Future Suicide Attempt: Psychobiological Features

Columbia University Health Sciences, New York NY

Investigators

Linked publications & trials

Abstract

Project Summary Most studies evaluating risk for future suicide attempts in major depressive disorder (MDD) have methodologic limitations, e.g., no structured diagnostic interviews or restricted to one or other domain of potential predictors. However, determinants of suicidal behavior are derived from multiple domains including psychopathologic, neurobiological, neurocognitive, familial and psychosocial. As such, development of predictive or explanatory models of suicide attempt behavior requires assessment of potential risk factors from multiple domains in the same patient population in a prospective follow-up study. This prospective study of suicidal behavior in major depression utilizes a unique cohort of patients undergoing a baseline evaluation for MDD, comprising two comparable groups of past attempters and nonattempters. Patients receive extensive clinical and biological baseline assessments funded by the MH062185 Conte Center for the Neuroscience of Mental Disorders at NYSPI/Columbia University. This application provides funding to assess patients for clinical state, life events, treatment and suicidal behavior at 3, 12 and 24 months after discharge. The initial major emphasis had been to build a clinical predictive model and begin testing that model. In the current funding period, the utility of biological predictors was assessed. Promising results identify stress reactivity, neurocognitive indices, and PET quantification of the serotonin system as potential risk factors that principally target traits underlying the diathesis for suicidal behavior. The current application shifts the emphasis to an integration of these biological and neurocognitive predictors with previously identified clinical predictors into a explanatory AND predictive model of suicidal behavior. To that end, the application adds an evaluation of early childhood trauma, neurocognitive tests (decision-making and problem solving) and stress reactivity (Trier Social Stress Test) measures, to complement PET scans and clinical measures obtained in the Conte Center at baseline. Statistical model building methods and prediction strategies will construct optimal models with both heuristic and clinical value.