Skip to main content

Inclusion of Semi-Structured and Unstructured Electronic Health Record (EHR) Data in Confounding Adjustment and Outcome Ascertainment

Basic Details
Date Posted
Status
In progress
Description

The MOSAIC-NLP (Multi-source Observational Safety study for Advanced Information Classification using NLP) project intends to build and utilize semi-automated methods and tools, including natural language processing (NLP), to enable outcome identification, confounder feature extraction, and longitudinal contextualization of electronic health record (EHR) data. The project will demonstrate how unstructured EHR data could help improve the validity of population-based pharmacoepidemiologic studies within linked health care claims and EHR data environments. Using longitudinal data from the Cerner Enviza Electronic Health Records (CE EHR) linked to claims and leveraging NLP technology, the overall goal is to show the usability and value of structured, semi-structured, and unstructured EHR data to improve the identification of study outcomes and covariates for use in pharmacoepidemiologic studies. This Innovation Center (IC) project is led by a team at Cerner Enviza in collaboration with Harvard Pilgrim Health Care Institute (HPHCI) and Brigham and Women’s Hospital (BWH). 

Deliverable(s) (1)
Use of Natural Language Processing in a Pharmacoepidemiology Study: The Examination of Neuropsychiatric Events and Incident Use of Montelukast Among Patients with Asthma Protocol
Information
Time Period
October 15, 2022 - September 30, 2024
Data Source(s)
Cerner Enviza Electronic Health Records (CE EHR), US claims data, National Jewish Health (NJH) clinical data
Related Links
Related Link(s)
Sentinel Innovation Center
Press Release: Cerner Enviza Collaborates with FDA to Develop Innovative AI Tools for Drug Safety and Real-World Evidence Studies
Presentation: Unexpected Issues in Pharmacoepidemiology Studies Applying Natural Language Processing to Clinical Notes
Presentation: Natural Language Processing in Pharmacoepidemiology: Lessons from the Multi-Source Observational Safety study for Advanced Information Classification Using NLP (MOSAIC-NLP)
Workgroup Leader(s)

Elise Berliner, PhD; Dena Jaffe, PhD, MS; Cerner Enviza, an Oracle company, North Kansas City, MO

Workgroup Member(s)

Darren Toh, ScD; Jenna Wong, PhD, MSc; Kim Gegear, MSc; Meighan Rogers Driscoll, MPH; Department of Population Medicine, Harvard Pilgrim Health Care Institute and Harvard Medical School, Boston, MA

Richie Wyss, PhD, MSc; Josh Lin, MD, MPH, ScD; Mass General Brigham, Boston, MA

Jenny Cai, MA, MBA; Sonam Lama, MPH; Nathan Vavroch, BA; Mike Jones, MBA; Bridget Balkaran, MPH; Austin Yue, MS; Cerner Enviza, and Oracle company, North Kansas City, MO

Hoang “Wayne” Nguyen, MD; Michael Chu, MD; Heather Huszti, PhD; Louis Ehwerhemuepha, PhD; Olga Guijon, MD; Children’s Hospital of Orange County, Orange, CA

Hasham Ul Haq, BS; Veysel Kocaman, PhD, MS; Andre Marian Feier, MD; Christian Kasim Loan, BS; Jiri Dobes, PhD; Enes Yamul, MS; David Talby, PhD, MBA, MS; John Snow Labs, Lewes, DE

Michael Wechsler, MD, MMSc; David Beuther, MD, PhD; National Jewish Health; Denver, CO

David Carrell, PhD; Kaiser Permanente Washington Health Research Institute, Seattle, WA

Sarah Dutcher, PhD, MS; Jummai Apata, MBBS, DrPH; Sai Dharmarajan, PhD; Robert Lim, MD; Jie Li, MBBS, PhD; FDA, Silver Spring, MD