Psychological Assessment 2015, Vol. 27, No. 1, 31-41
© 2014 American Psychological Association 1040-3590/15/$ 12.00 http://dx.doi.org/! 0 .1037/pas0000021
Assessing DSM-5 Nonsuicidal Self-Injury Disorder in a Clinical Sample Jason J. Washburn
Lauren M. Potthoff
Alexian Brothers Behavioral Health Hospital, Hoffman Estates, Illinois, and Northwestern University Feinberg School of Medicine
Northwestern University Feinberg School of Medicine
K. R. Juzwin
Denise M. Styer
Alexian Brothers Behavioral Health Hospital, Hoffman Estates, Illinois, and Illinois School of Professional Psychology, Schaumburg
Alexian Brothers Behavioral Health Hospital, Hoffman Estates, Illinois
The entry for nonsuicidal self-injury (NSI) disorder in the 5th edition of the Diagnostic and Statistical Manual o f Mental Disorders (DSM-5) provides a criterion-based definition of clinically relevant NSI. NSI disorder is currently classified in the DSM-5 as a condition requiring further study. The present study aimed to examine the reliability, validity, and clinical utility of a self-report measure of NSI disorder, the Alexian Brothers Assessment of Self-Injury (ABASI). The sample included 511 patients admitted to an acute care treatment program designed to treat NSI. Patients were administered the ABASI as part of a clinical assessment and routine outcome evaluation. The sample included a broad age range, as well as sufficient numbers of males and Hispanics to examine sociodemographic differences. The ABASI demonstrated adequate internal consistency and test-retest reliability, and the factor structure reflects NSI disorder criteria. Among patients being treated for NSI, 74% met criteria for NSI disorder. No differences in the rate of NSI disorder were observed by sex, ethnicity, or age. Although NSI disorder is associated with a worse presentation of self-injurious behavior, NSI disorder provides limited clinical utility as a dichotomous diagnosis, at least when compared with common NSI characteristics such as number of methods of NSI and the urge to self-injure. Instead, findings support a dimensional approach to NSI disorder. Analyses of specific symptoms of NSI disorder indicate concerns with Criterion B as currently defined by the DSM-5. Recommendations for a more parsimonious revision of NSI disorder are discussed.
Keywords: self-injury, DSM-5, psychometrics, clinical utility Supplemental materials: http://dx.doi.org/! 0.1037/pas0000021.supp
rates of at least one lifetime incidence of NSI range between 7.6% to 23.2% among children, adolescents, and emerging adults (Barrocas, Hankin, Young, & Abela, 2012; Jacobson & Gould, 2007; Whitlock, Eckenrode, & Silverman, 2006), dropping to 5.9% among adults (Klonsky, 2011). It remains unclear, however, if these rates reflect clinically relevant NSI given that there is no agreed-upon definition of when NSI requires clinical attention. The entry for NSI disorder in the recently published fifth edition of the Diagnostic and Statistical Manual o f Mental Disorders (DSM-5) provides a criterion-based definition of clinically relevant NSI (Amer ican Psychiatric Association, 2013; see Table 1 for specific criteria). NSI disorder was recently found to reflect the behavior of a “prototypic” self-injurer by 119 researchers and clinicians with expertise in NSI (Lengel & Mullins-Sweatt, 2013). Despite general agreement on the criteria, NSI disorder was classified in the D SM -5 as a condition requiring further study due to insufficient evidence supporting its designation as an “official mental disorder” (American Psychiatric Association, 2013). One concern with NSI disorder surfaced from the D SM -5 field trials, which failed to show acceptable interrater reli ability using clinician-administered diagnostic interviews (Regier et al., 2013). Those conducting the D SM -5 field trials, however, strug-
Nonsuicidal self-injury (NSI) refers to the direct and deliberate infliction of damage to one’s bodily tissue without suicidal intent and includes a variety of behaviors such as cutting, burning, scratching skin, biting, and hitting oneself (International Society for the Study of Self-injury, 2007). As an isolated set of behaviors,
This article was published Online First September 29, 2014. Jason J. Washburn, Center for Evidence-Based Practice, Alexian Broth ers Behavioral Health Hospital, Hoffman Estates, Illinois, and Department of Psychiatry and Behavioral Sciences, Northwestern University Feinberg School of Medicine; Lauren M. Potthoff, Department of Psychiatry and Behavioral Sciences, Northwestern University Feinberg School of Medi cine; K. R. Juzwin, Self-Injury Recovery Services, Alexian Brothers Be havioral Health Hospital, Hoffman Estates, Illinois, and Illinois School of Professional Psychology, Schaumburg; Denise M. Styer, Self-Injury Re covery Services, Alexian Brothers Behavioral Health Hospital, Hoffman Estates, Illinois. Correspondence concerning this article should be addressed to Jason J. Washburn, Center for Evidence-Based Practice, Alexian Brothers Behav ioral Health Hospital, 1650 Moon Lake Blvd., Hoffman Estates, IL 60302. E-mail: [email protected] [email protected]
31
32
WASHBURN, POTTHOFF, JUZWIN, AND STYER
Table 1 DSM-5 Criteria fo r Nonsuicidal Self-Injury (NSI) Disorder and Corresponding Methods o f Measurement in the Current Study DSM-5 description
Methods of measurement
Criteria A. In the last year, the individual has, on 5 or more days, engaged in intentional self-inflicted damage to the surface of his or her body of a sort likely to induce bleeding, bruising, or pain with the expectation that the injury will lead to only minor or moderate physical harm (i.e., there is no suicidal intent).
Met Criterion A if reported engaging in 5 or more days of a specific NSI method using the following question on the ABASI: “In the last year, how often did you intentionally (on purpose) do the following things to hurt yourself and/or cause pain?” See Table 2 for a list of the specific NSI behaviors included in the ABASI. The ABASI does not assess the total number of days of NSI across all types. Consequently, only those who reported 5 or more days of NSI within a specific type of NSI were included as meeting Criterion A. Met Criterion B if responded with Agree or Strongly agree to at least one of the following items: “When I self-injure, I expect that it will . . .” 1. Provide relief from negative feelings or thoughts. 2. Fix or resolve problems with other people. 3. Create or increase positive feelings (happy, joyful, excited, cheerful, etc.). Met Criterion C if responded with Some o f the time, Half o f the time, Most o f the time, or All o f the time to at least one of the following items: “In the past 12 months, how often did you . . .” la. Have negative feelings or thoughts (distress, anger, sadness, anxiety, self-criticism, etc.) immediately before self-injuring. lb. Have difficulties or problems with other people immediately before self-injuring. 2. Experience a strong desire or urge to hurt yourself that was difficult to resist before self-injuring. 3. Think about hurting yourself.
