Identifying Native American Youth at Risk of Suicide

In 2017, Senior Scientist Allison Barlow and Associate Scientist Emily Haroz explored whether a mobile health tool used by outreach workers could reduce suicide and self-harm behaviors among American Indian youth. Suicide prevention is a major priority in Native American communities. Working with the White Mountain Apache tribe in Arizona, the researchers used machine learning with community-based suicide surveillance data to better identify those most at risk. 

First, the team created an algorithm based on characteristics of over 4,000 incident reports to help community mental health workers prioritize and optimize follow-up with high risk individuals. They then embedded this algorithm in the m-health platform (mobile health app) and pilot tested it to measure its accuracy in identifying those at most imminent risk and improving care.  

Their research, funded by the Initiative, found 

• Their algorithms performed substantially better (AUC = 0.75-0.80) than a history of a previous suicide attempt, the best single predictor of suicide risk from the literature (AUC = 0.57). In addition, it may be more acceptable to staff in their day to day care of clients. 
• The use of tribal sovereignty to mandate suicide surveillance and maintain locally relevant data is helping address one of the most urgent behavioral health disparities for American Indian or Alaska Native communities. 
• The research team presented its findings at the American Association of Suicidology annual conference in April 2018, the International Academy of Suicide Research annual meeting in 2021, and received a federally funded grant to continue to expand this work from the National Institute of Mental Health.  

Read more about their work in Suicide and Life-Threatening Behavior and JMIR Public Health and Surveillance.