March 3, 2019
Last weekend I traveled to the lovely city of New Haven for a mental health hackathon hosted by Hack Mental Health Care. I was very pleasantly surprised by the experience, which proved interesting, fun and invigorating (with a few healthy dashes of disappointment and horror).
I was mostly expecting undergraduate participants with ideas for mood tracking apps, but the event drew over 200 people, and was quite diverse. In addition to programmers, designers, product folks and business people showed up. Genders were closely balanced and minorities were represented. Crucially, over 30% of the participants had clinical or lived experience. The event also featured a therapy dog, yoga sessions and a guided mediation. Peer voices and ethics were featured in some of the talks, although due to time constraints, project design was complete and implementation was already underway. And, kudos on the Code of Conduct… next year I would also love to see consent-based photography and sponsored childcare.
The organizers worked hard to prompt the participants in advance with these challenges:
- Challenge 1: How can we help to reduce the rates of suicide?
- Challenge 2: How might technology increase access or improve treatment for people with substance use disorders?
- Challenge 3: How might we use data to drive meaningful insights for patients and clinicians that improve mental health?
The challenges were a bit solutionist in their framing for my taste, but were generally crisply formulated and well researched.
A recap of the event is posted on medium: HackMentalHealth Yale’s Collegiate Mental Health Hackathon Recap, along with links to the winning projects.
I floated across a few teams including with a group who had the insight that a person’s musical listening preferences might reveal something important about their mood – Team Moodify was born. The theme of risk assessment prevailed across many of the projects I encountered. To speculate, risk assessment probably feels like a more tractable problem than crafting an intervention, and one that is amenable to statistical analysis.
Our team had a great conversation about what the team might do with a risk assessment assuming it could be computed – should Spotify summon law enforcement if it detects you are at risk? Should they contact your friends? Alert you to your own mood? Would a user be aware of this monitoring, and could they configure it according to their own preferences?
Team Moodify valiantly worked all weekend and came up with a prototype that allows users to log in to their Spotify account, and create a mood boosting playlist based on their own listening history (Spotify itself has a notion of valence and energy for a song that was used to find similar songs).
We didn’t have time for it over the weekend, but someday I would love to add a social component and create a mechanism to solicit a mood boosting mix tape from your friends 😉
For more about Moodify, see https://devpost.com/software/spotify-mood-detector-nqcdlj/ and https://github.com/jacsonding/Team_Banana)
Across projects, common themes emerged. Teams identified the need for better directory and referral services (an area where OpenReferral could really help), and gravitated towards monitoring and risk assessment. The studio-style format allowed for wide variation in these designs based on small differences in inflection. So, for example, Flip, a winning project about a Chrome extension designed to flag upsetting content, could benefit greatly by putting more agency back into the user’s control – allowing the user to configure and customize settings and thresholds, as well as actions to take when upsetting content is identified.
At the next Hack Mental Health session I would love to see the community continue to grow. I think the projects would benefit from a more structured design phase – intentional, community driven, persona driven, etc (BigApps does a great job with this now). I would also love for more open source and open standards to be represented and used – new coders could join an existing community and contribute towards a longer term goal. I would also enjoy more unconference style collaborations – more ice-breaking and exchange during team formation, lightning talks given by participants, self-organized birds of a feather conversations. Finally, all the teams should have a chance to present their work to each other – I fear many participants left the hackathon without appreciating much of the worth done by other teams.
Thanks to all the organizers and sponsors for pulling this together. Next year in New York City?