Infectious Diseases Grand Rounds

TOPIC: "Generating Data from Free Text in EHR (with Autism Spectrum Disorders as an example)"
SPEAKER: Gondy Leroy, PhD

WHERE: Join Zoom Meeting - https://uahs.zoom.us/j/2837376590

WHEN: Tuesday, April 7, 2020 | 8:00 a.m. - 9:00 a.m.

About the Speaker
Gondy Leroy, PhD, is Professor in MIS at the University of Arizona’s Management Information Systems. Her research focuses on the design, development and evaluation of artifacts that support, facilitate and improve information exchange between people. She has worked on apps to facilitate communication with children with autism, search engines for biomedical information, interview systems for crime witnesses and recently she has focused on automating text simplification in healthcare and text mining of electronic health records (ERH). She has won grants from NIH, AHRQ, NSF, Microsoft Research and several foundations, totaling more than $2.4M as principal investigator. She earned a combined BS and MS (1996) in cognitive, experimental psychology from the Catholic University of Leuven, (1996) and a MIS (1999) and PhD (2003) in management information systems from the University of Arizona. She serves on the editorial board of the Journal of Database Management, International Journal of Social and Organizational Dynamics in IT, Health Systems, Journal of Business Analytics, and co-chairs several sessions, tracks, workshops, and conferences focusing on design science and healthcare IT. She is the author of the book “Designing User Studies in Informatics (Springer, 2011). Finally, she is an active contributor to increasing the diversity and inclusion in computing and founded and leads the “Tomorrow’s Leaders Equipped for Diversity” program at the University of Arizona’s Eller School of Management.

A Summary from Dr. Leroy
In this presentation, I will provide a high level overview of natural language processing (NLP) techniques that can help generate data from free text in EHR. I will show the approach and results from two projects that focus on autism spectrum disorders (ASD). The first is an entity extraction project where we automatically identify individual diagnostic criteria in the EHR free text that match the Diagnostic Statistical Manual of Mental Disorders (DSM). The second is a classification project where we assign case labels to EHR. The projects combine both rule-based approaches and several machine learning algorithms. I will discuss current results and problems encountered because this is a low resource area. I will also show examples of new analyses made possible with this type of data creation and biases to be taken into account. I hope to conclude with a discussion with the audience of new promising areas and potential extensions, applications, and collaborations.