Can Digital Therapeutics Help Solve the Mental Health Crisis?

Published: November 26, 2024

By Jeremy Martin

4 min read

Few aspiring psychotherapists plan on building apps that mirror human sentience. “I originally set out thinking I was going to be a clinician,” says Nick Jacobson, PhD, an associate professor at the Geisel School of Medicine at Dartmouth, reflecting on his student years. “That was going to be my impact on the world.”

Although Jacobson didn’t become a therapist, his original goal did lay the groundwork for a career in mental and behavioral health. Jacobson now works at Dartmouth’s Center for Technology and Behavioral Health (CTBH), where he designs cutting-edge digital therapeutics and AI-powered technologies to address the challenges in mental health treatment and accessibility. Jacobson’s work in digital mental health has garnered significant attention both locally and nationally, and he was a featured speaker at Dartmouth’s inaugural Innovation in Medicine & Healthcare Summit in September.

Jacobson’s journey from therapy to technology began in graduate school, where he started developing Mood Triggers, a smartphone app that helps users identify triggers for their anxiety and depression by tracking daily moods and behaviors. Using smartphone sensor data, the app tracks symptoms and delivers personalized feedback to users on the factors influencing their mental health.

Shortly after launching the project in 2012, Jacobson had a major epiphany—one that drives his work at CTBH to this day. “I realized with Mood Triggers, I’d [already] treated more people than I could have over an entire career as a full-time clinician,” he says. “That, to me, really spoke to the salience of how accessible and scalable these technologies are—and the potential impact they can have on society.”

Help for Everybody—Anywhere, Anytime

For many patients, the limited availability of therapists can be an insurmountable challenge to receiving mental healthcare. Roughly 122 million Americans live in areas without enough mental health professionals. And more than 28 million U.S. adults with a mental illness are not getting treatment.

Those who do manage to connect with a therapist also often struggle to get help at critical times. “A lot of that, I think, is a scale problem,” Jacobson says.

While some evidence suggests technology is harmful to mental health and Jacobson openly acknowledges the risks, to him the ubiquity of devices is an opportunity for digital therapeutics to satisfy an urgent, often unmet need. “If that’s where [people] are, I’d rather meet them where they’re at than try to deliver something outside of it,” he says.

Such digital tools are typically designed to complement, not replace, healthcare professionals. As Jacobson explains, these tools aim to “be constantly available” and provide support “when you would not be able to reach a provider.” They augment therapists’ work by providing additional resources between sessions, addressing the reality that in traditional care, you’re typically “seeing somebody less than 1% of their life,” he says. While adhering to the same ethical guidelines, confidentiality, and privacy principles as traditional therapy, these digital interventions incorporate data protection protocols and are developed based on established therapeutic approaches.

With this approach, CTBH’s digital health tools can provide timely, personalized support to a wider audience. “The goal is to disseminate evidence-based interventions to those who need it—when they need it. And using generative AI in psychotherapy offers a great solution to the scalability and availability issues in mental health.”

Quality vs. Quantity

For a machine to accurately emulate human psychotherapy requires researchers to train models on good data. Such input could run the gamut from online peer support interaction logs to psychotherapy training transcripts. After years of iterating with multiple forms of data, Jacobson had another epiphany. “It became clear to me that the data was the biggest governing problem more than anything else,” he says.

Quality data, he learned, was far more important than quantity in order to create safe, helpful content in the style and tone that adheres to evidence-based therapeutic principles. As a result, Jacobson organized a team to create the content for models to learn from so that they could guide and control the responses. “That’s when things got better and better and better. So our secret sauce is ultimately the data.”

With this same ethos, Jacobson and his team trained the generative AI model for his newest venture, Therabot, beginning in 2019. While not a replacement for a human provider, Therabot is a chatbot that offers empirically supported treatment in a manner similar to a human therapist. So if it’s the middle of the night and you can’t sleep; if you’re in recovery but can’t connect with your sponsor; or if you’re grieving the loss of a loved one, but your next sessionisn’t for a few days, Therabot is available to you—24/7.

Stopping Mental Health Crises Before They Start

At the heart of Jacobson’s technology is the recognition that traditional mental healthcare models both work and often fall short in meeting modern society’s needs. “A lot of the theories and models we work with are established in cognitive-behavioral therapy (CBT). It really mirrors how people actually use psychotherapy, which is incredibly broad in terms of what concerns people from day to day,” he observes, underscoring the versatility and adaptability of digital interventions.

Predicting mental and behavioral health patterns is another advantage of digital therapeutics. While human therapists can and do learn about their clients over time, predictive AI models can analyze data and behavioral activity to identify, manage, prevent, and treat mental health disorders and potential crises. Using passively collected data from smartphone sensors or wearables, “We can predict the severity, symptoms, and course of mental illness,” he says. “We can know when folks are in the greatest need, when things escalate or before they escalate, and then provide an empirically proven intervention strategy.”

The implications of these predictive capabilities are profound. One major goal of Jacobson’s lab is to “accurately predict the ebbs and flows of anxiety and depressive symptoms before they occur,” he says. This early detection could allow more timely interventions to prevent crises, reduce hospitalizations, and improve patient outcomes overall. With widespread adoption, such technology could enable proactive rather than reactive mental health care on a large scale.

How the Apps Change Users’ Habits

The real challenge, in human and digital therapy, is changing behavior. This is why Jacobson deliberately designed Mood Triggers and Therabot to incorporate micro-doses of CBT techniques. Delivered through short videos and questions that can be easily integrated into daily life, both apps break down interventions into bite-sized steps that make it easier to see tangible progress. Like traditional therapy, Mood Triggers and Therabot encourage pro-social behavior, exercise, journaling, and scheduling time for activities they enjoy.

Trusting the process and harboring the expectation that an intervention could work, Jacobson says, are an absolute necessity for behavior change. “The more a person understands what benefits them, the more they start to consider behavior change and are willing to give it a try,” he says. “Starting to get the ball rolling is often the harder thing than keeping it rolling.”

Once the ball does get rolling, users will often “develop interesting attachments to these digital tools.” Jacobson refers to this curious new rapport between humans and health technologies as “the digital working alliance.” Along with the belief that treatment will work, a healthy attachment to one’s therapist—or machine—is a strong indicator for successful treatment and a crucial factor in facilitating behavior change.

To strengthen this digital alliance, Jacobson has released and plans to continue releasing digital interventions to the public free of charge. Meanwhile, Therabot is currently in its first clinical trial with several hundred test participants. “The ultimate goal is to disseminate this to as many people as possible,” Jacobson says.

With demand for mental health services higher than ever, and millions of Americans forgoing treatment, digital therapeutics may not be the singular solution to the mental health crisis, but according to Jacobson, “They’re a solution.”