/, Health, Technology/Will machine learning stop suicide?

Will machine learning stop suicide?

By  Shane G. Owens, Ph.D., ABPP

 

We are making great strides in preventing the things that kill Americans, except in stopping suicide. Since the turn of the century, we’ve gotten worse at suicide prevention. It is the 10th leading cause of death, claiming 44,193 lives in 2015, and is the second leading cause of death for people between 15 and 24 years old. Suicides among teenage girls have doubled over the last decade, and have increased 30 percent in teen boys over that same period.

While many believe social media is partially to blame for mental health issues in kids, sites like Facebook, Instagram, and Twitter may be fertile ground for suicide prevention, and the algorithms that drive them may help us to detect risk.

Facebook recently started allowing friends to report concerning content on users’ feeds and to use Messenger to contact the Crisis Text Line and the National Suicide Prevention Lifeline. Some researchers examined language and emojis in tweets for signals of risk, and one study of veterans’ social media use found that how often a person posts about depression and alcohol use may predict suicide attempts.   Work at the Cincinnati Children’s Medical Center utilizes kids’ language, social media posts, and other behaviors to predict suicide attempts. Several school districts are test sites for this program.

Given the difficulty that mental health professionals have predicting suicide attempts, any tool that helps to identify signs of crisis are welcome. In addition, online assessment and treatment can increase access to and utilization of vital mental health services. Over half the counties in the US have no access to competent mental health professionals.

Online assessment tools and algorithms are not without their detractors. Many psychologists are skeptical of these platforms. Some believe that putting two screens and thousands of miles of fiber optics between a person and her therapist removes something crucial from the therapeutic relationship.

It is unlikely that anyone who truly needs face-to-face help will be satisfied with online treatment if effective, in-person alternatives are available.

While machines may be more accurate than clinicians in predicting suicidal behaviors, they are not perfect. There is a danger in false positives €”saying that a kid is at-risk when he isn’t. Well-intentioned schools may require assessment at a hospital before the kid is allowed to return to classes. While seeking treatment for real crises is a good idea, over-reacting to them can make the problem worse.

Whether or not machines can accurately predict risk, families and communities are essential in suicide prevention. Even if a kid presents a significant risk, maintaining close ties to the community and receiving competent psychological treatment will keep her alive, safe, and healthy. Adults who know how to spot signs of distress and how to talk to kids who in trouble are the most powerful defense  we have against suicide.

 

Author Biography:

Shane G. Owens, Ph.D., ABPP  is an authority on college mental health practice and policy, including college readiness and behavioral risk management. As a college administrator and in private practice, he works primarily with adolescents and emerging adults. He is a board-certified behavioral and cognitive psychologist.

Follow Dr. Owens on Twitter:  @drshaneowens

For more helpful information, visit:  drshaneowens.com

Shane Owens

By | 2018-05-29T15:01:16+00:00 October 20th, 2017|Communication, Health, Technology|Comments Off on Will machine learning stop suicide?

About the Author: