SIG | Data Science in Epilepsy: Natural Language Processing and EMR Phenotyping for Precision Epilepsy

Sunday, December 5, 2021
-

You may need to log in to view video on this page.

OVERVIEW:

This Special Interest Group examines the benefits of natural language processing in electronic health records for precision epilepsy treatment and care.

A wealth of information is contained in electronic health records (EHRs) that can be harnessed to improve diagnostics, antiseizure medicine (ASM) or device discovery, monitoring of ASM safety and efficacy, and efficiency of clinical trial design for people with epilepsy. However, EHRs are difficult to model due to noise, sparseness, incompleteness, high dimensionality, and biases. A large amount of information is hidden in free-text clinical notes, which remains the most common way of documenting events in outpatient/EMU notes and surgical case conferences.

In this session, speakers investigate:

  1. How natural language processing (NLP) can be used to advance precision epilepsy efforts through EHR phenotyping, including a basic NLP primer
  2. State-of-the-art NLP applications for advancing clinical care and drug/device discovery in epilepsy
  3. Applications of NLP in epilepsy, including phenotype detection, improvement of efficiency in clinical trials, pharmacovigilance, Sudden Unexpected Death in Epilepsy (SUDEP) risk evaluation, and prediction of high-risk clinical events

Learning Objectives:

Following participation in this activity, participants will be able to:

  • Describe basic methods in natural language processing (NLP) analysis
  • Explain NLP's role in EMR phenotyping to advance precision epilepsy, including phenotype detection, efficiency of clinical trial design, pharmacovigiliance, and identification of high-risk patients
  • Discuss challenges in applying NLP to EMR phenotyping and methods for addressing these challenges

Program:

SIG Coordinators: Daniel M. Goldenholz, MD, PhD, Gregory Worrell, MD, PhD, and Sharon Chiang, MD, PhD 

Chair: Daniel M. Goldenholz, MD, PhD  

Using Natural Language Processing to Improve Efficiency of Clinical Trials and Quality Measurement | Xiaoxi Yao, PhD, MPH, MS, FACC, FAHA 

Identifying High-Risk Patients Using Natural Language Processing Based on Unstructured Clinical Notes | Wendong Ge, PhD 

Epilepsy Needles in EHR Haystacks — Computable Phenotypes, Natural Language Processing, and Learning Healthcare Systems | Zachary Grinspan, MD, MS, FAES 

Activity Type
Special Interest Group
Credit
Non-CME
Format
On-demand
Career Stage
Early Career (typically 0-5 years from completion of training)
Mid-Career (typically 6-15 years from completion of training)
Senior (typically >15 years from completion of training)
Audience
Advanced Practice Providers
Behavioral Health Providers
Clinicians
Fellows/Trainees
Nurses
Pharmacists
Scientists/Researchers
Demographic
Clinical
First-time Attendees
Research
Young Professionals