Criteria B. The individual engages in the self-injurious behavior with one or more of the following expectations: 1. To obtain relief from a negative feeling or cognitive state. 2. To resolve an interpersonal difficulty. 3. To induce a positive feeling state. Criteria C. The intentional self-injury is associated with at least one of the following:
1. Interpersonal difficulties or negative feelings or thoughts, such as depression, anxiety, tension, anger, generalized distress, or selfcriticism, occurring in the period immediately prior to the selfinjurious act. 2. Prior to engaging in the act, a period of preoccupation with the intended behavior that is difficult to control. 3. Thinking about self-injury that occurs frequently, even when it is not acted upon. Criteria D. The behavior is not socially sanctioned (e.g., body piercing, tattooing, part of a religious or cultural ritual) and is not restricted to picking a scab or nail biting. Criteria E. The behavior or its consequences cause clinically significant distress or interference in interpersonal, academic, or other important areas of functioning. Criteria F. The behavior does not occur exclusively during psychotic episodes, delirium, substance intoxication, or substance withdrawal. In individuals with a neurodevelopmental disorder, the behavior is not part of a pattern of repetitive stereotypes. The behavior is not better explained by another mental disorder or medical condition
Picking scabs and nail biting were excluded as NSI behaviors, and the phrase “to hurt yourself or cause pain” was added to clarify any behavior that may be used for purposes other than selfinjury (See Table 2) Met Criterion E if rated as 50 or below on the Global Assessment of Functioning. Met Criterion F if not clinically diagnosed with any schizophrenia spectrum disorder, delirium, substance intoxication or withdrawal, or any developmental or intellectual disorder.
Note. Description of nonsuicidal self-injury disorder from Diagnostic and Statistical Manual o f Mental Disorders (5th ed., DSM-5), by the American Psychiatric Association, 2013. Copyright 2013 by American Psychiatric Association. Reprinted with permission. ABASI = Alexian Brothers Assessment of Self-Injury.
gled to obtain sufficient samples sizes for NSI disorder to generate accurate kappa estimates; one site failed to recruit any patients, and the other two sites recruited only seven pediatric patients each (Clarke et al„ 2013). Acceptable interrater reliability (k = .90) of NSI disorder using a semistructured interview was recently demonstrated with a sam ple of 73 adolescent females recruited from inpatient units in Switzerland and Germany (In-Albon, Ruf, & Schmid, 2013). The interview questions, however, were based on the initial proposed criteria presented on the DSM website before publication of the DSM-5. Additional recent studies provide further support for the proposed NSI disorder criteria; however, they also failed to assess all of the current DSM-5 NSI criteria (Glenn & Klonsky, 2013; Odelius & Ramklint, in press; Selby, Bender, Gordon, Nock, & Joiner, 2012; Zetterqvist, Lundh, Dahlstrom, & Svedin, 2013). This study builds upon the available literature by examining the reliability and clinical utility of a self-report measure in assessing
the current criteria of NSI disorder proposed in Section III of the DSM-5. Using a large sample of people admitted to an acute care program designed specifically to treat NSI, this study also ad dresses the limitations of prior studies by including adequate numbers of males, adolescents and adults, and Hispanics for ex amination of potential sociodemographic differences. Finally, by examining NSI disorder in a sample of people being treated specifically for NSI, the current study contributes to the evaluation of the clinical utility of the NSI disorder criteria proposed in the DSM-5.
Method Sample The sample included 511 patients consecutively admitted to an acute care program designed to treat NSI. Approximately one third
NONSUICIDAL SELF-INJURY DISORDER
(33.2%) of the sample received inpatient treatment at the time of assessment, with the remaining sample receiving a combination of partial hospitalization and intensive outpatient day treatment. The sample was predominantly female (90.0%) and non-Hispanic W hite (83.2%), with some representation of Hispanics (9.5%), but minimal representation of African Americans (1.8%) and other race/ethnicities (2.0%); race/ethnicity data were not available for 3.5% of the sample. Ages ranged from 12 to 52 years old, with a mean age of 17.3 years (SD = 6.2); the majority o f the patients were either adolescents (72.0%) or emerging adults (15.5%). So cioeconomic status, which was estimated by examining the pro portion of publically funded health insurance (i.e., Medicaid) in the sample and by matching patients’ geocodes to median household incomes reported in the 2011 American Community Survey (U.S. Census Bureau, 2011), indicates an economically diverse sample, with 24.8% receiving Medicaid and a mean household income of $78,327 (SD = $20,727). Nearly the entire sample (95.0%) received a primary diagnosis of an affective disorder from attending psychiatrists using a nonstandardized clinical evaluation. Diagnoses included major depres sive disorder (52.6%), depressive disorder not otherwise specified (NOS; 18.0%), mood disorder NOS (17.5%), and bipolar spectrum disorder (6.9%). Nonaffective primary diagnoses included eating disor ders (2.1%), anxiety disorders (1.7%), and impulse control disorder (0.4%). Partly due to the age of the sample, personality disorders were not routinely diagnosed by the attending psychiatrists and therefore not included in this study.
Procedures The measures used in this study were administered to patients at the time of admission to the treatment program. Patients completed self-report measures as part of routine clinical assessment and clinical outcome monitoring. Completed measures were main tained in archived clinical outcome databases and combined with sociodemographic and clinician-generated data from the medical record for internal organizational improvem ent purposes. N SI dis order was identified by combining the self-report measures with clinician-generated data, as described in Table 1. All individuallevel data were de-identified before analyses were conducted, following the federally defined de-identification standard for pro tected health information. Procedures for this research study were reviewed by both the hospital and university systems’ institutional review boards (IRBs) and found to be exempt from further IRB review pursuant to the Protection o f Human Research Subjects Policy (2009) of the U. S. Department of Health and Human Services.
