SBIR Phase I: Automated Emotional Distress Severity Classification Using Speech Analytics and SFSS for SUD and OUD-Related ACE and Trauma

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project will result from a focus on families with low socioeconomic status (SES), in whom adverse childhood experiences (ACE) and opioid use disorder (OUD) are common. Children of parents with OUD have higher incidences of ACE due to neglect, physical abuse, or domestic violence. ACE have life-long adverse consequences, including drug abuse, early disability, and death. By improving the timely and objective measurement of mental health issues in children and adolescents using innovative technology, the company aims to help providers and public funding sources, such as Medicaid, meet the triple aim of improving care, quality and experience at a lower cost. The analytic solutions are intended to help mental health professionals working with children and adolescents identify problems and emotional disorder severity early and efficiently, and track treatment progress and outcomes systematically. This Small Business Innovation Research (SBIR) Phase I project aims to develop a machine learning (ML) algorithm that detects clinically relevant emotional distress in speech samples from at-risk youth receiving mental health and family preservation services. Following preliminary work with two community behavioral health organizations that serve rural and urban families in Georgia, the company has an operational cloud-based digital health platform that integrates the collection of voice samples from minors with a validated youth mental health survey. In collaboration with a voice and speech signal processing expert at the Georgia Institute of Technology, early work has advanced two ML models that will be further developed and validated in the proposed project. In Aim 1, Symptoms and Functioning Severity Scale (SFSS) and voice data will be systematically collected by therapists using the company app at the point of care in a community-based sample of youth receiving mental health services. EMR clinical data will help categorize voice sample data to train, validate, and test the ML algorithms. In Aim 2, independent evaluators will observe and make redesign recommendations of the therapist training and implementation protocol. The deliverable is a stakeholder centered ML treatment outcome tracking platform for the behavioral healthcare system. 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.