Marblesapp and Analytics: Algorithms Improving Mental Health

Program Description

Marblesapp and Analytics is providing a free, virtual medium to address, understand, and refer individuals with mental health problems 24/7. Since October 2015, Marbles iOS and android apps have gathered over 38,000 unique data points from our users self-reporting their mental health status and facilitated nearly 800 conversations. Users can anonymously converse about mental health challenges and build support. We are currently researching whether Marbles can be a viable evidence-based intervention for undergraduates experiencing anxiety and/or depression (Study 1). We are also building an algorithm to understand what and how varying types of user posts (i.e. a user posts something like “I just don’t want to go on anymore”) are categorized by sentiment, solved, and referred to additional mental health services (Study 2) using quantitative and qualitative data. We want to figure out what types of mental health issues users can address virtually through peer support and how to escalate more serious or unsolvable issues to mental health professionals. Our goal is to eliminate student suicide at colleges by improving the identification, resolution, and referral of mental health issues by crowd-sourcing support and using big data and artificial intelligence to efficiently connect users to additional services.

Creativity

We use a peer-to-peer model to crowd-source support and mental health recommendations – anyone can download, post, or reply. This makes the support and diversity of thought shared on the app extremely wide and deep because it can come from anyone. There is some added risk to this model, however, the tone of the discussions is one of genuine care, concern, and support – we have never had any complaints of abuse or bullying. In addition to providing support for users acute, episodic mental health crises, we hope to leverage technology and artificial intelligence get the right person to the right level of therapy, at the right time, for the lowest cost. We are hoping this algorithm built and tested in Study 2 can eventually be applied to other mental health diagnoses like anxiety, BPD, PTSD, and more.

Leadership

Marbles is built for colleges that are struggling to address student mental health needs – Adam struggled with depression and suicide in college. Marbles is a more than just a virtual mental health safety net, it gathers valuable mental health data on the real-time changes in campus mental health and shares that data with administrators and decision-makers via an online web portal. We want to re-define mental health policy, outreach, and health promotion from blanket approaches to data-driven, highly nimble, customizable, and personalized engagement to improve overall effectiveness and eliminate campus suicide. Our two studies and pilot sites will validate if and how this may be possible. We will be publishing the results in academic journals and speaking at industry conferences. Adam is scheduled to present Study 1 preliminary findings at the Building Healthy Academics Community Conference in April and will be applying to more conferences and journals this summer.

Sustainability

Marbles founder Adam currently invests a portion of his paychecks from his full-time job to sustain the efforts. The University of Minnesota has also contributed $10,000 in grants and has allocated $25,000 for evaluating Marbles as a viable precision medicine intervention for depression (Study 2). We hope to sign other colleges as clients to implement Marbles and turn it into a sustaining effort across the US. Marbles has secured an additional pilot site at Cleveland State University and is actively engaging other colleges. 1,100 college students die by suicide each year and we are confident Marbles can help reverse this trend. Using Marbles as a virtual safety net to address acute, episodic mental health crises, colleges could avoid needing to hire additional staff. We believe adding Marbles to a college campus could replace at least 1 mental health professional per 10,000 students and are actively validating that assumption.

Replicability

We’ve built a plug-and-play solution for any college (or high school, employer, health system, or insurer) and the apps and web platform are currently operational. Any institution across the US could start using the platform tomorrow. Because the solution is entirely virtual and not constrained by any staffing or personnel, it’s extremely scalable and only gets stronger as more individuals use the apps. The algorithm also has incredible potential to be applied not just to Marbles, but other text (social media, messaging apps, or text crisis lines) and/or voice mediums (coaching, tele and virtual therapy, and crisis lines). We can apply the algorithms identification, resolution, and referral capabilities across the internet and smartphones to improve people’s health, well-being, and access to mental health services by empowering care providers with more information.

Results/Outcomes

In our first project researching undergraduate students, we observed a statistically significant decrease in self-reported levels of depression at the p = .01 level and anxiety at the .001 level for Marbles users. Study 1 is designed to control for environmental factors to see if those findings can be confirmed. In our first pilot with a employer, we implemented Marbles during their busiest season and noticed a 25% increase in Reflection scores 1.5 months after implementation. We have also witnessed over 50 conversations between individuals where one has explicitly expressed suicidal intent and have had reinforcing discussions to keep them from attempting suicide. Marbles is currently being used as an outlet for the unheard voices that are disenfranchised by the school, mental health, and/or social services. Scrolling through the conversations on Marbles demonstrates the humanity, care, and genuine conversation that still can exist between people.