A globe sitting in a book with coding digits surrounding it.

Partnering with Lyssn and the University of Denver, we’re deploying AI-assisted multicultural competency training to find out.

It remains difficult to access therapy. According to NAMI, less than 50% of adults experiencing mental illness received treatment in 2021. And if you were from a community of color, it is even harder to access care. In 2021, fewer than 40% received treatment and for some historically marginalized communities, it was less than 25%. While so many companies, health plans, and providers are rightfully working on the issue of increasing access to and quality of mental health care, one critical way to achieve this shared goal is upping clinicians’ cultural competency skills to meet the needs and expectations of clients from all cultural backgrounds.

Right now, most clinicians receive training on multicultural competence, but only limited opportunities for feedback on how well they are applying the skills in real world patient and client encounters. Anyone can sit through a training, but the goal has to be demonstrated improvements. Historically, receiving feedback on multicultural competency skills required a human evaluator to manually observe the clinician and offer feedback, which is not scalable nor could produce useful data across organizations. And the status quo for multicultural training has largely been unsuccessful for changing racial/ethnic disparities in therapy outcomes, or to offer any innovative solutions to evaluate or improve population-wide quality. Given the vast need for this type of training, we have to find impactful solutions that can be deployed to thousands–not dozens–of clinicians at a time, and that actually make a difference for individuals.

Technology as a cultural competency trainer

This is where AI comes in. SonderMind and Lyssn.io will be partnering with Dr. Jesse Owen at the University of Denver on his grant, just shy of $500,000 from the John Templeton Foundation, to increase mental health clinician cultural competence using AI/ML and deliberate practice via a randomized controlled trial. 

We have brought together an interdisciplinary team of intellectual humility researchers, computer scientists, psychometricians, and Ph.D level counseling/clinical psychologists to conduct a comprehensive project designed to improve multicultural training. We are proposing that intellectual humility is an ideal framework for training the acquisition of knowledge regarding therapists’ multicultural orientation (i.e., cultural humility, cultural opportunities, cultural comfort). The training, delivered through Lyssn’s innovative platform (validated across 60+ peer reviewed studies), will include learning modules delivered by a multicultural orientation instructor, written content, followed by deliberate practice videos with immediate intellectual humility feedback. This feedback will be delivered via artificial intelligence, informed by previously established machine learning models and based on industry wide gold standard quality metrics.

Clinicians as trial participants, and discovering if AI is a better teacher

We expect this innovative approach to produce new training outcomes for two reasons:

First, randomized control trials rarely engage with independent practice therapists. In our study, 175 active SonderMind providers will participate. Second, participating clinicians will practice in the privacy of their own spaces without fear of saying the wrong thing to another human. Through the trial, they will get repeated, real time feedback and coaching via the AI tools developed for this study. We expect skills improvement when clinicians are comfortable saying what they would naturally say to a client, combined with actionable and appropriate feedback.

We believe that involving practicing clinicians in this trial and offering them this low-stakes way to improve their cultural competency skills will demonstrate the efficacy of this training and the impact we can have on therapy outcomes for individuals from all backgrounds and walks of life.

How we’ll know if AI is a good teacher

The goal of our trial is to improve access to quality care for individuals from historically marginalized groups, ethnic backgrounds, cultures, and identities. Ideally, the learnings from this research will be scaled to thousands of clinicians in just a matter of months after the study is completed. 

So how will we know if it works?

We will assess outcomes for the participating SonderMind clinicians both at the therapist level, as well as their client therapy outcomes. We will also conduct qualitative interviews with select participants about their training experiences. Ultimately, our hope is that we will empower therapists to increase their multicultural orientation and intellectual humility, thereby improving their clients’ therapy outcomes–and allowing them to serve a more diverse population of clients.

Addressing the mental health provider shortage and improving access to quality care requires innovative, outside-the-box thinking, one tailor made for a partnership between academia, healthcare technology companies, and a philanthropic organization. By thinking about ways to increase cultural competency amongst practicing clinicians and those who are currently in training to become clinicians, we can better meet the needs of individuals who are acutely underserved right now.

We look forward to undertaking this important work.