AOA-OMED Research Posters 2024-OMED24-POSTERS

Video Transcription

Video Summary

Drew Patel presents a study titled "Predicting 30 Days in Hospital Mortality in Surgical Patients" using a logistic regression model with comprehensive preoperative data. Conducted with collaborations from several universities, the research uses Seoul National University Hospital's INSPIRE dataset of 130,000 surgeries. The model outperformed traditional methods in predicting mortality within 30 days post-surgery, achieving an ROC-AUC score of 0.978. Key predictors include emergency surgery status and preoperative health indicators. The model aims to enhance surgical planning and patient care, requiring further validation across diverse settings for integration into clinical practice.

Keywords

hospital mortality prediction

logistic regression model

INSPIRE dataset

surgical planning

ROC-AUC score