Can technology help predict who will attempt suicide? – partial transcript


By Melanie Saltzman and Saskia de Melker

Tens of millions of Americans struggle with mental illnesses, and knowing who, among these individuals, is at risk for suicide is one of the biggest challenges psychologists and psychiatrists face.

There are a growing list of groups considered at high risk for suicide — for example, people with medical conditions, mental health issues, or military veterans — but those groups are too large to effectively narrow down. And with few diagnostic tools beyond their clinical judgement, doctors mostly rely on individuals verbally expressing their thoughts and plans to kill themselves.

But now, researchers are developing technological innovations to go far beyond talk therapy and more accurately diagnose imminent suicide risk.

“Traditionally, if someone was suicidal out in the world, we would wait for them to come into our office or into our lab and tell us they’re suicidal and we’d ask them about what their experience was like,” suicide researcher Dr. Matthew Nock said. “But now with new technologies, we can take the lab and bring it to the person.”

At Massachusetts General Hospital, Nock is using computer and smartphone-based tests that unveil a person’s subconscious associations with suicide. He’s also using bracelets that track an individual’s biological signals so he can search for physical and behavioral patterns among those who are suicidal. And at Bradley Hospital in Providence, Rhode Island, Dr. Daniel Dickstein has developed video games that tap into the inner workings of the brains of bipolar children.

Scientists say the results from all these tools are promising, but they need more funding and support to continue developing these innovations.

Read the full transcript below.

ALISON STEWART: In New York City this summer, more than two thousand people gathered for the biannual “out of the darkness” overnight walk. It covers 16 miles from dusk to dawn, in memory of those who have died from suicide. Rosemary Fuss traveled from Boston. It’s the tenth time she’s walked for her son, Tommy.

ROSEMARY FUSS: We are part of a family. We are survivors. And we are committed to saving other precious lives. We can’t bring our loved ones back, but we can certainly do something to help others.

ALISON STEWART: The event raises awareness and money to research suicide prevention. Suicide deaths in the United States rose from 29,000 people in 1999, to nearly 43,000 in 2014. And the overall suicide rate between those years is up 24 percent. Suicide is now the nation’s 10th leading cause of the death, and the second leading cause for young people between 10 and 24 years old. Like Tommy Fuss.

ROSEMARY FUSS: He was a 17-year-old kid who presented himself in a confident, independent way. A kid who had everything to live for.

ALISON STEWART: At what point in this process, in your journey, did you realize that Tommy was struggling?

ROSEMARY FUSS: It’s hard to know what is a struggle that is associated with a mental illness and what’s the normal adolescent finding his own self. We did receive a call from someone at the school who said, “You know, maybe Tommy should see someone, talk to someone, I think Tommy is depressed.” My husband and my first reaction was, Tommy? Depressed? You’ve got to be kidding.

ALISON STEWART: Tommy did see a therapist, but no serious issues were revealed. A couple of months later, rosemary found her son’s suicide plan. It was foiled and tommy spent one night in a hospital.

ROSEMARY FUSS: The psychiatrist and others thought get him back into his functioning environment; we’ll treat him as an outpatient. But Tommy did not share his inner self.

ALISON STEWART: Two months later, Tommy turned on his family’s car, left the engine running inside the garage, and killed himself by carbon monoxide poisoning.

JEFF HUFFMAN: We don’t have any way particularly of knowing what a person’s thoughts or planned actions are going to be.

ALISON STEWART: Doctor Jeff Huffman is the director of inpatient psychiatry at massachusetts general hospital. He was not involved in Tommy’s care.

JEFF HUFFMAN: During our training, a lot of the focus on assessing people’s suicide was asking people directly but also kind of using your quote unquote clinical judgement. And there was a sense that as you grew as a psychiatrist, you would just know that somebody is at risk or not at risk.

ALISON STEWART: But studies, like this one from the Journal of Psychological Science, show clinicians have about a 50/50 chance – no better than a coin toss – of predicting who will attempt suicide.

