Two months ago, Connecticut Senator Christopher Murphy told Secretary Eric Shinseki, chief of the Department of Veterans Affairs, that “every veteran suicide represents a collective failure of our nation to properly care for those who have borne the battle.” He concluded: “We can do better.”
The statistics are alarming: suicide rates among U.S. veterans are almost double those of the general U.S. adult population. Twenty-two veterans commit suicide every day. Time has called it a “problem of epidemic proportions.” Reducing the incidence of suicide among U.S. veterans has proven to be a complex and challenging battle: no initiative or program has worked to reverse this trend.
Now it’s technology’s turn to try.
Last week, an initiative called The Durkheim Project announced the availability of an opt-in veteran’s database that will analyze veterans’ social-media posts to identify suicide risk factors. The project is named in honor of Emile Durkheim, whose book, Suicide – published in 1897 — consisted of surveys and suicide statistical analysis that classified victims into categories, and found links between their daily actions and the likelihood of committing suicide.
Phase One of The Durkheim Project, conducted with The Geisel School of Medicine at Dartmouth and the U.S. Department of Veterans Affairs, and completed in 2013, the predictive analytics and text-mining technology methods used to power the database were statistically significant in predicting suicide-risk factors with 65% accuracy.
Phase Two of The Durkheim Project, now underway, includes scaling the database to more than 100,000 veterans and encouraging all veterans – and their friends and families — to see the value in having social-media behaviors observed for risk factors. At this time, the project is authorized as observation only and non-interventional, with the hope of eventually providing a real-time triage of interventional actions (Phase Three). The database is entirely opt-in and complies with federal health-privacy standards.
The Durkheim Project has been a labor of love for all involved, especially Chris Poulin, a predictive-analytics guru and founder of the company Patterns and Predictions. Poulin began working with Dartmouth researchers to address the veteran suicide problem in 2010. He is steadfast in his belief that the next generation of predictive analytics software tools “gives scientific and clinical investigators new hope and resources for solving even the most intractable problems.”
In 2011, Poulin began enlisting the help of other technology companies, including my company, Attivio, to help join information together at a massive scale, and make it highly accessible so it can be analyzed in real time. We were on board immediately. After securing a $1.7 million contract from the Defense Advanced Research Projects Agency, the project was tested, validated and is now ready for wide-scale deployment.
In so many ways, The Durkheim Project is an example of Big Data done right. It’s an inspiring example of technology companies coming together to solve a social and government problem that has thus far eluded all other solutions. With so much recent negative attention given to privacy and the obtaining of personal information, The Durkheim Project is completely transparent, opt-in/opt-out, and committed to following all medical protocols for privacy.
The hope is that with awareness and success at helping identify veterans at risk, more and more people will feel encouraged to participate.
The solution is as much about helping veterans, as it is about putting veterans in the best possible position to be helped.
Ali Riaz is chief executive officer at Attivio.