Case Study: Emory University predicts sepsis earlier with Google for Education

Emory University researchers use Google Cloud Platform to predict sepsis in intensive care patients

Emory University’s Department of Biomedical Informatics needed a faster, more reliable way to detect sepsis in ICU patients, a condition that is difficult to diagnose early but can be fatal and costly if missed. Working with Google for Education, the team used Google Cloud Platform, TensorFlow, and App Engine to build an AI-driven prediction engine from anonymized electronic health records and real-time patient data.

Google for Education helped Emory University deploy a scalable, secure sepsis-risk system that analyzes 65 variables and updates every five minutes on a clinician dashboard. The model achieved 85% accuracy in predicting sepsis four to six hours before onset, giving doctors more time to intervene with treatment and potentially improve outcomes while reducing medical costs.

Emory University

Shamim Nemati

Assistant Professor, Department of Biomedical Informatics