Premium
P4‐555: EHR‐BASED PATIENT RISK STRATIFICATION TOOL FOR PROBABLE AD
Author(s) -
Tjandra Donna,
Migrino Raymond,
Giordani Bruno,
Wiens Jenna
Publication year - 2019
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2019.08.102
Subject(s) - interquartile range , medicine , receiver operating characteristic , covariate , predictive power , medical diagnosis , retrospective cohort study , cohort , population , medical record , electronic health record , emergency medicine , health care , machine learning , computer science , philosophy , environmental health , epistemology , pathology , economics , economic growth
Background: To date, predictive modeling for Alzheimer’s disease (AD) risk has focused on data not routinely collected in clinical care and is limited to short prediction horizons (e.g., 2-4 years). Given the limitations of existing datasets, we sought to leverage electronic health records (EHRs) that contain decades of longitudinal data for thousands of patients, with the goal of developing and validating a predictive model for AD onset with a 10-year prediction horizon.