There are certain groups with higher risk for suicide: like people with medical conditions; mental and substance use disorders; or military veterans. But it’s difficult to narrow down who – of the millions of people in those high risk groups – will try to kill themselves. And there are people who don’t fall into those groups who may also attempt suicide.

MATTHEW NOCK: So we have this growing list of risk factors. Where we’re limited is in our ability to put them together in a way that can tell us which people are at greatest risk.

ALISON STEWART: Doctor Matthew Nock is a leading suicide researcher and professor at Harvard University. He is developing new diagnostic tools — using technology and advances in science.

ALISON STEWART: What does technology offer in this field that traditional methods of diagnosis doesn’t allow for?

MATTHEW NOCK: Traditionally, if someone was suicidal out in the world, we would wait for them to come into our office or into our lab and tell us their suicidal and we’d ask them about what their experience was like. But now with new technologies, we can take the lab and bring it to the person.

ALISON STEWART: In one study at Mass General, Nock is using real-time monitoring to try and predict suicidal thoughts and behaviors. He has patients respond throughout the day to questions on a smartphone. Like how nervous or abandoned have you felt in the past 24 hours? Or how many hours did you sleep last night?

Patients also wear bracelets that collect “biomarkers” — their physical information — like sweat, skin temperature, and heart rate. When a patient feels suicidal, he or she presses a button on the bracelet.

MATTHEW NOCK: We can then look back – are there certain types of experiences or patterns or signatures that tell us when those button presses are going to occur. So can we get better at identifying when a person is at risk.

ALISON STEWART: Sam Bamford has had suicidal thoughts since high school. He’s now 34 years old and has Huntington’s, the degenerative brain disease with no cure.

ALISON STEWART: He’s says he’s been hospitalized twice for suicidal thoughts. We met him after he came to Mass General, following a suicide attempt that, he says, came out of nowhere.

SAM BAMFORD: The first hospitalization was all nonstop thoughts about it. Months of it. And then this time around, everything was going great. I had an awesome night, hung out with some really great friends that are super supportive. It was just really strange.

ALISON STEWART: So this crisis was one of impulse.

SAM BAMFORD: Yeah, it came out of the blue.

ALISON STEWART: After arriving in the ER, Bamford took what’s called an implicit association test, or iat. Iats are often used to examine racial or gender bias. Nock has designed this IAT to analyze subconscious thoughts associated with suicide. To see how it works, I took the fifteen minute test.

I was instructed to read a word flashing on the screen, and place it either in the me or not me column. Some words were suicide related—such as death or funeral. Some words were life affirming —such as breathe or living.To get the subjects unfiltered thoughts, there’s only a fraction of a second to respond.

ALISON STEWART: it was tough. It was tough, I couldn’t put myself and death next to each other.

MATTHEW NOCK: Exactly the idea. And what we found is that people who are suicidal are faster than those who are not suicidal pairing death and me on the same side, cause they identify with death or suicide. And so we want to compare one set of reaction times to the other in milliseconds.

ALISON STEWART: Nock is still working on making the test more accurate, but says he’s already seen strong results from the thousands of people who have taken it.

MATTHEW NOCK: It predicts better than a person’s own report of whether they’re going to make a suicide attempt. It predicts better than clinicians’ reports.

ALISON STEWART: Nock says most patients are open to interacting with technology-based tools. And some even say it’s preferable.

SAM BAMFORD: There’s actually parts of me that wish some of this stuff was already shared with my doctors, because I’d just have an easier time answering the questions with the phone than I would with a room full of people.

MATTHEW NOCK: Sure. Can be easier that way, less intimidating.

ALISON STEWART: And Jeff Huffman says these new tools can also help determine who needs to be in this psychiatric unit, which like so many across the country, has a small number of beds compared to the need.

JEFF HUFFMAN: We all sit and think and worry about our next patient we discharge attempting suicide or God forbid, completing suicide. If we can develop tools to make us more confident about patients who are safe to leave. We desperately need those, we desperately want those.

