Principal Investigator: Vivek Srikumar
CoPrincipal Investigator(s): Zachary Imel, Michael Tanana, Eric Poitras
Organization: University of Utah
Millions of Americans have been diagnosed with mental health or substance abuse problems. While conversational interventions like psychotherapy and other forms of counseling are among the most effective treatment options available, less than half of those in need receive care. One problem is that it is difficult to provide therapists in training with regular feedback based on direct observation of their work. Accordingly, training therapists is expensive and time consuming, leading to a shortage of expert counselors. Furthermore, due to difficulties in obtaining care, patients are turning to online sources of support, where quality may be difficult to ascertain. This project examines this timely question of how to use technology to improve the training of mental health counselors at all levels. This project will develop a novel intelligent tutoring system to capitalize on developments in natural language processing and also facilitate collaboration among trainees to enhance learning. Better trained therapists and improved interactions on mental health forums should improve the quality and timeliness of mental health care for everyone. Furthermore, this work will support lifelong and collaborative learning for licensed professional mental health counselors.
This project will develop and evaluate technological tools that facilitate new models for training tomorrow’s mental health workforce. Specifically, this will involve creating and studying a novel text-based platform with the goal of training mental health counselors. Within this platform, two broad research questions include investigating the impact of (a) natural language processing driven helpers that provide feedback in real time, and, (b) crowd-sourced counseling using individuals with minimal training. To this end, several statistical models will be designed and trained to operate within the proposed text- based platform to interact with novice therapists. The efficacy of the two kinds of feedback (automatic and crowd-based) in terms of how well they can train different kinds of trainees (lay support providers in online forums, novice therapists in training) will be compared to models where the learner practices on their own and/or without specific feedback.
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.