Instrumentation Alexian Brothers Assessment of Self-Injury (ABASI). The ABASI was created as a self-report clinical assessment in antici pation of NSI disorder being included in the DSM- 5. Specifically, the ABASI was designed to identify NSI disorder in a population o f individuals already identified as engaging in NSI behavior. As such, the ABASI is an assessment o f clinical severity o f NSI, rather than of NSI behavior per se. Criterion A was assessed by asking patients to record the num ber o f days in the past year in which they engaged in each of 21
33
specific NSI behaviors. In addition to typical NSI behaviors that result in damage to the surface of the body, 10 o f the 21 NSI behaviors focused on inducement o f pain or harm below the surface o f the body. Although not consistent with Criterion A in the DSM - 5, these 10 behaviors were included in the ABASI because o f the researchers’ clinical observation that they are used as form s o f self-injury. B ecause the A BA SI did not include a question asking about the total num ber o f days self-injured across all types o f N SI, only patients who reported 5 or more days o f N SI w ithin a specific type o f NSI were included as m eeting C riterion A. Criterion B was assessed with a set o f three items, rated on a 5-point (0 - 4 ) Likert scale from Strongly disagree to Strongly agree, using the lead, “W hen I self-injure, I expect that it w ill. . for each o f the following items: “Provide relief from negative feelings or thoughts” (B -l); “Fix or resolve problems with other people” (B-2); and “Create or increase positive feelings (happy, joyful, excited, cheerful, etc.)” (B-3). Responses o f Agree or Strongly agree to any of these three items were counted as meeting Criterion B. The 5-point responses to the three items in Criterion B were also summed to create a total score for Criterion B ranging from 0 to 12. Criterion C was assessed with a set o f four items rated on a 5-point (0 - 4 ) Likert scale from None o f the time to All o f the time, using the lead, “In the past year (12 months), how often did you . . .” for each of the following items: “Have negative feelings or thoughts (distress, anger, sadness, anxiety, tension, self-criticism, etc.) immediately before self-injuring” (C -la); “Have difficulties or problems with other people immediately before self-injuring” (C -lb); “Experience a strong desire or urge to hurt yourself that was difficult to resist before self-injuring” (C-2); and “Think about hurting yo u rself’ (C-3). For Symptom C -l, the ABASI separates interpersonal and intrapersonal preceding problems into two sep arate items, consistent with findings differentiating interpersonal from intrapersonal functions of NSI (Klonsky, 2007). Responses of Some o f the time, H alf o f the time, Most o f the time, or All o f the time to any o f these four items were counted as meeting Criterion C. To maintain consistency with the NSI disorder criteria, we used the average o f the two items assessing Symptom C -l for any specific analysis of Symptom C-1. Further, the average of the two items assessing Symptom C -l was summed with the 5-point re sponses from the items assessing Criteria C-2 and C-3 to create a total score for Criterion C ranging from 0 to 12. The total scores for Criteria B and C were summed to create an NSI severity score ranging from 0 to 24. To address Criterion D, we did not include scab picking and nail biting as specific behaviors in Criterion A. Further, the phrase “to hurt yourself and/or cause pain” was added to the descriptions of the following NSI behaviors in Criterion A: tattooing/piercing, pulling hair/eyelashes/eyebrows, falling down, carved into skin, fighting, ingested/swallowed items, overexercised, restricted eating/purged/use laxatives, sexual behavior, and not following med ical advice or worsening medical condition. The phrase was added to clarify that the behaviors should only be endorsed if participants engaged in them for purposes o f self-injury, not for socially sanctioned purposes or purposes other than self-injury. For exam ple, the ABASI considers tattooing a form of self-injury when it is engaged in specifically to cause pain or harm (Stirn & Hinz, 2008).
34
WASHBURN, POTTHOFF, JUZWIN, AND STYER
The ABASI also provided data on the characteristics of NSI that are similar to measures already existing in the literature to identify more severe cases of NSI, such as frequency of NSI, number of methods of NSI, age of onset of NSI, and duration of NSI. Frequency of NSI was assessed by taking the average number of days of NSI in the past year across all 21 types of NSI behaviors. Number of methods of NSI was assessed by counting the number of NSI behaviors that were endorsed at least 1 day in the past year. Age of onset was assessed by asking patients to estimate the age of onset for each specific NSI behavior they engaged in during the past year, and taking the youngest age. Duration of NSI was assessed by subtracting patients’ current age from the age of onset of NSI. Global Assessment of Functioning (GAF). The GAF (Amer ican Psychiatric Association, 2000), a measure of overall symp toms and impairment with a range from 1 to 100, was obtained from licensed clinicians conducting a level of care assessment at the time of admission to the acute care treatment program. The GAF provided a measure of Criterion E, with a threshold of 50 or below as meeting the criterion. Clinical diagnoses. Clinical diagnoses obtained from nonstandardized clinical evaluations conducted by attending psychia trists at the time of discharge provided a measure of Criterion F. Individuals diagnosed with any schizophrenia spectrum disorder, delirium, substance intoxication or withdrawal, or any develop mental or intellectual disorder were identified as not meeting Criterion F. Alexian Brothers Urge to Self-Injure scale (ABUSI). The ABUSI is a five-item self-report scale assessing the degree to which an individual is motivated or impelled to self-injure. Re sponses are on a 7-point scale with higher scores reflecting more intense urges to self-injure. The ABUSI has adequate reliability and validity (Washburn, Juzwin, Styer, & Aldridge, 2010) and was included in this study as an existing measure of NSI characteris tics. Behavioral and Symptom Identification Scale (BASIS-24) total score. The BASIS-24 is a self-report inventory which as sesses symptoms of psychopathology across six domains: depression/functioning, interpersonal problems, self-harm, emotional lability, psychosis, and alcohol/drug use. The BASIS-24 has adequate reliability and validity (Eisen, Gerena, Ranganathan, Esch, & Idiculla, 2006). The BASIS-24 total score ranges from 0 to 4 (higher scores indicate greater psychopathology) and is obtained using a weighted sum derived by multiplying each item’s rating by the item’s weight and then totaling all 24 of the weighted ratings. Weights provide a linear approximation of the BASIS-24 scores obtained from item response theory (IRT) methods (McLean Hospital, 2006). The BASIS-24 total score was included in this study as a criterion measure of overall psychopathology. Behavioral and Symptom Identification Scale (BASIS-24) suicidal ideation. The BASIS-24 includes a two-item self-harm subscale, one of which specifically assesses for suicidal ideation (“Think about ending your life?”). Unweighted ratings on this item were used, ranging from 0 (None o f the time) to 4 (All o f the time). The BASIS-24 suicide ideation item was included in this study as a criterion measure of suicidal ideation. Borderline Evaluation of Severity over Time (BEST). The BEST is a 15-item self-report measure used to measure the sever