ALISON STEWART: These tools were among the techniques discussed at a suicide conference hosted by the National Institutes of Mental Health in June. All of these researchers are looking for innovative ways to more precisely predict suicide risk, and to diagnose and treat mental health disorders.

Doctor Daniel Dickstein is a child psychiatrist and Associate Director of Research at Bradley Hospital in Providence, Rhode Island. He studies bipolar disorder, a mental illness characterized by mood swings ranging from depressive lows to manic highs.

DANIEL DICKSTEIN: Bipolar disorder, of all psychiatric disorders, involve the highest rates of suicide. So about 50 percent of kids who have bipolar disorder think about killing themselves at least once, and about 30 percent of kids with bipolar disorder actually try to kill themselves, again, at least once.

ALISON STEWART: 15-year-old Mason is bipolar, and is a participant in Dickstein’s research study. His mother, Jessica, felt more comfortable talking to us about her son’s illness with her face obscured.

ALISON STEWART: When you heard, “Jessica, your son has bipolar disorder,” what did you think?

JESSICA: It’s just really hard to hear. You want an easy life for your kid. You don’t want anything to make it hard for him. And I knew nothing about bipolar. I knew nothing about what he was diagnosed with. So it was scary.

ALISON STEWART: Did Mason ever pose harm to himself?

JESSICA: When he was about 9. He had about a two hour long rage fit that we couldn’t control anymore. And he was in the kitchen by the knives, and he just said he was going to kill himself, and he kept saying it.

ALISON STEWART: You took him to the hospital.

JESSICA: Yes, we took him to the hospital. We were there overnight. And then at 3 in the morning, they woke him up, did a psych eval and sent us home an hour later.

ALISON STEWART: Dickstein says children with bipolar disorder have certain irregularities in brain activity. One is an impaired ability to quickly adapt to changing situations. So he’s designed a space-themed video game that, he hypothesizes, will “retrain the brain” and improve those skills.

DANIEL DICKSTEIN: Periodically throughout the game, they have to adapt their behavior to figure out a decision point. And they have to learn from their prior decisions as they’re flying through. And that’s sort of the heart of the game. Their initial choice once they figure out the right move may work for a while, but after a while it changes without warning.

ALISON STEWART: Participants in the study play the game twice a week for eight weeks. Then Dickstein measures their brain activity with an MRI while they play a version of the game.

If the scan looks more like that of a healthy child after the eight weeks, the brain retraining may be working. Dickstein says early results are promising but much more testing is needed. He says, if it works, it could be scaled up quickly and help thousands of children.

ALISON STEWART: Why hasn’t this kind of research been done before?

DANIEL DICKSTEIN: It’s harder and harder to get funding.

ALISON STEWART: In the past several years, large amounts of federal funding have been put toward researching causes of death like HIV, heart disease and various cancers. Mortality rates in all those areas have dramatically decreased.

Even as suicide deaths continue to rise, suicide research gets significantly less funding.

Dickstein is among many experts who believe that, if suicide research was better funded, the epidemic would also slow. He says people need to prioritize suicide and mental health research in the same way they did for childhood cancer.

DANIEL DICKSTEIN: Whereas before up to the ‘80s, it was an absolute death sentence. Now the five-year survival for childhood leukemia is over 90, 95 percent. Patients, parents, clinicians, researchers and funders got together and said, enough of our kids dying of cancer. Every kid who has cancer is now going to be part of research. That’s really what I’m trying to do for bipolar disorder and related conditions like suicide.

JESSICA: If today Mason playing games helps a kid 10 years from now, then it was worth it, because that’s somebody we helped. When you’re looking for answers and you don’t know where to turn, you need help.

ALISON STEWART: It’s 2AM, and RoseMary Fuss and her eldest son, Danny, are still walking.

Their team is committed to helping moms like Jessica, by removing the stigma attached to suicide, and supporting research for the technological tools that, RoseMary believes, would have helped her son.

ROSEMARY FUSS: They would have given caretakers more information.The information that Tommy wasn’t willing to share with them face to face. Nothing would bring Tommy back, but we could help others. It’s a feeling of, you know, we are making a difference.