U-M researchers aim to shorten the path to effective mental health care

By Don Jordan

For many seeking mental health care across the country, the first obstacle is simply getting an appointment, a process that can take weeks or even months as demand exceeds available clinicians.

But what if that wait wasn’t just lost time? What if patients could improve during the delay, and the information gathered during that period — and throughout the first year of care — could help clinicians connect patients with the support most likely to help them?

Researchers at the University of Michigan aim to answer those questions through COMPASS, or Comprehensive Mobile Precision Approach for Scalable Solutions in Mental Health Treatment, a five-year, $17.9 million study collecting genetic, behavioral, clinical and survey data from U-M patients to help clinicians more quickly match patients with care that fits their needs.

So far, 6,300 U-M Health patients have enrolled in the study before beginning mental health treatment. Participants receive access to mental health apps on their phones and wearable devices that track data on sleep, activity and heart rate. They also complete surveys and genetic testing that may reveal predispositions to response to certain treatments.

Researchers hope this vast dataset will help reduce some of the frustrating trial and error that often comes with treating depression, anxiety and other mental health concerns.

“We don’t know what’s going to work for (individual patients), so we try different treatments,” said Srijan Sen, the Frances and Kenneth Eisenberg Professor of Depression and Neurosciences and director of the Eisenberg Family Depression Center. “The first one often doesn’t work and people cycle through many treatments, often for years. Sometimes we find one that works, sometimes not, but it’s a really long journey. It’s frustrating, and as someone stays in the system for longer, a new person can’t come in. We’re hoping to help that process, and are encouraged with the early progress in being able to identify key factors for treatment matching and to help people before they come into the clinic.”

The goal is providing clinicians and patients with additional tools and data that will make it easier to predict which interventions – medications, exercise habits, sleep routines, etc. – will work best for each individual. The project tested digital interventions designed to support them during that down time and found that participants showed measurable improvements in anxiety, depression and suicidality, with over 20% achieving remission before they had their clinic appointment.

COMPASS is led by Sen, Amy Bohnert, professor of anesthesiology, psychiatry and epidemiology, and Lars Fritsche, research associate professor of biostatistics and anesthesiology. They work with a team of researchers across seven disciplines and departments.

That interdisciplinary focus leads to new ways of thinking, and is something unique about U-M, Bohnert said.

“A really special thing about being able to bring in collaborators outside of your own field is they bring in new ideas from a completely different perspective,” said Bohnert, who also serves as co-director of the university’s Opioid Research Institute.

Bohnert points to an example in which colleagues with expertise in bioinformatics have helped the team think about how to measure which behavioral factors are particularly important to mood for an individual and then using large language models to share that information with study participants.

“We find that some people are particularly sensitive to exercise, where that really impacts the mood, and other people are really sensitive to sleep, “ she said. “We even find some people are sensitive to exercise where more exercise makes their mood worse, which we didn’t expect to find. The data has brought a new perspective, and I’m really excited for what potential impact it has for our participants.”

COMPASS is funded by the National Institute of Mental Health, part of the National Institutes of Health, and builds on years of related mental health research at U-M using wearable devices, surveys and genetic data.

Most notably, COMPASS builds on the Intern Health Study, launched in 2007, which tracks stress and mood among medical interns in the United States and China. Using wearables, daily mood diaries and genetic samples, the study aims to better understand why some people develop depression under prolonged stress while others do not.

Sen said studies like COMPASS succeed because U-M brings together the technical expertise, clinical access and interdisciplinary collaboration needed to test ideas at scale.

“I don’t think the study could happen anywhere but Michigan, having world-class expertise in so many different areas from engineering and understanding how technology works, psychology and psychiatry, epidemiology, biostatistics, math and bioinformatics and computer science,” Sen said. “We have some of the world’s smartest people here and also an incredible atmosphere where people are willing to collaborate and work together.”

For Bohnert, the study’s ultimate promise lies in using AI-assisted tools to shorten the frustrating trial-and-error process that often defines mental health treatment, helping patients find effective care sooner.

“If we can get that choice right the first time by using these algorithms that match patients to treatments, we’ll be much more likely to get people better faster.”