ity of thoughts, emotions, and behaviors typical of borderline personality disorder (BPD). The BEST has adequate reliability and validity (Pfohl et al., 2009). The BEST total score was used in this study, which sums the two pathology subscales minus the positive behavior subscale, with higher scores reflecting greater BPD pa thology. The BEST total score was included in this study as a criterion measure of borderline personality traits. Quality of Life Enjoyment and Satisfaction scale—short form (QLES). The QLES (Endicott, Nee, Harrison, & Blumenthal, 1993) is a self-report scale designed to assess levels of enjoyment and satisfaction experienced over the past week in the following areas of functioning: physical health, mood, school/ learning, interpersonal relationships, economic status, and ability to complete work/hobbies. Responses are rated on a 5-point scale ranging from Very poor to Very good. The short form of the QLES, which comprises the first 14 items assessing the aforementioned various areas of daily functioning, were summed to compute a maximum score that is a proportion (0%-100%) of the maximum possible score of 70. The QLES has adequate reliability and validity (Endicott, Nee, Yang, & Wohlberg, 2006; Rapaport, Clary, Fayyad, & Endicott, 2005). The QLES maximum score was included in this study as a criterion measure of general quality of life. Work and Social Adjustment Scale (WSAS). The WSAS is a five-item self-report measure designed to assess functional im pairment associated with an identified disorder or problem. Re sponses are rated on an 8-point scale ranging from No impairment to Very severe impairment. The WSAS has adequate reliability and validity (Mataix-Cols et al., 2005; Mundt, Marks, Shear, & Greist, 2002). The WSAS total score, ranging from 0 to 40, was included in this study as a criterion measure of functional impairment. D ata A nalysis To examine the factor structure of the ABASI, we submitted a polychoric correlation matrix of the seven ABASI items assessing the core NSI disorder criteria (Criteria B and C) to an exploratory factor analysis using maximum likelihood factoring and oblique (promax) rotation. In addition to Kaiser’s eigenvalue of 1 criterion and examination of the scree plot, Horn’s parallel analysis with Glorfeld’s Monte Carlo extension (Dinno, 2009; Glorfeld, 1995; Horn, 1965) was used to identify the optimal number of factors to retain for the final factor solution. Through generation of a large number of random data sets based on the number of subjects and variables of the target data set, parallel analysis compares the eigenvalues from the random data sets with the target data set to determine the optimal number of factors to retain. For the current parallel analysis, the critical eigenvalue was set at the 95th per centile and 1,000 random permutation data sets were generated by matching distributions from the raw data of the seven variables of the ABASI. Test-retest reliability of NSI disorder was examined using kappa coefficients. For NSI severity and the subscales of NSI disorder, interclass correlation coefficients (ICC) using a two-way random effects model with absolute agreement were used to ex amine test-retest reliability. To examine the utility of NSI disorder in identifying a more severe presentation of NSI, we examined group differences on measures of psychopathology and impair ment using t test statistics and effect size (Cohen’s d) estimates,
NONSUICIDAL SELF-INJURY DISORDER
with effect size confidence intervals estimated using 1,000 boot strap replications. Hierarchical multiple regressions were used to examine the incremental validity of NSI disorder when compared to existing measures of NSI characteristics (i.e., frequency of NSI, number of methods of NSI, urge to self-injure, age of onset of NSI, and duration of NSI). To explore the benefit of the ABASI as a dimensional measure of NSI severity, we entered the NSI severity score, an ordinal scale representing the summation of the seven ABASI items assessing the Criteria B and C of NSI disorder, into the hierarchical multiple regressions instead of NSI disorder. Vari ables created from a summation of Likert-scale responses were assumed to approximate an interval scale and were treated as such in analyses. For suicidal ideation, an ordinal regression was also performed to determine if results differed substantially when treat ing the dependent measure as an ordinal-level variable; no sub stantive differences in results were observed, and the original results were retained. The alpha level was adjusted to .01 to account for multiple comparisons. All analyses were conducted using Stata Statistical software (Release 13; StataCorp, 2013).
Results Psychometrics Results of the parallel analysis indicated that the eigenvalues for Factor 1 (2.78) and Factor 2 (1.01) were significantly higher than those expected by chance. Based on the parallel analysis, Kaiser’s criterion, and examination of the scree plot, a two-factor solution was retained, explaining 89.6% of the variance. As shown in Table 2, all three items assessing Criterion B loaded on Factor 1 and all four items assessing Criterion C loaded on Factor 2. Criterion B and Criterion C factors were significantly correlated, r(509) = .32, p < .001. Internal consistency, as measured by Cronbach’s alpha, was a - .75 for all seven items, with a = .79 for the Criterion B subscale and a = .65 for the Criterion C subscale. Despite ade quate internal consistency, corrected item-total correlations show that the ABASI item assessing Symptom B-3 is only weakly associated with the NSI severity score (see Table 3). Test-retest reliability using kappa coefficients and ICCs were calculated for 95 patients who were re-administered the ABASI between 1 and 14 days (M = 6.8, SD = 3.1) due to transfer between levels of care (e.g., from partial hospitalization to inpa tient). Test-retest reliability was moderate for the full NSI disorder
35
( k = .432), and good specifically for Criterion A ( k = .784) and Criterion C ( k = .662), but poor for Criterion B ( k = .216). Test-retest reliability was good for the ABASI NSI severity score (ICC = .606) and was moderate for the Criterion B (ICC = .484) and Criterion C (ICC = .517) subscales of the ABASI.
NSI Disorder Criteria Most, but not all, patients met Criterion A, with 85.5% of patients self-injuring 5 or more days in the past year with at least one specific type of NSI behavior. Across all types NSI, patients self-injured a mean of 76.8 (SD = 102.7) days in the past year, with a mean of 9.7 (SD = 14.3) separate incidents for each day injured. Patients used a mean of 4.0 (SD = 3.0) different types of NSI in the past year. Mean age of onset was 12.5 years old (SD = 4.0), with 96.6% reporting onset of NSI prior to age 18 years. Among the 14.5% (n = 74) of patients who reported fewer than 5 days for a specific type of NSI behavior in the past year, the mean number of days of NSI was 1.9 (SD = 1.3), with 2.7% of the full sample reporting no NSI in the last year. Given the ABASI likely underreported total NSI in the past year, the total number of NSI days reported across all types of NSI was summed to deter mine the maximum range of possible patients meeting Criterion A. At most, 21 additional patients were eligible for Criterion A, making the uppermost estimate of Criterion A 89.6% of the sam ple. As shown in Table 4, cutting was the most common type of NSI, followed by scratching, skin carving, food restriction, binging/purging, and hitting oneself. Because the ABASI in cludes 10 behaviors that may not necessarily result in damage to the surface of the body, it is possible that the ABASI overestimated the rate for Criterion A. To address this potential, the following 10 behaviors were removed from the algorithm creating Criterion A for NSI disorder: tattooing/piercing, selfchocking, pulling hair/eyebrow/eyelash, falling down, fighting, ingesting/swallowing objects, overexercising, binging/purging/ restricting food intake, sexual behaviors, and ignoring medical advice or making medical conditions worse. Removing these 10 behaviors reduced the proportion of those meeting Criterion A by 2.5%, making the lowermost estimate of Criterion A 83.0% of the sample. Using the uppermost estimate, which assumes different types of NSI behavior occurred on different days, and the lowermost estimate, which provides a conservative listing of only 11 NSI behaviors, the true rate of Criterion A is likely
Table 2 Rotated Factor Loadings fo r Criteria A and B o f the Alexian Brothers Assessment o f Self-Injury (ABASI) ABASI item (DSM criterion)
Factor 1
Factor 2
Relief from negative feelings or thoughts (B-l) Fix or resolve problems with other people (B-2) Create or increase positive feelings (B-3) Negative feelings or thoughts immediately before self-injuring (C-l) Difficulties or problems with other people immediately before self-injuring (C-l) Strong desire or urge to hurt yourself that was difficult to resist before self-injuring (C-2) Think about hurting yourself (C-3)
0.14 -0.06 -0.06 0.74 0.54
0.63 0.91 0.51 -0.03 0.14
0.89 0.74
-0.02 -0.04
Note.
