Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Medicine Stats The Military United States Science

Machine Learning Used To Predict Military Suicides 74

Posted by timothy from the sobering-statistics dept.
HughPickens.com (3830033) writes David Wagner writes that a predictive computer model using machine learning methods is helping to identify soldiers in the United States Army most likely to commit suicide. Computers combed through data on more than 40,000 soldiers who'd been hospitalized for mental health problems looking at 421 variables on each soldier drawn from 38 military data systems. Using a method known as "machine learning," the researchers identified roughly two dozen factors that are most important in predicting soldiers most likely to commit suicide. The soldiers most likely to take their own lives were men with past suicidal behavior and a history of psychiatric disorders and criminal offenses, including weapons possession and verbal assaults. Soldiers with hearing loss also faced heightened risk — a strong indicator that they had suffered a head injury. So did enlisting in the Army after age 27, most likely because those soldiers had already experienced trouble finding their way in life. "There's this group that comes to the Army later in life — they're smart, they have skills, they tend not to be married and they have no career or have left a career to join," Dr. Kessler said. "We don't know why they should be at higher risk, but they appear to be."

Murray Stein, co-author of the new study, found that among soldiers recently discharged from psychiatric hospitals, more than half of suicides were committed by just five percent of patients. "The most impressive thing is that they identified this high-risk group in the hospital, and by just focusing on one in 20 of them, you're really dramatically improving your ability to predict," says Dr. Mark Olfson, a professor of psychiatry at Columbia University who was not involved in the study. "Clinicians don't do a very good job predicting suicide risk, even though we think we do."
This discussion has been archived. No new comments can be posted.

Machine Learning Used To Predict Military Suicides More

Machine Learning Used To Predict Military Suicides

Comments Filter:

  • "...more than half of suicides were committed by just five percent of patients."

    The proportion that weren't committed by patients are referred to as "homicides".

    • Re: (Score:1)

      by Anonymous Coward

      Are you retarded? ->Take 5 of the most suicidal veterans from a random sampling of 100 patients. If 10 patients kill themselves, 5 of them will be this group of suicide magnets.

      Back on topic:
      I'm not surprised by their findings since I meet almost every one of their selection criteria(and I am once again battling major depression). Sleep schedule was back on track but depression was over-whelming... So I took my antidepressants and now my depression is manageable and my circadian rhythm just rotated by 3

    • Re: (Score:2)

      by Z00L00K ( 682162 )

      If it's possible to find persons at risk and counter their behavior it's good. Most interesting are the high-skilled persons that joins late. It's not easy to put a diagnosis straight on each case, and if someone is depressed the cause may vary, ranging from bipolar disorder to post traumatic stress reaction and adrenaline junkies.

      Sudden relief from stress to a mode of passivity can be highly problematic for some persons.

  • 5% of patients? (Score:1)

    by Anonymous Coward

    If "more than half of suicides were committed by just five percent of patients", then all the suicides were committed by just ten percent of patients. Assuming at most one suicide per patient.

    • Re:5% of patients? (Score:5, Informative)

      by ShanghaiBill ( 739463 ) on Sunday November 16, 2014 @08:51AM (#48396255)

      If "more than half of suicides were committed by just five percent of patients", then all the suicides were committed by just ten percent of patients.

      No, that is not what they are saying. What they are saying is that if they screen for suicide likelihood, they can identify a sub-group consisting of 5% of the patients that commits 50% of the suicides. People in this subgroup are about 20 times more likely to commit suicide than the average patient. So it may make sense to focus intervention therapy on that subgroup.

  • Why doesn't machine learning predict suicides any longer?

  • The New Magic (Score:4, Insightful)

    by lorinc ( 2470890 ) on Sunday November 16, 2014 @08:57AM (#48396275) Homepage Journal

    Stop speaking of machine learning as if it's a new kind of black magic. I know it sounds better than "using a mathematical algorithm" or "performed statistical analysis", but to me it sounds as ridiculous as the "quantum whatever" of the 90's. Seriously, ML is being hyped beyond reasonable.

  • I don't think the state of AI will be anywhere near human thought for a long time. But dealing with huge sets of data containing many, many variables is a reasonable application for "Machine Learning". Although I kind of wonder why a conventional statistical analysis wouldn't have revealed the major factors [nytimes.com]. Anyone who has served in the military would know that an older, more intelligent enlistee probably has other issues (else why are they where they are?).

    Some of the factors associated with higher risk were expected, like previous suicide attempts, a history of using weapons and symptoms of severe traumatic brain injury, such as hearing loss. Others were less so, like a higher I.Q., and being older than 26 at enlistment.

  • When I look at the world today . . . I see Obama threatening to use dictatorial powers . . . which will cause the Republicans in Congress to shut the government down . . . while ISIS is making a mighty romp . . . and no Muslim countries seemed to be concerned about that . . . Putin, is, while just being Putin.

    And, and and more and. I'm trying hard do decide why I am not committing suicide myself.

    Oh, I need to teach a class at TU Darmstadt in a couple of weeks, and need to create a presentation. Never min

  • If someone's been searching for PTSD [wikipedia.org], Major Depressive Disorder [wikipedia.org] and Suicidal ideation [wikipedia.org], maybe they're at risk?

