Corresponding Author Information: Megan Keen
Session Abstract: Predicting Non-Suicidal Self-Injury and Suicidality in Adolescents Receiving Residential Treatment using the MMPI-A-RF
Megan A. Keen, Colette C. Delawalla, & Tayla T. C. Lee
Non-Suicidal Self-Injury (NSSI) and suicidal thoughts and behaviors (Suicidality) are critical threats to the health of adolescents that must be detected and addressed. The Minnesota Multiphasic Personality Inventory ‚Äú Adolescent‚Äú Restructured Form (MMPI-A-RF; Archer et al., 2016) could be used to detect risk for these types of behaviors, but no previous studies have examined this possibility. SUBJECTS, METHODS, & MATERIALS: We investigated whether MMIP-A-RF scale scores predicted NSSI/Suicidality in a sample of 147 adolescents in residential treatment who completed the MMPI-A-RF at intake. A series of t-tests and logistic regressions were used to determine MMPI-A-RF scores' abilities to predict NSSI/Suicidality recorded in the youths' medical records during treatment. ANALYSES AND RESULTS: Adolescents who experienced NSSI/Suicidality (N = 8) endorsed significantly more internalizing difficulties than did those who did not on EID, RCd, HLP, SFD, and MLS (d's = .63 -.80). Logistic regressions examining the ability of individual MMPI-A-RF scores to predict NSSI/Suicidality suggested higher scores on RCd, MLS, SFD, SUB, and BRF were associated with a 6 to7% increase in the odds a youth would experience NSSI/Suicidality. We also examined a model based on past studies suggesting psychological distress, specific internalizing and somatic difficulties, and substance abuse are related to NSSI/Suicidality in adolescents (e.g., Bohman et al., 2012; Soto-Sanz et al., 2019). This model included scores on RCd, internalizing and somatic SP scales, and SUB as potential predictors. In this model, higher scores on RCd [Exp(B) = 1.27], BRF [Exp(B) = 1.23], and SUB [Exp(B) = 1.14] were significantly associated with an increased odds of NSSI/Suicidality, while higher scores on ANX [Exp(B) = .73] and COG [Exp(B) = .84] were significantly associated with a lower likelihood of NSSI/Suicidality. This model correctly classified 138/139 individuals who did not experience NSSI/Suicidality (specificity = 99%) and 2/8 individuals who did (sensitivity = 25%). Overall, results indicate MMPI-A-RF scale scores do not improve on base rate-based determinations of risk for these behaviors. However, given that across all analyses scores on RCd, BRF, and SUB were associated with NSSI/Suicide risk, they warrant further investigation.
Megan Keen | Ball State University, Muncie IN
Megan A. Keen