Data-driven automated classification algorithms for acute health conditions: Applying PheNorm to Anaphylaxis

Poster Number: P187
Presentation Time: 05:00 PM - 06:30 PM

Abstract Keywords: Natural Language Processing, Real-World Evidence Generation, Machine Learning, Population Health
Primary Track: Foundations
Programmatic Theme: Public Health Informatics

Accurate identification of anaphylaxis using observational data is important for medical product safety surveillance, but difficult to both diagnose clinically and recognize algorithmically. Traditional phenotyping methods rely on expensive gold standard training data and manual feature engineering. Herein we apply an automated approach, PheNorm, to create a computable phenotype for identifying patients with anaphylaxis using NLP, machine learning, and low-cost silver-standard training labels. Performance was comparable to a recently published, higher-cost manual phenotyping effort.

Speaker(s):
Joshua Smith, PhD
Vanderbilt University Medical Center

Author(s):
Joshua Smith, PhD - Vanderbilt University Medical Center; Daniel Park, BS - Vanderbilt University Medical Center; Jill Whitaker, MSN - Vanderbilt University Medical Center; Michael McLemore, BSN - Vanderbilt University Medical Center; Robert Winter, BA - Vanderbilt University Medical Center; Arvind Ramaprasan, MS - Kaiser Permanente Washington Health Research Institute; David Cronkite, MS - Kaiser Permanente Washington Health Research Institute; Saranrat Wittayanukorn, PhD - US Food and Drug Administration; Danijela Stojanovic, PharmD, PhD - US Food and Drug Administration; Yueqin Zhao, PhD - US Food and Drug Administration; Sarah Dutcher, PhD - US Food and Drug Administration; Kevin Johnson, MD, MS - University of Pennsylvania; David Carrell, PhD - Kaiser Permanente Washington Health Research Institute; Brian Williamson, PhD - Kaiser Permanente Washington Health Research Institute;