The unfortunate truth when it comes to suicide is that we often can’t prevent it simply because we can’t predict the future.
For the past half century, researchers have made little progress in the area of suicide prevention—but all that is starting to change.
Through the use of machine learning, a research team at Florida State University has created an algorithm that can predict with 80-90% accuracy whether someone will attempt suicide as far off as two years into the future. As close to one week before a suicide attempt, the accuracy climbs to 92%.
They achieved these staggering numbers by turning their algorithms loose on detailed medical histories of 3,200 people leading up to their suicide attempt.
For decades, researchers have been assessing individual warning signs, but through these algorithms, they’ve found that identifying at-risk patients is more so about identifying hazardous relationships between different risk factors and then seeing who experiences those same factors.
Most importantly, this research team understands that risk factors change over time and can be vastly different among other populations.
So, maybe Facebook is better suited for this?
You read that correctly: Facebook is tackling suicide prevention, too.
Currently, they’re testing AI that uses pattern recognition to detect comments that are likely to include thoughts of suicide. Given that Facebook has some of the best algorithms in the world when it comes to suggesting content users will like, I would expect their suicide prevention algorithms to get pretty accurate over time.
Facebook has also created new Messenger tools in collaboration with the Crisis Text Line, the National Eating Disorder Association, the National Suicide Prevention Lifeline and other organizations. The idea is that these tools would help at-risk users or concerned friends to easily and quickly contact knowledgeable groups over chat.
Personally, I’d like to see Facebook make a better effort at eradicating the root cause of a lot of teen suicides, which is cyberbullying—especially since their platform is one of the reasons cyberbullying exists.
Nonetheless, it’s great to see digital health innovation cleaning up the mess which tech has accidentally created.