DSM = Diagnostic and Statistical Manual o f Mental Disorders.
WASHBURN, POTTHOFF, JUZWIN, AND STYER
36 Table 3
Corrected Item-Total Correlations o f Nonsuicidal Self-Injury Severity Score (NSI) NSI Criteria Criteria severity B C
ABASI item (DSM criterion) Relief from negative feelings or thoughts (B-l) Fix or resolve problems with other people (B-2) Create or increase positive feelings (B-3) Negative feelings or thoughts immediately before self-injuring (C-l) Difficulties or problems with other people immediately before self-injuring (C-l) Strong desire or urge to hurt yourself that was difficult to resist before self-injuring (C-2) Think about hurting yourself (C-3)
0.48 0.46 0.26 0.51 0.48
0.49 0.54 0.39 0.62 0.51
0.61 0.52
0.71 0.57
Note. ABASI = Alexian Brothers Assessment of Self-Injury; DSM = Diagnostic and Statistical Manual of Mental Disorders.
between 83.0% and 89.6%. For the remaining estimates and analyses, the original estimate of 85.5% for Criterion A was used because it represents a reasonable mid-point between the lowermost and uppermost estimates. Sensitivity analyses were conducted on all subsequent analyses; no differences were found in the results when either the lowermost or uppermost estimates were used in place of the original estimate of Crite rion A (analyses available in the online supplementary materi als). Most, but not all, patients met Criterion B, with 87.7% indicating at least one expectation of NSI. Specifically, most patients expected that NSI would bring relief from either a negative feeling or cognitive state (82.0%) or resolution of an interpersonal problem (57.1%). Slightly more than one third of the sample expected that NSI would create or enhance a posi tive feeling state (34.8%). The majority of patients (61.8%)
reported two or more expectations of NSI, with nearly one quarter (24.5%) reporting all three expectations of NSI. Nearly all of the patients met Criterion C, with 92.8% of patients indicating either a precipitant to NSI or thoughts about NSI. Specifically, nearly all patients reported interpersonal problems or negative thoughts or feelings prior to NSI (91.3%), whereas most reported a preoccupation or urge to self-injure prior to NSI (71.6%) and thinking about harming themselves (72.8%). The majority of patients (80.2%) reported two or more symptoms within Criterion C, with 62.4% reporting all three symptoms. Nearly all patients met Criteria E and F, with 98.2% of patients receiving GAF scores at 50 or below, and only four patients (0.8%) having a clinical diagnosis than could exclude them from NSI disorder (three patients had an autism spectrum disorder diagnosis, and one had a drug withdrawal diagnosis).
Table 4
Characteristics o f Nonsuicidal Self-Injury (NSI) Behaviors Assessed by the Alexian Brothers Assessment of Self-Injury (ABASI) No. of days
Times per day
Age of onset
NSI behaviors
n
%
M
SD
M
SD
M
SD
Banged head Binged/purged/starved for pain/harm8 Broke limbs Burned skin Carved skin for pain/harm Choked self® Cut skin Drawn blood Fell down for pain/harm8 Fought for pain/harma Gouged skin Hit self Ingested objects for pain/harma Inserted/embedded objects Over-exercised for pain/harm8 Prevented healing of injuries Pulled hair/eyelashes/eyebrows for pain/harm0 Scratched skin Sexual behavior for pain/harm8 Tattooed for pain/harm8 Worsened medical condition for pain/harm0
511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511
23.5 29.9 1.2 26.4 32.5 14.9 93.0 9.4 2.5 3.5 1.4 29.7 2.7 5.5 7.4 26.4 13.7 56.0 5.1 11.9 4.3
5.27 24.54 0.02 3.10 6.62 0.75 52.13 4.25 0.17 0.34 0.12 8.07 0.11 1.59 3.08 13.84 6.23 25.22 1.06 0.50 0.72
24.90 70.06 0.20 15.70 31.99 3.38 77.99 27.38 3.11 4.55 1.84 30.74 1.02 17.90 21.83 49.49 35.06 65.32 16.08 2.19 6.44
0.58 0.71 0.01 0.38 0.43 0.18 2.86 0.17 0.03 0.04 0.02 0.85 0.04 0.14 0.11 0.84 0.42 1.52 0.09 0.20 0.07
2.03 1.49 0.11 0.98 0.89 0.52 6.30 0.82 0.16 0.20 0.14 2.85 0.27 0.91 0.55 4.78 1.76 3.63 0.59 1.17 0.40
14.54 13.70 15.60 14.21 14.02 14.01 13.27 12.87 13.25 13.00 13.67 13.79 20.31 12.93 13.70 12.66 13.23 13.51 17.77 14.97 15.43
5.13 2.90 3.58 2.85 3.88 4.21 3.27 1.78 1.76 1.71 2.16 4.75 12.85 2.66 2.12 2.44 3.54 4.87 8.32 3.82 3.60
Note. ABASI = Alexian Brothers Assessment of Self-Injury. a Behaviors considered to be self-injurious in the ABASI, but may not result in damage to the surface of the body.
NONSUICIDAL SELF-INJURY DISORDER
37
Rates of NSI Disorder
Incremental Criterion Validity of NSI Disorder
Nearly three quarters (74.0%) of the sample met all criteria for NSI disorder, as described in the DSM-5. If the uppermost esti mate of 89.9% for Criterion A is used to identify NSI disorder, the rate of NSI disorder increases to 76.9%. In contrast, if the lower most estimate of 83.0% for Criterion A is used, the rate of NSI decreases to 71.8%. No significant differences were found in the proportion of patients meeting NSI disorder criteria by age, gen der, or ethnicity. As shown in Table 5, regardless of NSI disorder status, the ABASI item assessing C-l was the most commonly endorsed symptom, with all of the patients meeting criteria for NSI disorder endorsing this item. In contrast, the item assessing Symptom B-3 was the least commonly endorsed symptom. The item assessing Symptom B-3 was also the least useful in meeting criteria for NSI disorder; only seven patients with NSI disorder met Criterion B through the item assessing B-3 alone. It was also rare for patients to meet Criterion B without also meeting Criterion C. Across the entire sample, only five patients who met Criterion B did not meet Criterion C. In contrast, 62 patients who met Criterion C did not meet Criterion B. Among patients who did not meet criteria for NSI disorder, 82.7% met all but one of the NSI disorder criteria, most commonly failing to meet either Criterion A (48.2%) or Criterion B (40.0%); very few of these patients failed to meet Criterion C (1.8%).
