Impairment in risk-sensitive decision-making in older suicide attempters with depression

Abstract

Suicidal behavior is a potentially lethal complication of late-life depression. In younger adults, suicide has been linked to abnormal decision-making ability. Given that there are substantial age-related decreases in decision-making ability, and that older adults experience environmental stressors that require effective decision-making, we reasoned that impaired decision-making may be particularly relevant to suicidal behavior in the elderly. We thus compared performance on a probabilistic decision-making task that does not involve working memory ("Cambridge Gamble Task") in four groups of older adults: (1) individuals with major depression and a history of suicide attempt (n = 25), (2) individuals with major depression with active suicidal ideation but no suicide attempt (n = 13), (3) individuals with major depression without suicidality (n = 35), and (4) nondepressed control subjects (n = 22). There was a significant effect of group on quality of decision-making, whereby the suicide attempters exhibited poorer ability to choose the likely outcome, compared with the nonsuicidal depressed and nondepressed comparison subjects. There were no group differences in betting behavior. The suicide attempters differed in several aspects of social problem-solving on a self-report scale. Quality of decision-making was negatively correlated with the score on the impulsive/careless problem-solving subscale. These data suggest that older suicide attempters have a deficit in risk-sensitive decision-making, extending observations in younger adults. More specifically, older suicide attempters seem to neglect outcome probability and make poor choices. These impairments may precipitate and perpetuate suicidal crisis in depressed elders. Identification of decision-making impairment in suicidal elders may help with designing effective interventions.

Figure 1

Performance on the Cambridge Gamble Task in the depressed suicide attempters, depressed ideators, non-suicidal depressed, and non-psychiatric control participant groups. A) Quality of Decision-Making: the proportion of trials where the subject selected the box color in the majority, as a function of box ratio. B) Decision Latency: the latency (msec) to make the red/blue color decision, on trials where the subject selected the color in the majority. C) Betting Behavior: the percentage of the current points total wagered, as a function of box ratio, on those trials where the subject selected the color in the majority.