New AI Mental Health Tools

New AI Mental Health Tools

If you’ve ever texted with your kids from work or used your Starbucks app to pre-order your double latte for easy pickup, then you’ve already seen how smartphones and digital technologies have revolutionized the way we communicate with each other.

With the help of artificial intelligence (AI) technologies to process the flow of information between providers and patients, could these tools revolutionize the way we deliver mental health services as well?

That intriguing prospect was the topic of a panel discussion, Mental Health, Smartphone Apps and the Promise of AI, at the 2019 World Medical Innovation Forum hosted by Partners HealthCare Innovation this spring.

According to the National Institutes of Mental Health, tens of millions of people suffer from mental illness in the United States each year, yet only half of those affected receive treatment.

This gap is due in part to a lack of appropriately trained mental health providers, inaccurate or outdated patient assessments, and the cost and logistical challenges of attending regular psychotherapy sessions, said panel moderator Sabine Wilhelm, PhD, Chief of Psychology in the Department of Psychiatry at Massachusetts General Hospital (MGH).

So, how could AI-powered technologies combine with smartphones and other digital technologies help to fill that gap?

Panel participants—comprising leaders from academic medicine, biotech and digital health—identified a variety of opportunities, including increasing access to care, improved stratification of patients based on illness type and current risk status, and opportunities to identify and engage patients at crucial points of need within the care process.

 

Powerful Potential

The biggest role for AI in this changing landscape will be to help providers make the best use of the large amount of patient data and feedback that can be passively collected via smartphone apps and digital devices, participants agreed. AI algorithms could be used to process those large data sets and identify actionable steps for intervention and treatment.

“I think our field is undergoing a remarkable evolution in terms of the taxonomy of these disorders,” said David Silbersweig, MD, Chairman, Department of Psychiatry, Brigham Health (BH). “What was based on subjective reporting will start to be supplemented by more objective data, measurements and quantification.”

AI could also help primary care providers keep tabs on the patients that they refer to mental health services to ensure follow-up.

“I had a very personal experience being in the Army, where I had a soldier who was being seen for chronic pain and was also diagnosed with depression,” said Christopher Molaro, CEO of Neuroflow, an integrated behavioral health platform.

“He was referred to a specialist, a psychiatrist, and three months later he ended up losing his battle and died by suicide. We investigated what happened, and it turned out not once in those three months did he follow up and go see a psychiatrist.”

“If we had AI tools there to understand that the soldier wasn’t on the right track, or if we could have been proactive in reaching out to him, I think it could have saved his life.”

 

Considerations and Concerns

While there were reasons to be optimistic about the potential of AI-powered technology to improve digital and app-based mental healthcare, there were reasons for caution as well.

Providing mental health services through mobile and digital platforms introduces the potential for data breaches and misuse, participants acknowledged. “As long as that data exists somewhere in the world, it has the capacity to be used for good stuff of bad stuff,” said Thomas McCoy, MD, Director of Research at the Center for Quantitative Health at MGH.

Participants also cited the need for more rigorous standards of quality control to ensure that the care delivered through mobile apps meets high clinical standards.

“It’s very important that it is evidence-based care that occurs,” said Jennifer Gentile, PsyD, SVP, U.S. Clinical Operations, of Ieso Digital Health, an online platform for cognitive behavioral therapy. “[It should not be] provided by an unlicensed clinician, or job coach, or by someone who really wants to give advice…but rather someone who has training and education and there’s a monitoring body assuring there’s good care [being provided].”

Apps should also be rigorously tested to ensure they can engage patients from a wide spectrum of socioeconomic, cultural and ethnic groups. And while some have worried that older patients may have a harder time accessing the technology, providers could help by providing a hand-on introduction.

“It’s not about technology or age,” said Jeremy Sohn, VP, Global Head of Digital Business Development and Licensing for Novartis, noting how quickly those aged 65+ were to pick up on applications such as Facebook and Uber. “What you see in terms of age [difference] is actually variation in terms of how they engage technology.”

 

Accessible and Engaging

Perhaps the biggest hurdle will be to keep patients engaged in the services once they gain access to them.

“I think [with] solely AI-based tools, we’ve shown over and over again that they work beautifully in clinical trial settings, but then in real life that they don’t engage, and we need that human component guiding that care,” said Gentile.

“It doesn’t mean that we should eschew these technologies,” Gentile added. “Hopefully, we understand what the existing boundaries are, we accept them, and then we try to work to improve them, right?”