As shown in Table 7, commonly assessed characteristics of NSI explained between 17% and 49% of the variance across the crite rion measures of psychopathology and impairment. Among the characteristics of NSI, urge to self-injure was significantly asso ciated with all the criterion measures. The number of NSI methods was also significantly associated with overall psychopathology, borderline personality traits, and functional impairment. In con trast, frequency of NSI was only associated with borderline per sonality traits, and neither age of onset of NSI nor duration of NSI was associated with any of the criterion measures. Adding NSI disorder as a dichotomous variable to the model (see Step 2a in Table 7) failed to account for any additional variance in the criterion variables. In contrast, a model including NSI severity instead of NSI disorder in Step 2 accounted for additional variance beyond the characteristics of NSI for three of the criterion measures (see Step 2b in Table 7). Specifically, NSI severity accounted for an additional 2% of the variance for overall psychopathology, an additional 1% of variance for borderline personality traits, and an additional 1% of variance for quality of life. A model including the items assessing the specific criteria of NSI disorder instead of either NSI Disorder or NSI severity in Step 2 was also examined. The item assessing Symptom C-l was significantly associated with all of the criterion measures of psy chopathology and impairment, except for suicidal ideation. For suicidal ideation, only the item assessing C-3 remained signifi cantly associated after accounting for the characteristics of NSI. Given the conceptual and empirical overlap between the ABUSI and the item assessing C-2 (/is = .55, p < .001), we repeated the analyses with the ABUSI total score removed from Step 1. The item assessing C-2 was not found to account for any significant variance across any of the criterion measure, even with the ABUSI total score removed from the model.
Correlates of NSI Disorder Table 6 summarizes the differences between patients with and without NSI disorder on common characteristics of NSI, as well as criterion measures of psychopathology and impairment. In contrast to those without NSI disorder, those with NSI disorder experienced significantly greater urges to self-injure, more days of self-injury in the past year, and a greater number of methods used to selfinjure in the past year. No differences between those with and without NSI disorder were found on age of onset or duration of NSI. NSI disorder was also associated with significantly worse overall psychopathology, suicidal ideation, traits of borderline personality disorder, quality of life, and functional impairment.
Table 5 Endorsement o f Specific NSI Disorder Criteria by NSI Disorder Status NSI disorder Criteria Criterion Criterion Criterion Criterion Criterion Criterion Criterion Criterion Criterion Note.
A B1 B2 B3 Cl C2 C3 E F
Positive for (%)
Negative for (%)
100.0 93.4 67.5 41.3 100.0 96.8 98.4 100.0 100.0
44.4 49.6 27.8 16.5 92.5 74.4 86.5 93.2 97.0
NSI = Nonsuicidal Self-Injury.
Discussion The ABASI adds to a limited but growing body of literature examining the utility of a criterion-based approach to identifying clinically significant NSI. This study demonstrates that NSI dis order can be reliably identified among people who self-injure using a brief self-report instrument. The findings also indicate that NSI disorder, as identified by the ABASI, represents a clinically significant level of NSI. NSI disorder characterizes nearly three quarters of adolescents and adults being treated for NSI at an acute level of care and is indicative of a more severe case presentation when compared with simply identifying self-injurious behavior as a clinical concern. Specifically, NSI disorder is associated with significantly higher frequency of NSI, a greater number of NSI methods, a stronger urge to self-injure, more severe psychopathol ogy, greater suicidal ideation, more borderline personality traits, and more impaired quality of life and functioning than found in people who self-injure but do not meet criteria for NSI Disorder. As a dichotomous diagnosis, however, NSI disorder appears to have limited utility above and beyond NSI characteristics that are commonly assessed in the literature. Results of the hierarchical regression analyses indicate that NSI disorder fails to account for any additional variance in criterion measures of psychopathology and impairment beyond the urge to self-injure, number of NSI methods, and frequency of NSI. In contrast, an ordinal level NSI
WASHBURN, POTTHOFF, JUZWIN, AND STYER
38
Table 6 Correlates o f Nonsuicidal Self-Injury Disorder (NSI) NSI disorder Positive for Criterion NSI characteristics Frequency Methods ABUSI total score Onset Duration Psychopathology BASIS total Suicidal ideation BEST total Impairment QLES WSAS
Negative for
M
SD
M
SD
Cohen’s d
88.72 4.29 21.06 12.52 4.72
104.80 2.78 7.86 3.53 6.23
42.91 3.21 16.83 13.07 3.89
88.31 3.40 8.62 4.21 3.87
0.45 0.37 0.52 0.15 0.14
[.27, [.14, [.32, [.08, [.02,
.64] .60] .73] .37] .31]
© 2014 American Psychological Association 1040-3590/15/$ 12.00 http://dx.doi.org/! 0 .1037/pas0000021
Assessing DSM-5 Nonsuicidal Self-Injury Disorder in a Clinical Sample Jason J. Washburn
Lauren M. Potthoff
Alexian Brothers Behavioral Health Hospital, Hoffman Estates, Illinois, and Northwestern University Feinberg School of Medicine
Northwestern University Feinberg School of Medicine
K. R. Juzwin
Denise M. Styer
Alexian Brothers Behavioral Health Hospital, Hoffman Estates, Illinois, and Illinois School of Professional Psychology, Schaumburg
Alexian Brothers Behavioral Health Hospital, Hoffman Estates, Illinois
The entry for nonsuicidal self-injury (NSI) disorder in the 5th edition of the Diagnostic and Statistical Manual o f Mental Disorders (DSM-5) provides a criterion-based definition of clinically relevant NSI. NSI disorder is currently classified in the DSM-5 as a condition requiring further study. The present study aimed to examine the reliability, validity, and clinical utility of a self-report measure of NSI disorder, the Alexian Brothers Assessment of Self-Injury (ABASI). The sample included 511 patients admitted to an acute care treatment program designed to treat NSI. Patients were administered the ABASI as part of a clinical assessment and routine outcome evaluation. The sample included a broad age range, as well as sufficient numbers of males and Hispanics to examine sociodemographic differences. The ABASI demonstrated adequate internal consistency and test-retest reliability, and the factor structure reflects NSI disorder criteria. Among patients being treated for NSI, 74% met criteria for NSI disorder. No differences in the rate of NSI disorder were observed by sex, ethnicity, or age. Although NSI disorder is associated with a worse presentation of self-injurious behavior, NSI disorder provides limited clinical utility as a dichotomous diagnosis, at least when compared with common NSI characteristics such as number of methods of NSI and the urge to self-injure. Instead, findings support a dimensional approach to NSI disorder. Analyses of specific symptoms of NSI disorder indicate concerns with Criterion B as currently defined by the DSM-5. Recommendations for a more parsimonious revision of NSI disorder are discussed.
