(PS2-A9) Validity of the Death Implicit Association Task Among Inpatients at a High Risk for Suicide

Explicit measures of suicide risk can be inaccurate because individuals may lack insight or be hesitant to disclose suicidal thoughts for fear of hospitalization or other negatively perceived outcomes. Implicit assessments such as the Death Implicit Association Test (D-IAT), which do not rely on self-report, may provide additional clinical utility in suicide risk assessment. The D-IAT is a computerized test in which participants categorize word pair stimuli (death\life and me\not me) as accurately and quickly as possible. A d score is calculated from the difference between reaction times for death\me pairings and life\me pairings. Positive d scores (i.e., faster reaction times to death\me pairings) indicate stronger death\me implicit associations and are proposed to be associated with suicide risk. The D-IAT has been inconsistently related to explicit measures of suicide risk in outpatient and inpatient samples, and therefore replication is needed to validate the D-IAT as a measure of suicide risk. The present study assessed the validity of the D-IAT in an adult inpatient sample at high risk for suicidal behavior (n = 97; age M = 32.8 years, SD = 12.59; 38% racial/ethnic minority, 34% sexual/gender minority). All participants reported either a history of at least one suicide attempt in the previous two years with current ideation and plan or a suicide attempt in the previous week before admission. Participants completed the D-IAT and self-report questionnaires as part of a baseline assessment for a larger suicide prevention treatment study, which also included explicit assessments of suicide-related thoughts (Adult Suicidal Ideation Questionnaire [ASIQ], Suicide Cognitions Scale [SCS]), suicide attempt history (Columbia Suicide Severity Rating Scale), and emotional distress (Depression Anxiety Stress Scales [DASS]). Spearman-Brown coefficient indicated very strong split-half reliability, r(95) = .893. It was hypothesized that d scores would be significantly and positively associated with these explicit measures. Correlational analyses partially supported these hypotheses. Specifically, d scores were positively and significantly associated with suicide-specific measures including the ASIQ [r(97) = .314, p = .002], SCS-Unbearability Subscale [r(97) = .232, p = .022], and number of previous suicide attempts [r(97) = .239, p = .018], but not general distress as measured by the DASS. These preliminary results support the validity of the D-IAT as a measure related specifically to suicidal thoughts and behavior rather than general emotional distress in this high-risk sample. Data collection is ongoing and analyses from the larger sample will be presented.