Advanced Detection of Nausea/Vomiting and Anxiety in Patients with Cancer

Poster Number: P82
Presentation Time: 05:00 PM - 06:30 PM

Abstract Keywords: Natural Language Processing, Deep Learning, Data Mining, Clinical Decision Support, Bioinformatics, Machine Learning, Informatics Implementation, Nursing Informatics
Working Group: Knowledge Discovery and Data Mining Working Group
Primary Track: Applications
Programmatic Theme: Clinical Research Informatics

This study leverages large language models (LLMs) to enhance the detection of nausea/vomiting and anxiety in cancer patients. By augmenting the Bio-Clinical BERT model with extensive clinical data and symptom-specific tuning, it introduces an improved method for identifying symptoms from clinical texts. Comparative evaluations reveal its enhanced efficacy over other models, particularly in recognizing physical symptoms. This underscores the utility of LLMs in detecting symptoms and elevating patient care in clinical settings.

Speaker(s):
Nahid Zeinali, PhD student
University of iowa

Author(s):
Alaa Albashayreh, PhD, MSHI, RN - University of Iowa; Weiguo Fan; Stephanie Gilbertson White, PhD, APRN-BC, FAAN - university of Iowa;