Delineating Empirically Plausible Causal Pathways to Suicidality Among People at Clinical High Risk for Psychosis

Suicidality is common among people at clinical high risk (CHR) for psychosis. Delineating causal pathways to suicidality and identifying its determinants would inform tailored intervention efforts for these individuals. To this end, we analyzed data on CHR samples from the second and third North American Prodrome Longitudinal Studies (NAPLS-2, n = 355; NAPLS-3, n = 266). Data on correlates of suicidality—including depression and attenuated psychosis symptoms, sleep, and childhood trauma—from two initial study timepoints were submitted to the greedy relaxations of the sparsest permutation algorithm. Intervention calculus was used to estimate the (lower bound) total empirically plausible causal effects of each variable on suicidality. Across both samples, greedy relaxations of the sparsest permutation suggested that symptoms of depression—particularly hopelessness, self-deprecation, and depressed mood—were likely direct causes of suicidality among people at CHR for psychosis. Across samples and measurement time points, intervention calculus indicated that depressed mood exerted the greatest influence over suicidality of all measured variables. This study provides data-driven, testable hypotheses about the causal pathways leading to suicidality among people at CHR for psychosis and suggests promising targets for interventions on suicidality tailored to these individuals. Future experimental research should test these hypotheses by, for example, comparing the suicide risk reduction afforded by interventions aimed at each aforementioned target.