Digital Mental Health Service for Non-Treatment Seeking Young Adults - PROJECT SUMMARY
Young adults aged 18-24 experience higher levels of mental health problems than any other adult age
group. Over one quarter of all young adults living in the United States suffer from a mental health condition.
Unfortunately, they are the adult age group who are least likely to seek or receive traditional face-to-face
treatments such as in-person psychotherapy or pharmacotherapy. There is evidence that they are, however,
interested in using digital mental health interventions (DMHIs), such as mobile and web-based apps, to support
symptom self-management and skill building. While evidence suggests mental health apps can effectively
reduce symptom severity, they require motivation from the user to open and use the app, which contributes to
the high dropout rates. The primary method of addressing dropout has been the use of human coaching, which
boosts engagement.
SMS text messages arrive through the most commonly used app on the phone, and are therefore likely to
be viewed. Initial work in text message interventions suggests good adherence, as effort is low, but
effectiveness has been inconsistent. Messages can be perceived as off-target or impersonal, and it is difficult
to convey more complex information. This project aims to address these problems by developing and piloting a
personalized text messaging platform that uses machine learning to tailor SMS messages to an individual’s
needs and preferences, and URL links to provide access to psychoeducational content to contextualize
messages, when the length of that content exceeds the limitations of messages. This project will include a
partnership with Mental Health America, the nation’s largest mental health advocacy organization.
The primary goals of the project are to: (1) Develop an adaptive messaging service for young adults that
personalizes messages and psychoeducational content to the needs and preferences of an individual, (2)
Conduct a feasibility trial using a sequential multiple assignment randomized treatment (SMART) design, which
will evalutate (a) the effectiveness of an adaptive, personalized messaging intervention in reducing
engagement relative to a static version; and (b) whether human coaching results in greater symptom reduction
and engagement, relative an unguided implementation.
This project will, in the near term, allow us to determine the feasibility of this intervention, including whether
our adaptive intervention affects treatment psychological and engagement targets, and reduces psychological
distress. It will also provide preliminary information on the feasibility of a scalable model of targeted, low-
intensity coaching for users who may require additional support above and beyond a fully automated
intervention. This will prepare us for our longer-term goal of conducting a fully-powered randomized controlled
trial of our adaptive intervention in an online community setting.
