Using AI to Improve Teen Mental Health Helplines

Dr. Sylvanna M. Vargas is a UC Chancellor’s Postdoctoral Fellow at the UCLA Psychology Department/ Chicano Studies Research Center. She is also a Clinical Instructor at the UCLA David Geffen School of Medicine’s Department of Psychiatry. She completed her clinical psychology training at the University of Southern California (PhD, MPH) and West Los Angeles VA Medical Center (pre-doctoral internship). The overarching goal of her research is to develop innovative and large-scale solutions that improve the landscape of mental health services for underserved youth experiencing severe mental health concerns, including depression and suicidality. Her research employs translational methods that span the development of basic science on social determinants of mental health, to the design and implementation of culturally- and contextually-responsive evidence-based interventions that mitigate inequities. She hasobtained support for her work from the National Academies of Sciences, Engineering, and Medicine’s (NASEM) Ford Foundation, National Institute of Mental Health, American Psychological Foundation, and Foundation for Psychocultural Research.

Project Summary:

Thousands of youths use peer-delivered mental health helplines every year. Demand, particularly over text and online chat, has surged. As a result, new helpline programs have proliferated, including ones relying on teen peer helpers. However, little is known abouthow peer helpers are trained. Helplines may be especially useful for reaching youth with unmet mental health needs, and provide an opportunity to mobilize youth towards seeking help in real life (IRL; e.g., from caregivers or specialty mental health care). Motivational Interviewing (MI) is an evidence-based intervention that has been shown to support help-seeking. It is possible that teen peer helpers delivering MI in a helpline setting may be able to encourage IRL help-seeking among youth users, but it is not known whether teen helpers can adherently deliver this intervention.

To support skill acquisition and intervention adherence, best practices for training on evidence-based interventions include ongoing performance-based feedback. However, performance feedback through live supervision is resource intensive and reduces feasibility within helplines at scale. Technological advances, such as Artificial Intelligence (AI), can help.Performance feedback delivered by AI has been shown to improve the use of effective counseling strategies among clinicians and online peer helpers. Therefore, performance feedback delivered by AI may support adherent MI use in teen peer-delivered helplines.

Fellowship:

My program of research aims to increase the quality and reach of mental health care, with an eye on improving services for the most vulnerable. This fellowship will support my research aims through a community-partnered pilot project. This effort willultimately adapt and test an MI training and AI-derived performance feedback tool for teen peer helpers on a chat-based mental health helpline.

Impact:

This research could capitalize on the promise of teen peer helpers by informing future evidence-based training standards. It will help address urgent questions regarding the role of emerging technology in supporting scalable training of the mental health workforce. Overall, a scalable solution would help us better understand how to integrate evidence-based practices in a large and wide-reaching, but understudied, service sector.