A Formal Test of the Interpersonal Psychological Theory of Suicide and a Closer Look at the Role of Social Support in Adolescent Suicidal Ideation and Behavior

Adam Miller

Advisor: Christianne Esposito-Smythers, PhD, Department of Psychology

Committee Members: Sarah Fischer, Rick Zimmerman

Psychological Clinic, 10340 Democracy Lane Suite 301E
June 17, 2014, 03:00 PM to 12:00 PM


The purpose of this dissertation project was to examine the theoretical and empirical importance of perceptions of interpersonal relationships to suicidal ideation (SI) and behavior in an adolescent clinical sample. The first study offered a rigorous test of the interpersonal-psychological theory of suicide as it applies to SI in a sample of adolescents at risk for suicide. Specifically, the association between the two interpersonal states central to this theory, perceived burdensomeness and thwarted belongingness, and severity of SI, was tested across three separate models. The first was a cross-sectional model, the second a short-term prospective model, and the third a competing developmentally sensitive mediational model that used a short-term prospective study design. Participants were 143 adolescents (64% female, 81% white, range = 12-18 years, M = 15.38, SD = 1.43) consecutively admitted to a psychiatric partial hospitalization program. Data were collected with paper and pencil surveys upon intake into the program (Time 1 [T1]) and discharge from the program (Time 2 [T2]). Youth SI was assessed with the Suicidal Ideation Questionnaire and depression (diagnosis and symptom severity) with the Youth Inventory-4. Youth also completed the Interpersonal Needs Questionnaire, which assesses for perceptions of burdensomeness and thwarted belongingness. Results of cross-sectional analyses showed an independent association of perceived burdensomeness on T1 SI, after controlling for depression diagnosis and sex. There was no main effect of thwarted belongingness and no significant interaction between the two interpersonal states. In a short-term prospective moderation model, neither interpersonal state predicted T2 SI after controlling for covariates. In a third mediation model, thwarted belongingness, but not perceived burdensomeness, had a significant indirect effect on T2 SI via T2 depressive symptom severity after controlling for T1 SI and sex. Results suggest that: 1) perceptions of burdensomeness may contribute to concurrent risk for SI; and 2) thwarted belongingness affects depression symptom severity over time, which indirectly predicts SI over a short follow-up time frame. These results only partially replicate findings with adults in prior tests of the interpersonal-psychological theory of suicide. Though results are preliminary, these data suggest that perceptions of burdensomeness and thwarted belongingness may function differently in adolescent relative to adult clinical samples.

The second study examined the relative contributions of perceptions of social support from parents, close friends, and school on current SI and suicide attempt (SA) history in a clinical sample of adolescents. It also explored whether interactions between these sources of support help explain additional variance in suicidality. Data were collected with paper and pencil surveys and a structured clinical interview from the 143 adolescents described in the first study. The well-validated Beck Scale for Suicidal Ideation, Child and Adolescent Survey of Social Support, Youth Inventory-4, and the Self-Injurious Thoughts and Behavior Interview were used to assess study constructs. Hierarchical linear and logistic regression analyses were conducted to examine main and interactive effects of perceptions of social support on SI and SA, respectively. Results from the linear regression analysis revealed that perceptions of low school support independently predicted greater severity of SI after accounting for parent and close friend support. Further, the relationship between low perceived school support and SI was strongest among those who perceived low (versus) high parental support. Results from the logistic regression analysis revealed that perceptions of low parental support independently predicted SA history after accounting for school and close friend support. A significant interaction between close friend and school support was also found. Those who perceived low support from school and close friends reported the greatest odds of a SA history. Results address a significant gap in the social support and suicide literatures by demonstrating that perceptions of parent, school, and close friend support play an important role in understanding adolescent SI and SA. Moreover, the confluence of low support across sources may place adolescents at heightened risk for more severe SI and/or behavior.

Overall, results of both studies suggest that perceptions of interpersonal relationships may play an important role in adolescent mental health, including SI and suicidal behavior. Adolescents who believe that they are a burden on others or lack close connections with family members, peers, and/or school staff school, may be at heightened risk for SI or behavior either directly or indirectly via co-occurring mental health problems. Clinically, results suggest that efforts to improve relationships and perceptions of support across all three social domains in the context of treatment with adolescents in clinical care may be important in suicide prevention efforts.