(PS9-B30) Developing a Hierarchical Model of Suicide

Background: Theories of suicide can help understand suicidality, however, each theory provides a limited picture of contributing factors (Joiner, 2011; Klonsky and May, 2015). Although there have been efforts to combine theories of suicidality (Joiner et al., 2002), there is little empirical research on the overlap between these theories. This study sought to develop and test the stability of a hierarchical model of factors that contribute to suicidality by applying a sequential factor analytic procedure (Goldberg, 2006) to components of several existing theories of suicide.
Methods: Participants (N = 344) at risk for or with current depression were recruited from either a depression/suicide screening study (n = 164; Leykin et al., 2012) or a managed recruitment platform (Prolific: n = 180). Measures captured two theories of suicidality: Interpersonal Theory of Suicide (IPST)(Interpersonal Needs Questionnaire; Van Orden et al., 2012; Acquired Capability for Suicide Scale–Fear of Death Scale; Riberio et al., 2013) and Defeat and Entrapment (DE; Entrapment and Defeat scale; Gilbert & Allan, 1998). Additionally, given the association of Self-Discrepancy Theory with suicidality (SDT; Cornette et al., 2001; Joiner et al., 2002), the Self-Discrepancy Questionnaire (Higgins, Klein, & Strauman, 1985) was used. The “Bass-ackwards analysis” (Goldberg, 2006), a sequence of principal components factor analyses and correlational analyses, was used to derive the hierarchical factor structure, which was tested on the overall sample, and subsamples (men, women, older (31+), younger). 

Results: Results suggest a nine-level hierarchical structure, with a single-factor suicidality at the top, and individual subscales from each theory at the bottom. For overall sample and subsamples, constructs from the IPST and DE theories (i.e., burdensomeness, entrapment, etc.) converged at 4th level into a single Intra/Interpersonal Dejection factor. SDT (ought, ideal, feared selves) converged in the 3rd level into a Self-Loathing factor, and joined Acquired capability at the second level. For men, acquired capability failed to join the rest of the hierarchy at the 3rd level (r = .17, below the .3 cut-off). At lower levels of the hierarchy, IPST and DE constructs (except acquired capability) had different patterns of interrelatedness at lower levels of hierarchy for men and for women. For older subsample, burdensomeness only converged with other IPTS and DE constructs at the 6th level (rather than at the 8th for younger participants). Finally, unlike in all other subgroups, for younger participants, acquired capability converged with feared self at the 3rd level (Antipathy factor), which appeared separate from the rest of the hierarchy (r = -.27).

Conclusion: ISPT constructs (except acquired capability) and DE constructs tended to group together, suggesting that these theories may involve similar constructs. Acquired capability appeared to be a unique construct in all groups, likely due to its instrumental function rather than an expression of preference. Interestingly, SDT constructs were likely to join acquired capability. Gender and age differences suggest a need to explore differences in predictors and factors that contribute to suicidality.