Keywords: self-injury, DSM-5, psychometrics, clinical utility Supplemental materials: http://dx.doi.org/! 0.1037/pas0000021.supp
rates of at least one lifetime incidence of NSI range between 7.6% to 23.2% among children, adolescents, and emerging adults (Barrocas, Hankin, Young, & Abela, 2012; Jacobson & Gould, 2007; Whitlock, Eckenrode, & Silverman, 2006), dropping to 5.9% among adults (Klonsky, 2011). It remains unclear, however, if these rates reflect clinically relevant NSI given that there is no agreed-upon definition of when NSI requires clinical attention. The entry for NSI disorder in the recently published fifth edition of the Diagnostic and Statistical Manual o f Mental Disorders (DSM-5) provides a criterion-based definition of clinically relevant NSI (Amer ican Psychiatric Association, 2013; see Table 1 for specific criteria). NSI disorder was recently found to reflect the behavior of a “prototypic” self-injurer by 119 researchers and clinicians with expertise in NSI (Lengel & Mullins-Sweatt, 2013). Despite general agreement on the criteria, NSI disorder was classified in the D SM -5 as a condition requiring further study due to insufficient evidence supporting its designation as an “official mental disorder” (American Psychiatric Association, 2013). One concern with NSI disorder surfaced from the D SM -5 field trials, which failed to show acceptable interrater reli ability using clinician-administered diagnostic interviews (Regier et al., 2013). Those conducting the D SM -5 field trials, however, strug-
Nonsuicidal self-injury (NSI) refers to the direct and deliberate infliction of damage to one’s bodily tissue without suicidal intent and includes a variety of behaviors such as cutting, burning, scratching skin, biting, and hitting oneself (International Society for the Study of Self-injury, 2007). As an isolated set of behaviors,
This article was published Online First September 29, 2014. Jason J. Washburn, Center for Evidence-Based Practice, Alexian Broth ers Behavioral Health Hospital, Hoffman Estates, Illinois, and Department of Psychiatry and Behavioral Sciences, Northwestern University Feinberg School of Medicine; Lauren M. Potthoff, Department of Psychiatry and Behavioral Sciences, Northwestern University Feinberg School of Medi cine; K. R. Juzwin, Self-Injury Recovery Services, Alexian Brothers Be havioral Health Hospital, Hoffman Estates, Illinois, and Illinois School of Professional Psychology, Schaumburg; Denise M. Styer, Self-Injury Re covery Services, Alexian Brothers Behavioral Health Hospital, Hoffman Estates, Illinois. Correspondence concerning this article should be addressed to Jason J. Washburn, Center for Evidence-Based Practice, Alexian Brothers Behav ioral Health Hospital, 1650 Moon Lake Blvd., Hoffman Estates, IL 60302. E-mail: [email protected] [email protected]
31
32
WASHBURN, POTTHOFF, JUZWIN, AND STYER
Table 1 DSM-5 Criteria fo r Nonsuicidal Self-Injury (NSI) Disorder and Corresponding Methods o f Measurement in the Current Study DSM-5 description
Methods of measurement
Criteria A. In the last year, the individual has, on 5 or more days, engaged in intentional self-inflicted damage to the surface of his or her body of a sort likely to induce bleeding, bruising, or pain with the expectation that the injury will lead to only minor or moderate physical harm (i.e., there is no suicidal intent).
Met Criterion A if reported engaging in 5 or more days of a specific NSI method using the following question on the ABASI: “In the last year, how often did you intentionally (on purpose) do the following things to hurt yourself and/or cause pain?” See Table 2 for a list of the specific NSI behaviors included in the ABASI. The ABASI does not assess the total number of days of NSI across all types. Consequently, only those who reported 5 or more days of NSI within a specific type of NSI were included as meeting Criterion A. Met Criterion B if responded with Agree or Strongly agree to at least one of the following items: “When I self-injure, I expect that it will . . .” 1. Provide relief from negative feelings or thoughts. 2. Fix or resolve problems with other people. 3. Create or increase positive feelings (happy, joyful, excited, cheerful, etc.). Met Criterion C if responded with Some o f the time, Half o f the time, Most o f the time, or All o f the time to at least one of the following items: “In the past 12 months, how often did you . . .” la. Have negative feelings or thoughts (distress, anger, sadness, anxiety, self-criticism, etc.) immediately before self-injuring. lb. Have difficulties or problems with other people immediately before self-injuring. 2. Experience a strong desire or urge to hurt yourself that was difficult to resist before self-injuring. 3. Think about hurting yourself.
Criteria B. The individual engages in the self-injurious behavior with one or more of the following expectations: 1. To obtain relief from a negative feeling or cognitive state. 2. To resolve an interpersonal difficulty. 3. To induce a positive feeling state. Criteria C. The intentional self-injury is associated with at least one of the following:
1. Interpersonal difficulties or negative feelings or thoughts, such as depression, anxiety, tension, anger, generalized distress, or selfcriticism, occurring in the period immediately prior to the selfinjurious act. 2. Prior to engaging in the act, a period of preoccupation with the intended behavior that is difficult to control. 3. Thinking about self-injury that occurs frequently, even when it is not acted upon. Criteria D. The behavior is not socially sanctioned (e.g., body piercing, tattooing, part of a religious or cultural ritual) and is not restricted to picking a scab or nail biting. Criteria E. The behavior or its consequences cause clinically significant distress or interference in interpersonal, academic, or other important areas of functioning. Criteria F. The behavior does not occur exclusively during psychotic episodes, delirium, substance intoxication, or substance withdrawal. In individuals with a neurodevelopmental disorder, the behavior is not part of a pattern of repetitive stereotypes. The behavior is not better explained by another mental disorder or medical condition
Picking scabs and nail biting were excluded as NSI behaviors, and the phrase “to hurt yourself or cause pain” was added to clarify any behavior that may be used for purposes other than selfinjury (See Table 2) Met Criterion E if rated as 50 or below on the Global Assessment of Functioning. Met Criterion F if not clinically diagnosed with any schizophrenia spectrum disorder, delirium, substance intoxication or withdrawal, or any developmental or intellectual disorder.
