can be effectively incorporated while ensuring physician oversight and documentation reliability.
David Pellow
Title of Talk
Development of a Readily Deployable Prognostication Tool for Predicting Major Adverse Cardiovascular Events Following Liver Transplantation
Description
Liver transplant (LT) recipients are at an increased risk of death over a 4-year period due to cardiovascular events compared with age- and sex-matched controls and over a quarter experience a major adverse cardiovascular event (MACE) within 10 years of transplantation. Importantly, risk scores used to assess cardiovascular risk amongst the general population do not perform as well in LT recipients and there currently are no validated tools for predicting the risk of MACE beyond the first post-transplant year. Using retrospective UHN data from approximately 4000 LT recipients, we developed a random survival forest model designed for the prediction of 10-year post-transplant risk of MACE.
