Abstract: Patients undergoing cancer therapy in the form of radiation therapy or systemic therapy may require evaluation in the emergency department or hospitalization due to treatment toxicity or symptomatic disease. Identification of patients likely to experience these toxicities could direct aggressive supportive care to decrease such events. The objective of this study is to apply machine learning to structured and unstructured electronic medical record data to generate actionable predictions to mitigate patient toxicities.

Learning Objective 1: Routinely collected EMR structured data prior to cancer therapy can provide predictive insights into rates of emergency department visits and hospitalization during treatment.


Julian Hong (Presenter)
Duke University

Manisha Palta, Duke University
Jessica Tenenbaum, Duke University

Presentation Materials: