I-Corps: Translation potential of an emotional intelligence platform for use in clinical note taking settings

Award ID: 2629680

This I-Corps project is based on the development of a software platform that automatically creates and double-checks written records of spoken conversations. The initial application is in clinical care. Documenting what was said during important conversations has become one of the heaviest burdens on clinical professionals, contributing to widespread burnout and reducing the time they have for the patients in their care. However, the current new generation of automated note-taking tools introduces a risk as the clinical notes they produce cannot be guaranteed to match what was said. This technology addresses these problems by drawing every part of the record directly from the conversation and automatically checking each statement before finalizing the record, giving users a fast way to trust and stand behind the record while reclaiming time for patient care. In addition, this technology goes beyond the note-writer function by examining, tracking and analyzing emotions in the clinical visit. It is designed to temporally trend emotions and responses to those emotions. Reliable records of spoken interactions are needed in many fields, including private and group medical practices, academic medical centers, training programs, and large health systems, with further potential anywhere accurate records of conversations carry legal, financial, or safety consequences. This technology may reduce physician documentation burden and help to reclaim meaningful face time with patients, improve note quality without requiring any change to clinical workflow, and improve patient satisfaction with the quality of care provided. This I-Corps project utilizes experiential learning coupled with first-hand investigation of the industry ecosystem to assess the translation potential of an evidence-grounded cognitive-intelligence system for documenting and analyzing spoken encounters. This technology converts a single recorded conversation into an organized note that captures both the actual content and additional context, including measures of empathy and patient-centered communication. It is based on the development of an automated verification method anchored to the encounter where the system breaks its own draft into individual statements, assembles supporting evidence drawn only from the conversation, tests whether each statement is supported, and then corrects, flags, reprocesses, or removes anything that is not supported. This verification step directly targets the reliability concerns that limit adoption in settings where errors carry real consequences. In a controlled, blinded study using simulated clinical visits, the approach produced records that were 93 percent free of fabricated or erroneous content, outperforming both popular automated tools and human-written notes and scored highest across the quality measures evaluated. This technology may provide clinician well-being, patient trust, operational efficiency, and quality care for patients. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Program Director: Ruth Shuman