Predicting suicidal ideation by interpersonal variables, hopelessness and depression in real-time. An ecological momentary assessment study in psychiatric inpatients with depression

Abstract Objective: To extend evidence on the short-term variability of passive and active suicidal ideation (SI) and the association with suggested proximal risk factors such as interpersonal variables (perceived burdensomeness [PB], thwarted belongingness [TB], hopelessness, and depression) in real-time. Methods: This is an observational study using a prospective design applying ecological momentary assessments (EMA). Eligible for study inclusion were inpatients with unipolar depression, current or lifetime suicidal ideation, and fluent German. Over six days, 74 participants rated their momentary level of passive and active SI, PB, TB, depressiveness, and hopelessness up to 10 times per day on smartphones. Data was collected from August 2015 to July 2017. Compliance was excellent (89.7%). Results: Mean squared successive differences supported temporal instability for all variables. According intra-class correlations, between 25% and 47% of variance was accounted for by within-person variability. Multilevel analysis demonstrated significant positive associations between hopelessness, depressiveness, PB, and TB with passive SI. Prospectively, hopelessness and PB remained predictors of passive SI. For active SI, hopelessness, depression, PB, and TB were significantly associated cross-sectionally. Prospectively, hopelessness, PB, and the interaction PBxTB predicted active SI. All models were controlled for previous level of SI. Conclusions: This study provides further evidence on the short-term variability of SI in very short time frames implying the need of assessing SI repeatedly in clinical and research settings. The associations between interpersonal variables and passive and active SI were only partial in line with assumptions of the Interpersonal Theory of Suicide. Overall, the effects were small warranting further investigation.