Project Grant R21DC022440

This $442,750 federal Project Grant awarded by the National Institutes of Health (NIH) Trans-NIH Research Support program (CFDA 93.310) aims to improve the identification of late-talking children and map their developmental trajectories using real-world data from electronic health records (EHRs). The research team from Duke University will employ novel machine learning approaches like natural language processing to more accurately identify late talkers within EHR databases and create open data resources to facilitate this process. The goal is to delineate distinct developmental trajectories associated with late language emergence, enabling more personalized early intervention approaches to improve outcomes for late-talking children. This research addresses critical gaps in current late-talking studies, which have relied on limited cohort data, and leverages the advantages of large-scale EHR data to gain better insights into this prevalent developmental concern affecting approximately one in five children in the United States.