(PS9-D91) The Vulnerability of Experiencing Suicidal Ideation in Response to Stress: A Proposed Indicator of Risk for Suicidal Ideation

Identifying suicidal ideation (SI) risk factors is important for early stage suicide prevention. Theoretical and empirical evidence suggests a cognitive response pattern underlies SI (Cha et al., 2010; Wenzel et al., 2008) and is activated by emotional fluctuations (Williams et al., 2008), such as those initiated by stress responsiveness—a proposed suicidal thinking subtype (Bernanke et al., 2017). Thus, frequently experiencing SI in response to stress may increase risk for future SI. We investigated the frequency, or vulnerability, of experiencing SI in response to stress and its relationship to past month SI.

Undergraduate students (UG; N=320, SI: 27.5% lifetime, 8.1% past month; 69.4% female, 78.8% white, M_age=19.38) and Amazon Mechanical Turk workers (MT; N=368, SI: 24.2% lifetime, 11.7% past month; 41.6% female, 89.1% white, M_age=37.47) were recruited. Measures of the vulnerability to experiencing SI in response to stress during a prior one year period were developed: (1) percentage of stressful events resulting in active SI; (2) 8 item measure of the frequency of suicidal responses when under stress (e.g., ‘When experiencing stress, I have thoughts of killing myself’); items were adapted from validated SI measures (Heisel & Flett, 2006; Liu et al., 2021; Nock et al., 2007) and responses (Never to Almost Always) summed as a total score; (3) 1 item yes/no measure (i.e., ‘When experiencing stress, I frequently have thoughts of killing myself’). SI during the past month (Nock et al., 2007) and SI risk factors (e.g., hopelessness, depression, distress tolerance) were evaluated. Psychometric properties of the 8 item scale and correlations among the vulnerability measures were evaluated within both samples. Hierarchical logistic regression evaluated the relationship of each vulnerability measure with past month SI, beyond SI risk factors, including lifetime SI.

Items from the 8 item scale were correlated (all p< .001) and showed good internal consistency (UG, r’s .35 to .85, Cronbach’s α = .93; MT, r’s .58 to .76, Cronbach’s α = .94); results were consistent among those with SI history. A 1 factor structure was indicated (UG, λ = 5.63, 70.31% variance, loadings: .60 to .89; MT, λ = 5.66, 70.79% variance, loadings: .74 to .85). All vulnerability measures were correlated (p< .001; UG: r’s = .47, .50, .67; MT: r’s = .40, .51, .59), and among those with SI history. Within the UG sample all vulnerability measures predicted past month SI beyond other risk factors (OR’s= 1.05, 1.12, 9.94) and among those with SI history (% value, OR=1.08 [1.03, 1.14]; 8 item, OR=1.20 [1.01, 1.34]; 1 item, OR=6.01 [1.63, 27.21]). Within the MT sample, the percentage value (OR=1.03 [1.01, 1.05]) and 8 item scale (OR=1.25 [1.11, 1.40]) predicted past month SI beyond other factors (1 item, p=.051), and among those with SI history (% value, OR=1.05 [1.03, 1.08]; 8 item, OR=1.24 [1.08, 1.41]; 1 item, p=.056),

The frequency of experiencing SI in response to stress may serve as an important SI risk factor, with evidence of the utility of three vulnerability measures for future use. Sample size may be a limitation in understanding the performance of these measures. Mediators of this relationship should be investigated (e.g., cognitive flexibility).