(PS9-6) Innovations in Mental Health Screening: Measuring Implicit Suicide Risk in a Community Sample of Adolescents

Suicide is a leading cause of death among adolescents, and the rate is increasing (Curtin & Garnett, 2023). Most screenings rely on subjective self-report to determine risk; yet, those who die by suicide have denied the experience of suicidality before their death (e.g., McHugh et al., 2009). This paradox presents an urgent need for objective measures of suicide risk to aid in prevention and early intervention (Sohn et al., 2021).

The Death Implicit Association Test (D-IAT; Nock et al., 2008) is one example of an objective measure that has been used to predict future suicide attempts in clinical settings. It involves categorizing words that appear on a computer screen (i.e.., words related to death/life and me/not me), with higher scores indicating greater risk in acute settings (Glenn et al, 2019). The D-IAT’s utility and feasibility has not yet been investigated in universal settings.

In this pilot study, we added the D-IAT to a universal screening tool to determine if adding it increased incremental validity. As a first step, we tested whether adding the D-IAT presented iatrogenic effects. We further investigated whether any order effects should be considered when adding the D-IAT to a universal screening tool.

After parental consent and assent, participants (N = 55; 49% girls, 79% White; Mage = 14.21, SD = 1.66) completed assessments including the D-IAT and a self-report screening tool comprising items on psychopathology and suicidal thoughts and behaviors (Hilt et al., 2018). They also participated in a brief clinical interview, (Self-Injurious Thoughts and Behaviors Interview; Nock et al., 2007). We administered measures of state emotion (Self-Assessment Manikin; Bradley & Lang, 1994) before and after participants completed the D-IAT to assess potential iatrogenic effects on mood, and we counterbalanced the order of D-IAT administration.

We are continuing to collect data for another 2 months. Preliminary findings show no iatrogenic effects on mood, comparing SAM scores pre-to-post D-IAT (valence M = -.02, SD = .89, t = -.16, p = 8.75; arousal M = -.02, SD = 1.39, t = -.10, p = .92; dominance M = -.22, SD = 1.17, t = -1.34, p = .188). When considering incremental validity, thus far, 6 participants have a D-score indicating potential risk in the absence of self-reporting any suicidal thoughts or behaviors. When data collection is completed, we will conduct a hierarchical logistic regression to analyze how self-report and D-IAT scores predict outcomes from the clinical interview. Finally, we did not find evidence of any order effects.

One potential limitation is that the D-IAT relies on the use of pronouns that some adolescents who belong to gender minority groups may use for themselves (i.e. them/them/theirs), but would be categorized as “other” on the test. This limitation poses concern as gender-minoritized adolescents are at an increased risk for suicide (Yıldız, 2016). This study lays the initial groundwork for prioritizing theoretical and evidence-based self-report and objective measures to provide early intervention and potentially save lives. Next steps include testing the D-IAT in a large, prospective design to determine predictive validity regarding future suicidal behaviors (e.g., attempts).