Abstract
Background
Along with other domains of psychopathology, adolescents are at increased risk for psychotic experiences (PE) as well as for suicidal ideation (SI) and suicidal attempts. Literature has found robust associations between PE and SI, with suicidal risk appearing higher in subjects with persistent PE, in general and clinical samples. However, the associations between psychotic experiences and suicidal ideation are not clearly understood, and the role of common risk factors in this link, such as depression (DS), is controversial. Moreover, causality has yet to be properly stablished between PE and SI. A study recently found psychosis may be consequential to suicidal behavior. Under the novel “suicidal drive hypothesis for psychosis” framework, we examined whether depressive symptoms mediate the association between psychotic experiences and suicidal ideation on cross-sectional data from a community sample of adolescents. Additionally, we tested this mediating role when both PE and SI were the outcome variables.
Methods
1708 Chilean school adolescents aged 13–19 years (M=15.68 + 1, 67, women= 39%) answered an online mental health screening between May and October 2019. The screening encompassed several questionnaires adapted for Chilean subjects, including the Columbia-suicide Severity Rating scale, the Community assessment for Psychic experiences (CAPE-P15) and the Patient Health Questionnaire (PHQ-9). Informed consent was granted by the parents and caregivers of adolescents.
Mediation analyses were executed with the PROCESS statistical package. PROCESS runs regression-based mediations based on the Baron & Kenny procedure. A 5000 bootstrap resampling was used to estimate 95% confidence intervals; these are used to identify an indirect effect that suggests mediation.
Results
Pearson’s correlations showed significant associations between all the variables in (SI-DS: r= .624, p<.001; PE-SI: r= .539, p<.001; PE-DS: r= .708, p<.001). No demographic variables (i.e age, gender, education) were correlated significantly to the dependent variable (DV), therefore no covariates were controlled for in the mediation analyses.
In the mediation analyses, model 1 showed SI as the DV. The link between psychotic experiences and suicide was mediated by depression (b= .3433, 95% BCa CI [.2981, .3885]). When the mediators were not included in the model, PE significantly predicted SI (b = .113, t = 26.45, p < 0.001). Additionally, in model 2, where PE was the DV, the link between suicidal ideation and psychotic experiences was mediated by depression as well (b= .3794, 95% BCa CI [.3431, .4150]). When the mediators were not included in the model, SI significantly predicted PE (b = 2.57, t = 26.45, p < 0.001). Both models exhibit robust and significant partial mediations.
Discussion
Our results indicate that depressive symptoms partially mediate the association between SI and PE. Moreover, either PE or SI could be outcome variables when depressive symptoms are mediators. This adds new evidence supporting that PE could consequential to SI as stated by the “suicidal drive hypothesis for psychosis”. Our exploratory findings must be carefully interpreted, mainly because of our cross-sectional design, and the fact that there could be unmeasured or non-controlled psychopathological confounder variables in our models.
What matters for adolescent suicidality: depressive symptoms or fixed mindsets? Examination of cross-sectional and longitudinal associations between fixed mindset and suicidal ideation