(PS7-97) The Effects of Thought Disturbance, Sleep Difficulties, and Sex on Suicidal Ideation Among Adolescents in the Juvenile Legal System: A Three-way Moderation Model

Background: Sleep difficulties are prevalent among adolescents with psychosis-spectrum experiences, and both phenomena have been linked to increased suicidal ideation (SI). The associations between these symptoms have not been well-explored in juvenile legal system (JLS)-involved youth, a population that is at increased risk for mental health concerns including suicidality. The current study sought to explore the hypothesis that psychosis-spectrum symptoms would be positively associated with SI among JLS-involved youth. Furthermore, given evidence that sleep difficulties may exacerbate mental health concerns and biological sex has been differentially linked to psychosis, sleep, and SI, we hypothesized that sleep quality and sex would moderate the association between psychosis-spectrum symptoms and SI.

Methods: Participants (n = 1290, aged 10-18) were diverted youth who, upon JLS intake, completed the Massachusetts Youth Screening Inventory- 2 (MAYSI-2), which probes suicidal ideation (SI), thought disturbance (TD; i.e. potential psychosis-spectrum symptoms), sleep difficulties, and other mental health symptoms in the following domains: alcohol/drug use (AD), angry-irritable (AI), depressed-anxious (DA), and traumatic experiences (TE). Individual questions are scored yes (1) or no (0), and each scale includes five to nine items. Prior to analyses, all MAYSI-2 scale scores were log transformed due to the non-normal distribution of several scales (i.e. there was a large proportion of “0” responses).

Results: Simple linear regression indicated that TD, sleep difficulties, and female sex were all positively associated with SI. In a test of robustness, the other MAYSI-2 scales and demographic characteristics were added to the model. Results indicated that TD and sex, but not sleep difficulties, remained significantly linked to SI, beyond the significant effects of the AD, DA, and TE scales and the non-significant effects of AI, age, race, and ethnicity. A subsequent moderation analysis identified a significant three-way interaction between TD, sleep difficulties, and sex on SI, controlling for the AD, DA, and TE scales. Probing the interaction indicated that among males, TD was not significantly linked to SI at low and moderate levels of sleep difficulties. At the highest level of sleep difficulty, however, TD was positively associated with SI. Among females, TD was not associated with SI at the lowest level of sleep difficulty, but a positive effect was significant and incrementally stronger at increasing levels of sleep difficulties.

Conclusions: Findings indicate that TD symptoms are associated with SI in this JLS sample, and sleep quality moderates this association for both males and females in distinct ways. Females may be particularly vulnerable to the impact of minimal sleep difficulties on psychosis-spectrum experiences and suicidal thoughts. Overall, these results suggest that better sleep may buffer the effect of TD on SI, whereas poorer sleep may exacerbate this association. Interventions targeting sleep health may be particularly helpful for JLS youth with psychosis-spectrum symptoms to mitigate suicide related risk.