This research quantified physiologically acceptable PICU discharge vital signs and developed machine learning models to predict these values for individual patients throughout their PICU episode. EMR data from 7256 survivor PICU episodes collected between 2009-2017 at Children’s Hospital Los Angeles was analyzed. Each patient’s heart rate, systolic blood pressure, and diastolic blood pressure between medical clearance and physical discharge was predicted using recurrent neural networks. RNN predictions better approximate patient-specific PASS values than age-normal values.
Describe the new knowledge and additional skills the participant will gain after attending your presentation.: Attendees will gain an understanding of using EMR data to train deep learning models to predict clinically relevant information.
Cameron Carlin (Presenter)
Children's Hospital Los Angeles
Long Ho, Children's Hospital Los Angeles
David Ledbetter, Children's Hospital Los Angeles
Melissa Aczon, Children's Hospital Los Angeles
Randall Wetzel, Children's Hospital Los Angeles