Automating pediatric audiology reporting to public health using AI
Jurisdictional Early Hearing Detection and Intervention (EHDI) programs track newborn hearing screenings, audiological results and intervention referrals to ensure infants suspected of being deaf or hard of hearing (DHH) receive timely care and connection to services. EHDI programs frequently face under-reporting or delayed data from audiology providers due to voluntary reporting laws and time-consuming, duplicative data entry processes.
PHII led a multi-sector collaboration involving the Centers for Disease Control and Prevention (CDC), Boston Children’s Hospital, Mass Eye and Ear, the Massachusetts Department of Public Health’s Universal Newborn Hearing Screening (UNHS) program, and the Digital Transformation Hub (DxHub) at California Polytechnic State University, San Luis Obispo (Cal Poly). The collaboration led to the development of an open-source generative artificial intelligence (gen AI) solution using large language models (LLMs) and a configurable logic engine.
You’ll find resources on this page designed to support the implementation of processes to reduce lost to follow-up (LTFU) and lost to documentation (LTD), and reduce the burden of reporting to audiologists.
Why it matters
By converting text reports into a format that state EHDI-IS platforms can automatically ingest, this technology delivers a two-fold benefit:
- For audiologists: eliminates the burden of manual, dual data entry.
- For State EHDI Programs: reduces delayed or missing data, enabling faster, more efficient and complete tracking of early audiologic assessments for infants.
Read a blog post about how GenAI was leveraged to automate the classification of pediatric audiology reports.
Interested in doing a similar project?
These resources can support you in AI-related audiology work:
These slides introduce the Automated Audiology Concept Extraction (AACE) project and provide background information, an overview, and next steps for implementation.
This document summarizes the Automated Audiology Concept Extraction (AACE) project, highlighting its primary objectives, public health benefits, operational approach, and potential for scalability. It also provides technical project details and definitions of commonly used audiology-related acronyms.
View the open source code and technical specifications for this GenAI solution.
This document provides lessons learned during the implementation the Automated Audiology Concept Extraction (AACE) project, a pediatric audiology reporting effort that used GenAI to translate clinical notes into categorical values.
Watch a video demonstration of the solution:
This scoping review aims to identify gaps and promising practices and to propose recommendations for EHDI programs integrating new data sources for the timely identification of infants who are DHH.
This self-assessment helps EHDI programs strengthen their ability to exchange electronic health information with electronic health record systems (EHRs). It supports strategic planning, facilitates discussions with leadership and IT partners and clarifies current capabilities and success factors. It also outlines steps programs can take to improve interoperability with clinical partners.
Video: Bridging the gap in hearing loss intervention with data
For questions about EHDI resources on this page, please contact us (opens in new tab).
These resources were developed by PHII, through funding from the Centers for Disease Control and Prevention (CDC) of the U.S. Department of Health and Human Services (HHS) under grant number 5 NU38PW000002.
PHII is a program of The Task Force for Global Health, a 501(c)(3) nonprofit organization that was founded as The Task Force for Child Survival in 1984. The Task Force is affiliated with Emory University.