Skip to main content

Improving Probabilistic Phenotyping of Incident Outcomes through Enhanced Ascertainment with Natural Language Processing

    Basic Details
    Date Posted
    In progress
    Health Outcome(s)

    Using a representative use case of suicidality (i.e., suicide ideation and suicide attempts), this project will advance several themes related to Sentinel’s computable phenotyping strategy for electronic health record (EHR) data, including focusing on phenotyping methods for identifying incident (versus prevalent) conditions, focusing on a health outcome of interest that relies predominantly on unstructured EHR data and for which a clearly defined reference standard is absent, the need for rapid (near-real-time) natural language processing (NLP), and the evaluation of the generalizability of EHR-based phenotyping to neuropsychiatric events. The specific aims of the Workgroup are to:


    • Aim 1: Adapt NLP to identify lifetime suicidality to classify suicidality at the level of an individual clinical event to identify incident suicidality. 
    • Aim 2: Validate the suicide risk models to identify suicide as cause of death by obtaining and linking NDI records for deaths that are identified in our EHR from 2015-2020 but for which direct cause of death information is not available in the EHR.
    • Aim 3: Evaluate the generalizability of this approach by characterizing its applicability to neuropsychiatric events as a second outcome.
    Time Period
    2015 - 2022
    Population / Cohort
    Academic medical center adult patient population
    Data Source(s)
    Electronic Health Records
    Workgroup Leader(s)

    Colin G. Walsh, MD, MA, Department of Biomedical Informatics, Vanderbilt University Medical Center, Nashville, TN

    Workgroup Member(s)

    Sai Dharmarajan, PhD, MS; Andrew Mosholder, MD, MPH; Danijela Stojanovic, PharmD, PhD; Office of Surveillance and Epidemiology, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, MD

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

    Adi Bejan, PhD; Cindy Chen, PhD; Daniel Fabbri, PhD; Kevin Johnson, MD, MS, FAAP, FAMIA, FACMI; Michael Matheny, MD, MS, MPH; Michael Ripperger; Katelyn Robinson; Drew Wilimitis; Department of Biomedical Informatics, Vanderbilt University Medical Center, Nashville, TN

    Sruthi Adimadhyam, PhD; Adee Kennedy, MS, MPH; Department of Population Medicine, Harvard Pilgrim Health Care Institute and Harvard Medical School, Boston, MA