Providence will recruit patients to help develop a new AI framework to manage and predict mental health crises, using an app and patient health records
RENTON, Wash. [May 14, 2026] – Providence is joining a nationwide, federally funded research project to develop a Large Health Behavior Model, a new class of artificial intelligence designed to advance the prediction, prevention and management of mental health and substance use disorders.
The Proactive Health Office of the Advanced Research Projects Agency for Health (ARPA-H) has awarded a $17.9-million contract award under their EVIDENT initiative to Ksana Health of Eugene, Ore. to develop the model. Providence is partnering with Ksana to recruit up to 25,000 adult patients to participate in data collection to inform the model.
“Research participants will download Ksana's health behavior app on their smartphone and participate in the app activities for three months,” said Staci Wendt, Ph.D., director of research for the Providence Health Research Accelerator. “Participants will also consent to have their health records included in the project.”
Recruitment will begin this summer across Providence’s seven-state footprint. All Providence patients 18 years or older, will be eligible to participate in the two-year project.
Despite the enormous burden of behavioral health conditions, which contribute to an estimated 8% loss of annual GDP, the current tools for screening and early intervention are limited. Measurement-based care is rarely implemented in behavioral health settings, and the approaches that do exist rely heavily on the patient discussing their situation with their provider. Patients and providers then navigate a time-consuming trial-and-error process to find effective treatments, with clinicians often unable to reliably predict patient outcomes using clinical judgment alone.
Providence, Ksana and other project partners want to change this by developing AI models trained on continuous behavioral data from smartphones and wearables, including patterns of physical activity, sleep, mobility, social connection, and language use linked to large-scale electronic health records. The approach draws on three converging technological advances: the ability to measure behavior at population scale through ubiquitous mobile devices, the aggregation of digitized medical records, and breakthroughs in foundation model architectures that can learn powerful representations from multi-modal data.
“Behavioral factors are a leading cause in a large majority of patient conditions that the health care industry treats,” said Bill Wright, Ph.D., chief research officer, Providence. “Given the proliferation of smart phones and watches, if we can harness the sensing capabilities of those devices and link behavioral signals to patient health records, we can build a new model of personalized, proactive behavioral health care. This innovation be a major component of the delivery model of the future.”
MedStar Health Research Institute of Columbia, Md., will recruit patients from its East Coast footprint to bring geographic diversity to the project. The University of Washington’s Paul Allen School of Computer Science & Engineering will spearhead the computational modeling effort.
About the Author
More Content by Providence News Team
