UCHealth and New York-based startup HealthRhythms are harnessing the ubiquity of smartphones, insights from decades of clinical research, and artificial intelligence to identify the signatures of mania, depression, schizophrenia, and other mental illnesses.

The goal is to identify the telltale changes in movement patterns and the circadian (sleep-wake) cycle that can trigger behavioral-health episodes early enough to steer clear of the depths or even avoid them entirely.

The science rests upon a simple idea: We are creatures of habit. We tend to feel better when we keep a consistent sleep schedule, when we exercise regularly, and when we have events to attend and friends to meet up with. Disruptions to these life patterns – not to mention major upheavals such as illnesses and deaths in the family, job losses, and divorces – can reverberate through our general sense of well-being.

These patterns matter more to some people than others. Research going back 35 years has found that those with bipolar disorder are particularly vulnerable when their circadian and social rhythms, as these life patterns are variously called, miss a beat. They can find themselves soaring into a manic episode or plunging into depression as a result.

Researchers have found that identifying the disrupted life pattern(s) that trigger a manic or depressive episode can speed recovery and even stop future episodes from happening at all. That identification and intervention has been formalized through something called interpersonal and social rhythm therapy, or IPSRT. Its pioneers included emeritus University of Pittsburgh psychiatrists Dr. Ellen Frank and Dr. David Kupfer.

Frank and Kupfer have since cofounded HealthRhythms. Its goal is more ambitious than just partially automating social rhythm therapy. They’re using smartphone-sensor data in a privacy-protected way to passively and continuously monitor mental health. As HealthRhythms’s Chief Medical Officer Dr. Gabriel Aranovich put it, “We aim to make precision psychiatry a reality.”

A partnership with the UCHealth CARE Innovation Center is central to that effort.

“UCHealth is a dream partner. They’re visionaries, but they also know how to turn that vision into reality,” Aranovich, himself a psychiatrist, said. “Taking this vision and translating it into reality and scaling it means saving lives.”

Movement tells a story

A HealthRhythms-UCHealth pilot study kicked off in February with 10 volunteers and will continue through the summer with 90 more. All are UCHealth patients being treated for bipolar disorder or major depression. The effort focuses not on proving whether HealthRhythms’ approach works – that’s been shown already – but rather how to build it into the electronic health record, mobile-phone app, and patient-care processes of a major health care system, says Dr. Richard Zane, UCHealth’s chief innovation officer and chair of the CU School of Medicine’s Department of Emergency Medicine.

“What we are working on is that last mile of integration and bringing it into practice,” Zane said.

Integration of a tool aimed at improving patients’ behavioral health makes sense for UCHealth, Zane adds. Foremost, psychological well-being is as central to overall health as is well-managed blood pressure or blood sugar. But also, patients with chronic health problems such as diabetes, chronic heart failure, multiple sclerosis and other diseases do worse when behavioral health issues are part of the mix: When you don’t feel well, it’s easy to feel down, and if you feel down, you’re less likely to take good care of yourself.

While the artificial intelligence that powers the HealthRhythms app is far from simple, the information collected is surprisingly straightforward. The system is interested in movement – how often you’re moving, how far you’re moving, and how fast you’re moving. It’s not interested in your precise location – just the movement related to that location. That also helps ensure confidentiality, a paramount concern for a company with this target audience and a leadership team stocked with psychiatrists.

“We don’t care specifically where you are. We’re not looking at your specific location – we’re looking at patterns of movement,” Aranovich said.

Early warning system

It may seem far-fetched that a better understanding of the sleep-wake cycle and movements throughout the day can provide valuable insight into a patient’s mental state, much less predict a crisis a solid week in advance, as HealthRhythms says its system can do. But peer-reviewed studies have shown it to be true. To choose a straightforward example, a patient amid a manic episode typically sleeps much less, moves around a lot more, and moves faster and to more novel destinations than when in a steady state, Aranovich says. Conversely, a depressed patient moves around less.

“They’re holed up, and they also move more slowly,” Aranovich said.

The idea isn’t to replace provider-to-patient psychiatric care with a phone app, says Dr. C. Neill Epperson, chair of the University of Colorado School of Medicine’s Department of Psychiatry and a champion of the HealthRhythms-UCHealth program. Rather, it’s to deliver ongoing, real-time insight into the patient’s state of mind.

These insights feed into an advance-warning system with automated and human elements. Epperson, who has HealthRhythms on her own phone, periodically gets a brief spot-check survey based on established mood-measuring questionnaires such as the PHQ-9 and the GAD-7. In Epperson’s case, they’re merely periodic, but HealthRhythms also sends them out when it detects a patient’s social rhythms to be changing.

Based on the social-rhythm data and answers to the survey questions, human intervention can come into play, and that’s one place where integration with UCHealth’s Epic electronic health record (which powers MyHealthConnection online and on the UCHealth app) is important. HealthRhythms can send an alert to the 24/7 UCHealth Virtual Behavioral Health Center, triggering a call to a patient, as well as to the patient’s psychiatrist. Part of the work of the pilot study is to fine-tune the thresholds of when to send an alert and to which providers, says Pam Villareal, the UCHealth CARE Innovation Center’s project manager for the HealthRhythms program.

Doctors may benefit

A core UCHealth CARE Innovation Center precept is to introduce technologies such as HealthRhythms as part of the UCHealth app patients are already familiar with and to keep additional screens for doctors and other caregivers to a minimum. The pilot-study team is also working to create a seamless experience for both users and providers by integrating with the UCHealth app on the patient side and, for providers, with the Epic electronic health record and a HealthRhythms provider dashboard.

“We want to gather input from providers as well,” Villareal said. “How is this additional data helping them interact with patients and giving them a better picture of the rhythms of those patients?”

Aranovich, for example, uses HealthRhythms with his own psychiatric patients. He’ll bring up the data onscreen in clinic and point out patterns that show sleep-quality deterioration, mood improvements, and other factors.

More broadly, Epperson says, “the biggest reason that this is so important is that we can’t treat our way out of this mental health crisis. We have to be more proactive in terms of prevention.”

“Given limited resources, we have to treat people who need the highest level of care,” she said.

While the pilot study involves UCHealth patients with behavioral health diagnoses, Villareal, Epperson and others say HealthRhythms could, with time, provide early-warning signals for those with no history of mental health issues – or even for patients with other conditions. Epperson says discussions with leaders of clinics treating diabetes and multiple sclerosis patients have drawn interest both with respect to possible mood-related diagnoses, but also with respect to clues HealthRhythms could deliver about, say, a multiple sclerosis flare-up that a sudden lack of movement might indicate.

“I think this is going to be an amazing tool, and I think it’s going to make a tremendous difference in how we help our patients,” Epperson said.