Predictive Analytics using the PEDSnet Common Data Model

Brown Bag Lunch – Please RSVP.

For the past year we have been developing Machine Learning (ML) predictive analytic tools to characterize disease course and identify adverse outcomes for the Nemours patient population. Central to that effort is a database containing research-ready EHR data on all Nemours patients for whom health data exists in machine readable form (over two million distinct patients). These data have been extracted from our EHR system, transformed into  the PEDSnet Common Data Model (CDM) and loaded into an internal database available to Nemours investigators for research and quality improvement. This presentation will provide an overview of the PEDSnet CDM, a description of the tools and methods developed for access to, and analysis of these data. Examples of how predictive analytics can inform our precision medicine initiative will be presented. 

Meet the Speaker

Tim Bunnell, PhD

Dr. Bunnell received his Ph.D. in Experimental Psychology from The Pennsylvania State University in 1983, concentrating in human speech perception and the acoustic properties of speech (acoustic phonetics). From 1983 to 1989, he worked as a Research Scientist in the Sensory Communication Research Laboratory (later Center for Auditory and Speech Sciences) at Gallaudet University conducting research on the application of digital speech processing techniques to hearing enhancement, primarily for acoustic hearing aid users. In 1989 he became the director of the Speech Processing Laboratory at the Alfred I. duPont Hospital for Children where his research interests expanded to include text to speech synthesis and speech recognition. Dr. Bunnell is now the director of the Nemours Center for Pediatric Auditory and Speech Sciences (CPASS) and head of the Speech Research Laboratory within the CPASS. Since 2000, he has also served as director of the Nemours Biomedical Research Department’s Bioinformatics Core (now the Biomedical Research Informatics Center — BRIC). BRIC maintains computing infrastructure for Nemours Biomedical Research, and provides expertise in applications development, biostatistics, data mining, and numerical analysis. Dr. Bunnell’s primary research interests are in biomedical and clinical applications of speech technology for the diagnosis and remediation of hearing and speech disorders, particularly in pediatric patients. However, in recent years he has also actively explored the application of machine learning and data mining to the analysis of electronic health records data.

I am interested in leveraging AI to accelerate knowledge discovery through the application of innovative, distributed and scalable approaches that expand our understanding of health and healthcare with a focus on cancer. Specifically, I examine novel hybrid clustering approaches such as combining clusters with network analysis, ontological knowledge, and risk models. In addition, I explore integrated AI learning applied to large repositories of medical data for knowledge discovery.  I use infrastructure-as-code to model the integration of medical multi-omic data beyond institutional boundaries to generate reproducible models across healthcare networks.

This activity has been approved for AMA PRA Category 1 Credit

Stay Connected

Join the ACCEL newsletter and keep up to date with the latest news, events and funding opportunities!