|
Cincinnati Enquirer
Detecting the
language of suicide: There's an app for that
Chris Graves
November 21, 2016
Any parent of a teen knows all too well the ups and downs of those
tumultuous years.
At best, we feel utterly infuriated. At worst, we feel completely
helpless and shaken to our core — especially if our kids harm, or
threaten to kill, themselves.
As parents, we are conditioned to look and listen for their cries for
help.
Turns out we may miss a lot.
Suicide still remains the third leading cause of death of kids age 10
to 14, and is second among people aged 15 to 34 years old. Every 14
minutes someone commits suicide.
'Amazing, really amazing'
Enter John Pestian and his team of researchers with big brains and even
bigger hearts.
Pestian, a professor in the divisions of Biomedical Informatics and
psychiatry at Cincinnati Children’s Hospital Medical Center, has spent
nearly a decade immersed in the language of suicide in an effort to try
and keep kids alive.
What he has found is there are indicators in spoken language that can
help school counselors and medical professionals identify when kids are
at risk for suicide or when they may be suffering from a mental illness
but are not at risk for suicide.
Essentially, there is science behind the words we use, in additional to
non-verbal clues like pauses, tone and pitch, that are what Pestian
calls "thought markers," according to his most recent work published
earlier this month by the official journal of the American Association
of Suicidology.
Using that information, technology wizards built algorithms to power an
app being studied in eight Cincinnati-area schools this year. The app,
which looks incredibly simple, is called SAM and stands for Spreading
Activation Mobile. SAM records a teen's conversation during counseling
sessions. It uses Pestian's technology to measure the words the teens
use to determine if the language is similar to someone who is at risk
for suicide. The app also detects when a teen's language is just the
speech of a typical teen: It might be full of angst, but it's not akin
to suicidal speech.
"It's amazing, it's really amazing," Pestian says.
His work has been repeatedly called groundbreaking.
Machine can't replace humans
In the recently published study, conducted from October 2013 to March
2015 in three different emergency departments, suicidal subjects were
found to laugh less and sigh more. They also expressed more anger and
exhibited less hope than those who were deemed to suffer from mental
illness or were neither.
The study showed that the computer technology, known as machine
learning, is up to 93 percent accurate in correctly classifying a
suicidal person and 85 percent accurate in identifying a person who is
suicidal, has a mental illness but is not suicidal, or neither.
An ongoing experimental phase of his research includes video recording
the conversations to study facial movements to see if there are further
clues embedded in the face. Early research seems to indicate, for
example, that non-suicidal teens showed more teeth in a 90-minute
conversation than suicidal teens. And suicidal teens gazed down six
times during a 90-minute interview, while their non-suicidal
counterparts looked down just twice, he said.
"These are things you just don't pick up in conversation,'' he says.
Pestian is quick to point out, however, that artificial intelligence
can only go so far. It is not - and never will be, nor should be - a
replacement for clinicians, therapists or others who work with teens.
"The technology is not going to stop the suicide, the technology can
only say: "We have an issue over here,'' says Pestian, whose work has
won patents and has been published around the world. "Then we
have to intervene and get a path to get to care.
"If it's just a machine it is useless,'' he says.
Suicide every 14 minutes
Pestian first started his work by studying language written in notes of
1,319 people shortly before they committed suicide. Then they moved
onto interviewing 379 teenage patients at three hospitals, including
Cincinnati Children's. Some of the teenage patients were in the
emergency rooms because they were exhibiting suicidal thoughts or
tendencies, while others went there for other conditions.
He still shakes his head at how prevalent suicide is among teens.
Based on a 2013 Centers for Disease Control report:
17 percent of high-school age kids seriously considered attempting
suicide in the previous 12 months;
13.6 percent had made a plan about how they would attempt suicide and 8
percent of students attempted suicide one or more times. Suicide
results in an estimated $51 billion in combined medical and work loss
costs.
"The purpose of SAM is to find the kids earlier, because suicide is an
avoidable death; it's preventable death. If we just get them early and
if we know what to look for; if the parents know what to do," Pestian
says. "It's miserable on the family. It hurts just to call."
"It's extremely easy to use. It's super simple,'' Crotte says, adding
that all of his clients know he is using it and they are involved in
the study. "It's really a minimal thing we are incorporating into our
work."
If you or someone you know needs help: National Suicide Prevention
Lifeline, 800-273-8255
Read this and other articles at the Cincinnati Enquirer
|
|
|
|