On June 16, Mission Daybreak brought together suicide prevention, data, and privacy experts to discuss how to use technology and data to support mental health and suicide prevention initiatives. Speakers included:
- Dr. Munmun De Choudhury, Associate Professor, School of Interactive Computing, Georgia Tech
- Dr. Richard McKeon, Chief, Substance Abuse and Mental Health Services Administration (SAMHSA)
- Dr. Amanda Purnell, Director of Data and Analytics Innovation, VHA Innovation Ecosystem
Following the panel discussion, the speakers answered questions from attendees. Watch the session recording and read on for highlights from the virtual event. Also, see the presentation slides; the challenge FAQ will be updated later this week with questions and answers from the virtual session.
What are new insights from data innovations in suicide prevention?
“What is very promising are the kind of signals that we can get when we are combining disparate data sources together. We looked at it in a different way, in a public health style of work where we combined social media with health services data such as visits to the emergency department, calls placed on the National Suicide Prevention Lifeline… When we look at all of these data sources together, it allows us to reach maybe those pockets of the population that otherwise we are not able to when we look at a single source of data alone.”
— Dr. Munmun De Choudhury
“We obviously learned early on about the preference of young people for chat and text. It was a surprise to realize that their likelihood of currently being suicidal at the time that they were trying to access this service was so much greater than people calling on the phone. Regarding data, some of the work that is done with the PHQ-9 showed that the answer to the question on that screen about whether you’re currently thinking about harming yourself was associated with later suicide attempts.”
— Dr. Richard McKeon
What is synthetic data and what is the VA’s capability in creating it?
“Synthetic data is derived using machine learning, so it’s a basic computational algorithm that’s run on real data that kind of derives the relationships between the variables within the data and generates an output that does not contain any real data or real identifiable data, meaning people’s privacy is preserved. You’re not sharing patient X’s or person X’s information, while still sharing meaningful information to allow for early stage development, hypothesis testing, and an ability to really see if there’s something meaningful there.”
— Dr. Amanda Purnell
Enter Mission Daybreak by July 8
Solvers should submit detailed concept papers to Phase 1 by 4:59 p.m. EDT on July 8, 2022. Some innovators will have market-ready solutions, while others will have early concepts — in Phase 1 of Mission Daybreak, the quality of the idea and potential for viability is more critical than present-day maturity.
30 finalists will be selected to participate in a virtual accelerator, which will provide resources — from exclusive data sets to networking and mentorship — to help solvers prepare their final submissions.
To receive all challenge updates, subscribe to the Mission Daybreak newsletter. All challenge questions should be sent to hello@missiondaybreak.net.