Non-suicidal self-injurious behavior (NSI) among adolescents is a signi.cant and growing problem. Estimates of the prevalence of NSI range from 2.4% to 25% in community samples and up to 61% among psychiatric inpatients (Diekstra et al., 1995). Many adolescents who engage in NSI have comorbid mood and conduct disorders, as well as may meet criteria for select personality disorders. Borderline personality disorder (BPD) is a common diagnosis among adults who engage in NSI. Many of the key symptoms of BPD have been found to correlate with NSI such as emotional and interpersonal lability, impulsivity, and poor self-concept (Linehan, 1993). Little research has examined the in.uence of BPD symptoms on NSI within adolescent populations. It is important to determine whether characterological symptoms predict NSI in adolescents beyond Axis I diagnoses. The purpose of the current study was to determine whether specifc symptoms associated with BPD were predictive of lifetime and repetitive NSI in adolescents after controlling for comorbid diagnoses. We hypothesized that a greater number of BPD symptoms would be predictive of both lifetime occurrence and frequency of NSI. We also examined the unique contribution of each separate BPD symptom to predicting lifetime and frequency of NSI. Due to the exploratory nature of this part of the study, no speci.c hypotheses were made.
Participants & Procedure
A chart review was conducted on 457 patients consecutively admitted to the Adolescent Depression & Suicide Program at Monte.ore Medical Center between December 1998 and February 2006. NSI data was available for 441 adolescents.
As this is a chart review study, no formal informed consent was obtained, but participants and their guardians signed consent to complete a series of psychological measures as part of their treatment. Adolescents were administered the Schedule for Affective Disorders and Schizophrenia, Child version; Structured Clinical Interview for DSM-IIIR Personality Disorders, Borderline Personality Module; and Lifetime Parasuicide Inventory.
Of the total sample, 146 (33.1%) reported at least one act of NSSI. The mean frequency of NSSI was 8.24 (SD = 15.61) with a modal frequency of one. Females comprised 84.9% of those reporting NSSI. Mood disorders were the most common DSM IV Axis I diagnoses (50.3% w/ MDD; 27.7% other mood disorders). 32.2% of the adolescents were diagnosed with a minimum of three BPD symptoms.
Statistical Analyses & Results
A series of logistic regression models were run: 1) gender and Axis I diagnostic groups predicting lifetime incidence of NSI, 2) gender, Axis I, and # BPD sx predicting lifetime incidence of NSI, 3) Gender, Axis I, # BPD sx, and each BPD sx entered separately into a logistic regression predicting life incidence of NSI. The same set of analyses were then conducted predicting frequency of NSSI (high vs low frequency, where high frequency = 3 or more acts of NSI).
Separate logistic regression models for each symptom of BPD along with Gender, Axis I d/o, and # BPD sxs were conducted to predict presence of lifetime NSI:
Results: None of the specific symptoms added significantly to the model.
Impulsivity approached significance (adj. OR = 1.88, p = .05)
Separate logistic regression models for each symptom of BPD along with Gender, Axis I d/o, and # BPD sxs were conducted to predict frequency of NSI (high vs low):
Results: Only the symptom of affective instability added significant variance to the model
(adj. OR = 7.66, p = .001)
Logistic Regression Models including Axis I d/o, # BPD sxs, & each BPD sx simultaneously
1) LR model predicting Lifetime Presence of NSI included: # BPD symptoms, gender, Axis I diagnoses, ALL
BPD symptoms as predictors. ONLY impulsivity significantly predicted NSI.
2) LR model predicting Frequency of NSI included: # BPD symptoms, gender, Axis I diagnoses, All BPD
symptoms as predictors. ONLY affective instability significantly predicted frequency of NSI.
- Female gender and total number of BPD symptoms were predictive of lifetime incidence of NSI, suggesting that adolescents who have greater borderline personality features may be at heightened risk for engaging in NSI.
- When each specific symptom BPD was entered into a separate logistic regression model predicting lifetime incidence of NSI controlling for each of the other variables, only impulsivity added significantly to the model. When all the symptoms of BPD were entered into the model together, along with gender, Axis I d/o, and # of BPD symptoms, impulsivity was the only significant predictor of lifetime incidence of BPD.
- Within the logistic regression model including gender, # of BPD symptoms, and Axis I diagnoses, only the number of BPD symptoms was predictive of frequency of NSI. Again, this suggests that greater borderline pathology is important to understanding repetitive acts of NSI above and beyond gender and Axis I diagnosis.
- In examining the influence of each specific symptom of BPD on frequency, only affective instability
added significantly to the model. Adolescents with affective instability are at 7.6 time greater risk of engaging in frequent NSI. Lending further support to this finding, when all symptoms of BPD were entered simultaneously into the model, affective instability emerged as the only significant predictor of frequency of
- Our results suggest that specific symptoms of BPD are relevant to understanding NSI in adolescents.
Specifically, it appears as though impulsivity may be important to initiating NSI. Among those who try NSI,
it appears that affective instability plays a critical role in determining the frequency, or repetition, of NSI.
- Our results support the importance of targeting impulsivity and emotion regulation in treatment with self-injuring adolescents.
For further information, please contact:
Jennifer J. Muehlenkamp, Ph.D. Dept. of Psychology
University of North Dakota Grand Forks, ND 58202