Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 435 - Making the Most of Wearable Device Data: Functional Data Approaches
Type: Invited
Date/Time: Thursday, August 6, 2020 : 10:00 AM to 11:50 AM
Sponsor: Biometrics Section
Abstract #309416
Title: Predicting GI Bleeds Using Wearable Device Data
Author(s): Julia Wrobel*
Companies: Colorado School of Public Health
Keywords: functional data analysis; wearable devices; time series forcast; GI bleeds; multiview data

Gastrointestinal (GI) bleeding, or hemorrhage, is a symptom of a disorders in the digestive tract including hemorrhoids, ulcers, colitis, and Crohn's disease. GI bleeds can be severe and even life threatening, causing a significant blood loss over a short time. The current standard of diagnosis is the alarming appearance of blood in the stool or vomit. Our goal is to use a multistream wearable device to predict when high-risk subjects are having or are likely to have a bleed, thus recognizing the problem earlier and reducing blood loss. The study uses data from a new wearable device called the E4 wristband, which simultaneously tracks heart rate, physical activity, skin temperature, and skin conductance over 48 hours. The E4 data streams, along with electronic health record data collected during a hospital stay, will be used to identify data signatures that are predictive of GI bleeds.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program