A tool for predicting suicide?

Suicide is both disturbing and perplexing to survivors, in part because it is so unpredictable. People who are intent on killing themselves often conceal their thoughts—or outright deny them—so family and friends are left puzzling over warning signs they might have missed.

Even experienced clinical judgment often misses the mark. As a result, suicide experts have long hoped and searched for a clear behavioral marker of suicide risk. Now they may have found one. Harvard University scientists are reporting that a tool widely used for probing unconscious thoughts might be used to spot suicidal intent—even if the suicidal mind is in denial—and offer new hope for timely intervention to keep people alive.

Psychologist Matthew Nock (working with colleagues at both Harvard and nearby Massachusetts General Hospital) decided to adapt a decade-old test called the Implicit Association Test, or IAT, to plumb for warning signs of suicide. Specifically, he wanted to see if people who are suicidal might have stronger implicit associations between themselves and death—associations that might point toward self-destructive intentions. To find out, he tested 157 people seeking treatment in a psychiatric emergency room. The patients were all emotionally distressed, but only some were in the hospital because of attempted suicide. The scientists wanted to see if the IAT could distinguish those who had attempted suicide from those who had not.

The IAT is a reaction time test. During their hospital stay, often while sitting in bed, the patients very rapidly classified words on a computer screen, words like: lifeless, thrive, myself, deceased, they, theirs, survive, breathing. And so forth. The idea is to see how rapidly patients connect identity-related words to either life or death words. And the findings were unambiguous. As reported on-line this week in the journal Psychological Science, patients who had attempted suicide prior to admission had much stronger unconscious associations between self and death.

But the study didn’t end there. Nock followed all the patients for six months to see how they fared, and he found that the patients with a powerful self-death association in the hospital had a six-fold increase in later suicide attempts. Six-fold is a dramatic difference, and what’s more, the unconscious associations were a much better suicide predictor than depression, previous suicide attempts, or the intuition of the attending clinician.

What about the patients’ own predictions? Fourteen of the emergency patients attempted suicide within six months of leaving the hospital. Their self-evaluations were an indicator of their future risk, but an imperfect indicator. The IAT results were a better prognosticator even than the patients’ self-evaluations. This suggests that unconscious thoughts might be a useful detector and predictor of intentions that patients are reluctant to discuss—or intentions of which they themselves are unaware.

Excerpts from “We’re Only Human” appear regularly in the magazine Scientific American Mind. Wray Herbert’s book, On Second Thought: Outsmarting Your Mind’s Hard-Wired Habits, will be published by Crown in September.