(PS11-50) Psychometric Characteristics of the Passive and Active Suicidal Ideation Scale

Suicide is a leading case of death worldwide. Suicidal ideation (SI) is a top risk factor for dying by suicide and of standalone clinical interest. SI ranges from passive (PSI; i.e., a desire to be dead) to active (ASI; i.e., desire to kill oneself) ideation in severity (Liu et al., 2020). Past work suggests potential differences in vulnerability and resiliency factors for PSI and ASI (Liu et al., 2021); micro-longitudinal timescales may also differ (Hallensleben et al., 2019). However, limited research disentangles PSI and ASI potentially due to a lack of measures that cleanly differentiate between them. The current study aims to provide preliminary validation of the Passive and Active Suicidal Ideation Scale (PASIS) which assesses PSI and ASI over the past 7 days. A PASIS past day (PASIS-PD) version is also examined.

Participants were recruited through Prime Panels as part of a larger study. Respondents failing data quality checks or reporting no lifetime history of suicidal thoughts and behaviors (STBs) were excluded from analyses. This resulted in a final sample of 672 adults (age = 42.4; 55.0% Female; 67.0% White). Participants completed both the PASIS and the PASIS past day. The present PASIS version has 17 items (9 PSI, 8 ASI). The PASIS-PD contained 18 items with 9 for each subscale. Participants also completed measures of hopelessness, social connectedness, perceived burdensomeness, among others.

The PASIS exhibited exceptionally high internal consistency for PSI (ω_h = 0.91, α = 0.98), ASI (ω_h = 0.94, α = 0.98), and total (ω_h = 0.90, α = 0.99) scores. The PASIS-PD showed high internal consistency (ω_h = 0.93, α = 0.98, ω_h = 0.93, α = 0.98; ω_h = 0.90, α = 0.98, respectively). PASIS-PD scores were highly correlated with corresponding PSI (r = 0.81), ASI (r = 0.79), and total (r = 0.83) PASIS scores. Confirmatory factor analysis (CFA) for ordinal responses supported a two-factor structure for the PASIS (CFI = .995, RMSEA = .103; SRMR = .026) and PASIS-PD (CFI = .992, RMSEA = .115; SRMR = 0.028). Factors were highly correlated for both (r > .94). Single factor models fit marginally worse.

PASIS PSI (p = .002), ASI (p < .001), and total (p < .001) scores were higher for males than females. The same pattern was observed for PASIS-PD scores (p = .011, p < .001, p = .003). PASIS and PASIS-PD subscales correlated with external variables in expected directions. History of past month SI (yes/no), suicide planning (SP), and suicide attempt (SA) were regressed on PASIS and PASIS-PD subscales. ASI scores were associated with past month SI (b = 0.14, p</em> < .001), SP (b = 0.26, p < .001), and SA (b = 0.18, p < .001) history but PSI scores were only associated with SI (b = .19, p < .001). The same pattern of findings was observed for PASIS-PD scores.

Results provided psychometric support of the PASIS and PASIS-PD as measures of PSI and ASI in an online sample of U.S. adults with self-reported STB history. The measures exhibited high internal consistency as well as convergent validity. CFA supported separate PSI and ASI subscales. Evidence for differential relationships between PSI and ASI with past month SI, SP, and SA was observed. Future work should explore observed group differences from a measurement invariance perspective, psychometric characteristics in clinical samples, and short-form development.