    People at risk tend to isolate themselves during bouts of depression (not just "the blues"), so they turn to the internet.

  • non-issue? (Score:3)

    by rubycodez ( 864176 ) on Sunday November 16, 2014 @11:03AM (#48396669)

    Once corrected for age demographics (which people tallying raw numbers usually forget to do), the suicide rate in US military is lower than civilian population. Rather impressive for organization whose purpose is to kill, maim and blow up shit.

    • Once corrected for age demographics (which people tallying raw numbers usually forget to do), the suicide rate in US military is lower than civilian population.

      You didn't actually read the articles [latimes.com], did you? That was only true up to the last decade.

      Traditionally, the Army has been psychologically healthier than the rest of society because of screening, fitness standards and access to health care. Soldiers committed suicide at about half the rate of civilians with similar demographics. But starting in 2004, the Army rate has essentially caught up.

      • No, you're citing the incorrect statistic of the article.

        http://www.nytimes.com/2013/05... [nytimes.com]

        • From your link:

          an analysis of Pentagon data shows that the Department of Defense uses numbers that may underestimate its suicide rate. A different methodology, like one employed by the Centers for Disease Control and Prevention, would result in a military rate equivalent to or above the comparable civilian rate, experts say.

          Bob Anderson, head statistician for mortality statistics at the C.D.C., said the Pentagon’s approach resulted in a suicide rate that “will be lower than it should be.”

          “It will underestimate the mortality rate,” he said

          ... and ...

          But over the years, senior military officials have regularly used suicide rates in testimony before Congress, medical journals and other public pronouncements to underscore the relative resiliency of the military population.

          A result may have been to play down the problem, some experts say.

          Dr. Elspeth Cameron Ritchie, a psychiatrist and retired Army colonel who was once a senior Pentagon adviser on mental health issues, said that when she raised concerns about suicide in 2005, senior military officers did not consider the problem significant partly because the suicide rate seemed lower than for civilians.

          Dr. Ritchie said she believed that the military is moving more aggressively now to stem the problem. But the prevailing attitude at the time, she said, “was kind of like, ‘This just isn’t a big problem.’ ”

          “They really did not focus on suicide,” she said.

          So, still the same.

    • Once corrected for age demographics (which people tallying raw numbers usually forget to do), the suicide rate in US military is lower than civilian population.

      There's probably all manner of "corrections" you can do to make yourself feel all superior and to denigrate the folks actually doing the work.

      Rather impressive for organization whose purpose is to kill, maim and blow up shit.

      Why do you think that's impressive? You think everyone who served is nothing but a suicidal maniac who wants nothing

      • Once corrected for age demographics (which people tallying raw numbers usually forget to do), the suicide rate in US military is lower than civilian population.

        There's probably all manner of "corrections" you can do to make yourself feel all superior and to denigrate the folks actually doing the work.

        Huh? How is it "denigrating" the military to note that they have a lower suicide rate than the average population? I would have thought that having a lower suicide rate would be a GOOD thing.

        (By the way, I'm ignoring the fact noted elsewhere in this thread that TFA basically says the military is now approaching the average suicide rate of the general population, even though it had been less in the past. Point is -- I can't possibly see how it's an insult to note that a group has a lower suicide rate th

    • Your grade school level of understanding as to the purpose of a military is impressive

    • Re: (Score:2)

      by AHuxley ( 892839 )
      It depends how the US military is counted. A huge number of civilian like staff, air force, army... could be used to hide deep counts.
      So what can be counted on? The Air Force has very expensive equipment to look after. So great care is taken to only allow selected, well tested people near that equipment.
      That would sort out most of the human issues long term and offer amazing statistics.
      Other areas of the US mil may also like to have well tested people near some of the more interesting systems. Grea

  • Instead of using computers to figure out who's most likely to take their own lives, how about we use machine learning to figure out why these people are killing themselves, and fix the actual problem?

    • So, you don't think that finding common characteristics that are good predictors of behavior provides information that can be used in researching the root cause?

    • Re: (Score:2)

      by Livius ( 318358 )

      Maybe figuring out why will be easier if we have an idea about who.

  • Why should it be illegal to commit suicide?

    I don't feel that if someone wishes to take their life we have a right to stop them. It is their existence, their experience, their decision, their right. We should not encourage it, but if we want to reduce the number of people who do it, then help them with their problems, don't go hunting them down, and locking them up in white room - mattress walled asylums in straight jackets, doped up on who knows what.

    It's dangerous to allow the government the ability

  • Or alcohol.

    Thank you, NRA and Congress, for giving our brave service men and women the freedom and personal responsibility to kill themselves when things get too tough.

    http://touch.latimes.com/#sect... [latimes.com]
    Programs to prevent psychological problems in troops questioned
    By Alan Zarembo
    February 20, 2014, 7:34 p.m.

    Many federal programs aimed at preventing psychological problems in military service members and their families have not been evaluated correctly to determine whether they are working and are not supported

Slashdot Top Deals

Single tasking: Just Say No.

Close