The Bridge2AI-Voice Application: Initial Feasibility Study of Voice Data Acquisition Through Mobile Health Bridge2AI Voice-App: Feasibility Study

Elijah Moothedan ¹ Micah Boyer ² Stephanie Watts ² Yassmeen Abdel-Aty ² Satrajit S Ghosh ³ Anaïs Rameau ⁴ Alexandros Sigaras ⁵ Olivier Elemento ⁵ Yael Bensoussan ^2*

• ¹ Charles E. Schmidt College of Medicine, Florida Atlantic University, Boca Raton, United States
• ² Department of Otolaryngology - Head & Neck Surgery, Morsani College of Medicine, USF Health, Tampa, Florida, United States
• ³ McGovern Institute for Brain Research, School of Science, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States
• ⁴ Sean Parker Institute for the Voice, Weill Cornell Medicine, Cornell University, New York, New York, United States
• ⁵ Englander Institute for Precision Medicine, Weill Cornell Medicine, Cornell University, New York, New York, United States

Bridge2AI-Voice, a collaborative multi-institutional consortium, aims to generate a large-scale, ethically sourced voice, speech, and cough database linked to health metadata in order to support AI-driven research. A novel smartphone application, the Bridge2AI-Voice app, was created to collect standardized recordings of acoustic tasks, validated patient questionnaires, and validated patient reported outcomes. Before broad data collection, a feasibility study was undertaken to assess the viability of the app in a clinical setting through task performance metrics and participant feedback.Participants were recruited from a tertiary academic voice center. Participants were instructed to complete a series of tasks through the application on an iPad. The Plan-Do-Study-Act model for quality improvement was implemented. Data collected included demographics and task metrics including time of completion, successful task/recording completion, and need for assistance. Participant feedback was measured by a qualitative interview adapted from the Mobile App Rating Scale.Forty-seven participants were enrolled (61% female, 92% reported primary language of English, mean age of 58.3 years). All owned smart devices, with 49% using mobile health apps. Overall task completion rate was 68%, with acoustic tasks successfully recorded in 41% of cases. Participants requested assistance in 41% of successfully completed tasks, with challenges mainly related to design and instruction understandability. Interview responses reflected favorable perception of voice-screening apps and their features.Findings suggest that the Bridge2AI-Voice application is a promising tool for voice data acquisition in a clinical setting. However, development of improved User Interface/User Experience and broader, diverse feasibility studies are needed for a usable tool.