What's the medical breakthrough that could save the most lives in the U.S. over the next 10 years? In the 2020s, medical research will likely inch forward when it comes to major killers like heart disease and cancer. But the biggest potential to save lives could lie in learning to prevent suicide.
The rates of reported suicides have been creeping up over the last two decades. Even more disturbingly, the Centers for Disease Control and Prevention reports that the suicide death rate for teens increased 56% between 2007 and 2017. Rising suicide rates might be a result of many things — rising levels of despair, the opioid epidemic, greater access to guns, even the proliferation of internet groups that offer people advice on how to kill themselves. It could also be that more people are reporting suicides instead of concealing such deaths as accidents.
It's a surprisingly common form of death — more prevalent than homicide or automobile accidents. Unlike cancer and heart disease, which are leading causes of death among the old, suicide robs people of decades of life. According to CDC statistics, it is the second most prevalent cause of death, after accidents, for people between 10 and 34 and fourth for people between 34 and 54.
Because it hasn't been all that thoroughly studied as a medical problem, there's room to cut down on that death toll even without any remarkable technological breakthrough. A streamlined three-digit suicide hotline number, approved last month by FCC, could become one of the great public health measures of the century. Further out on the frontier, researchers are having some success using artificial intelligence to identify suicidal people — those whose lives might be saved by talk therapy or drugs.
John Pestian, director of the computational medicine center at Cincinnati Children's Hospital, explains that there are different kinds of suicides. Some are driven primarily by chronic mental illness, while others are more impulsive. Those with chronic mental illness may make repeated attempts. People have a powerful instinct to live, he says, and for their psychological pain to override this, it must be incredibly intense. He's hoping to help such people through pharmacogenomics — finding drugs that will ease their chronic emotional pain.
The more impulsive cases are simpler to prevent — think of the teenager whose boyfriend or girlfriend just left, or a Wall Street trader who lost all his money, he says. If someone is going through an acute crisis and wants to jump out a window, the right words spoken at the right time might be the only treatment needed to save a life.
Renowned suicide researcher Edwin Schneidman writes in "The Suicidal Mind" that therapists can help people in this state by getting them to consider alternatives besides killing themselves — helping them see that they have choices. He describes how he helped a suicidal college student who felt hopeless after she found out she was pregnant. He got her to consider which of her options was least terrible, and she recovered.
In the memorable 2003 New Yorker story "Jumpers," Tad Friend describes conversations with several people who survived after jumping off the Golden Gate Bridge. They told him that they recognized their mistake before they hit the water: "I instantly realized that everything in my life that I'd thought was unfixable was totally fixable — except for having just jumped," one said. Another left a note saying 'I'm going to walk to the bridge. If one person smiles at me on the way, I will not jump.' "
To learn more about the reasons people decide to take their own lives, Pestian and other researchers are amassing troves of data. Right now, he says, he has the biggest collection of suicide notes in the country, as well as samples of speech and body language from suicidal patients. There are clues in these that therapists can look for — and patterns that algorithms can use to identify those most at risk.
He says suicide hotlines are crucial, and he approves of the idea of amending the current one, created in 2005, from the usual 10 digits to just three: 988. Even so, it's not quite as simple as it sounds, he says: There will also have to be the right kinds of resources at the other end of the line.
According to one news story, an FCC committee estimated that those resources would cost $570 million in the first year and $175 million the next. This is pocket change compared to, say, routine mammography, which costs Americans billions and the value of which has been called into question.
As for AI, using algorithms to predict anyone's behavior can sound scary — especially the use of systems that attempt to label Facebook users as suicidal from their posts, or others that gather data from users' smartphones. But Pestian convinced me that there's potential for AI to do much good in the area of suicide, as long as it's only used to support human decisionmaking, and humans don't delegate the decisionmaking to machines.
He has developed algorithms that work with what he calls sentiment data — acoustic, visual or language patterns that differ between the suicidal and non-suicidal. In a paper published in 2016 in the journal "Suicide and Life-Threatening Behavior," he applied an algorithm to interviews with a sample of 379 people, some known to be suicidal, some diagnosed with mental illness but not suicidal, and a healthy control group. The algorithm used speech, facial expressions and body language to identify the suicidal group with 85% accuracy.
That's not perfect, but it's better than doctors can do. Pestian has recently gotten a contract with Oak Ridge National Laboratory to apply AI to the rampant problem of suicide among veterans. According to a report from the U.S. Department of Veterans Affairs, more than 6,000 veterans die by suicide every year — a rate 50% higher than that in the general adult population.
We may not know the reason for the rising suicide rate, but we do know it is killing too many people — and that those deaths ought to be preventable. While science has a pretty detailed understanding of cancer and heart problems, suicide was studied by relatively few researchers until recently. It's good news that we're finally starting to learn more.
