Suicide Risk on Twitter

Introduction

Suicide ranks as the second leading cause of death among individuals 25–34 years old and the third leading cause of death among 15–25 years old (Sussman, 2002. Preventing suicide is inherently complicated by the heterogeneity of individuals who commit suicide and the lack of strong, reliable predictors of suicide. Less than 50% of suicide victims contact a mental health or primary care provider within one month of their suicide attempt (Luoma, 2002). As such, there is more interest in leveraging social media platforms to detect suicidality and intervene in high risk cases outside the healthcare delivery system (Robinson, 2015). To better detect suicide risk, previous research manually analyzed the contents of suicide notes/letters as they include thoughts and feelings of completers that may be indicative of their emotional and mental state directly before they die (Barr, 2007; Foster, 2003; Ho, 1998; Kuwabara 2009).

Recently, researchers investigated the utility of applying automated and computational methods to suicide notes to find patterns of behaviors or language associated with suicide. Ultimately, the objective is to describe patterns that would guide early interventions that would prevent active suicide. For example, in (Pestian, 2008; Pestian, 2010), natural language processing approaches were applied to distinguish between classes of suicide notes (of completers versus not). In a different study (Lewinsohn, 1994), a self-administered risk assessment tool has shown that adolescents with previous suicide attempts have many psychological risk factors (i.e. current suicidal ideation and depression, recent attempt by a friend, low self-esteem, and having been born to a teenage mother) in common. Although these studies are important, the reported results were based on small scale data; therefore, conclusions need to be further investigated with larger and other samples, perhaps using big data, before generalization. Social media, a big data resource, has been recently utilized for promoting positive behaviors such as help seeking for depression management (Guan, 2015), surveying social needs (Hui, 2015) and preferences on receiving mental health services using technology (Krueger, 2015). Social media has also been used to identify users with high suicide probabilities (Lal, 2015).