Note. Description of nonsuicidal self-injury disorder from Diagnostic and Statistical Manual o f Mental Disorders (5th ed., DSM-5), by the American Psychiatric Association, 2013. Copyright 2013 by American Psychiatric Association. Reprinted with permission. ABASI = Alexian Brothers Assessment of Self-Injury.
gled to obtain sufficient samples sizes for NSI disorder to generate accurate kappa estimates; one site failed to recruit any patients, and the other two sites recruited only seven pediatric patients each (Clarke et al„ 2013). Acceptable interrater reliability (k = .90) of NSI disorder using a semistructured interview was recently demonstrated with a sam ple of 73 adolescent females recruited from inpatient units in Switzerland and Germany (In-Albon, Ruf, & Schmid, 2013). The interview questions, however, were based on the initial proposed criteria presented on the DSM website before publication of the DSM-5. Additional recent studies provide further support for the proposed NSI disorder criteria; however, they also failed to assess all of the current DSM-5 NSI criteria (Glenn & Klonsky, 2013; Odelius & Ramklint, in press; Selby, Bender, Gordon, Nock, & Joiner, 2012; Zetterqvist, Lundh, Dahlstrom, & Svedin, 2013). This study builds upon the available literature by examining the reliability and clinical utility of a self-report measure in assessing
the current criteria of NSI disorder proposed in Section III of the DSM-5. Using a large sample of people admitted to an acute care program designed specifically to treat NSI, this study also ad dresses the limitations of prior studies by including adequate numbers of males, adolescents and adults, and Hispanics for ex amination of potential sociodemographic differences. Finally, by examining NSI disorder in a sample of people being treated specifically for NSI, the current study contributes to the evaluation of the clinical utility of the NSI disorder criteria proposed in the DSM-5.
Method Sample The sample included 511 patients consecutively admitted to an acute care program designed to treat NSI. Approximately one third
NONSUICIDAL SELF-INJURY DISORDER
(33.2%) of the sample received inpatient treatment at the time of assessment, with the remaining sample receiving a combination of partial hospitalization and intensive outpatient day treatment. The sample was predominantly female (90.0%) and non-Hispanic W hite (83.2%), with some representation of Hispanics (9.5%), but minimal representation of African Americans (1.8%) and other race/ethnicities (2.0%); race/ethnicity data were not available for 3.5% of the sample. Ages ranged from 12 to 52 years old, with a mean age of 17.3 years (SD = 6.2); the majority o f the patients were either adolescents (72.0%) or emerging adults (15.5%). So cioeconomic status, which was estimated by examining the pro portion of publically funded health insurance (i.e., Medicaid) in the sample and by matching patients’ geocodes to median household incomes reported in the 2011 American Community Survey (U.S. Census Bureau, 2011), indicates an economically diverse sample, with 24.8% receiving Medicaid and a mean household income of $78,327 (SD = $20,727). Nearly the entire sample (95.0%) received a primary diagnosis of an affective disorder from attending psychiatrists using a nonstandardized clinical evaluation. Diagnoses included major depres sive disorder (52.6%), depressive disorder not otherwise specified (NOS; 18.0%), mood disorder NOS (17.5%), and bipolar spectrum disorder (6.9%). Nonaffective primary diagnoses included eating disor ders (2.1%), anxiety disorders (1.7%), and impulse control disorder (0.4%). Partly due to the age of the sample, personality disorders were not routinely diagnosed by the attending psychiatrists and therefore not included in this study.
Procedures The measures used in this study were administered to patients at the time of admission to the treatment program. Patients completed self-report measures as part of routine clinical assessment and clinical outcome monitoring. Completed measures were main tained in archived clinical outcome databases and combined with sociodemographic and clinician-generated data from the medical record for internal organizational improvem ent purposes. N SI dis order was identified by combining the self-report measures with clinician-generated data, as described in Table 1. All individuallevel data were de-identified before analyses were conducted, following the federally defined de-identification standard for pro tected health information. Procedures for this research study were reviewed by both the hospital and university systems’ institutional review boards (IRBs) and found to be exempt from further IRB review pursuant to the Protection o f Human Research Subjects Policy (2009) of the U. S. Department of Health and Human Services.
Instrumentation Alexian Brothers Assessment of Self-Injury (ABASI). The ABASI was created as a self-report clinical assessment in antici pation of NSI disorder being included in the DSM- 5. Specifically, the ABASI was designed to identify NSI disorder in a population o f individuals already identified as engaging in NSI behavior. As such, the ABASI is an assessment o f clinical severity o f NSI, rather than of NSI behavior per se. Criterion A was assessed by asking patients to record the num ber o f days in the past year in which they engaged in each of 21
33
specific NSI behaviors. In addition to typical NSI behaviors that result in damage to the surface of the body, 10 o f the 21 NSI behaviors focused on inducement o f pain or harm below the surface o f the body. Although not consistent with Criterion A in the DSM - 5, these 10 behaviors were included in the ABASI because o f the researchers’ clinical observation that they are used as form s o f self-injury. B ecause the A BA SI did not include a question asking about the total num ber o f days self-injured across all types o f N SI, only patients who reported 5 or more days o f N SI w ithin a specific type o f NSI were included as m eeting C riterion A. Criterion B was assessed with a set o f three items, rated on a 5-point (0 - 4 ) Likert scale from Strongly disagree to Strongly agree, using the lead, “W hen I self-injure, I expect that it w ill. . for each o f the following items: “Provide relief from negative feelings or thoughts” (B -l); “Fix or resolve problems with other people” (B-2); and “Create or increase positive feelings (happy, joyful, excited, cheerful, etc.)” (B-3). Responses o f Agree or Strongly agree to any of these three items were counted as meeting Criterion B. The 5-point responses to the three items in Criterion B were also summed to create a total score for Criterion B ranging from 0 to 12. Criterion C was assessed with a set o f four items rated on a 5-point (0 - 4 ) Likert scale from None o f the time to All o f the time, using the lead, “In the past year (12 months), how often did you . . .” for each of the following items: “Have negative feelings or thoughts (distress, anger, sadness, anxiety, tension, self-criticism, etc.) immediately before self-injuring” (C -la); “Have difficulties or problems with other people immediately before self-injuring” (C -lb); “Experience a strong desire or urge to hurt yourself that was difficult to resist before self-injuring” (C-2); and “Think about hurting yo u rself’ (C-3). For Symptom C -l, the ABASI separates interpersonal and intrapersonal preceding problems into two sep arate items, consistent with findings differentiating interpersonal from intrapersonal functions of NSI (Klonsky, 2007). Responses of Some o f the time, H alf o f the time, Most o f the time, or All o f the time to any o f these four items were counted as meeting Criterion C. To maintain consistency with the NSI disorder criteria, we used the average o f the two items assessing Symptom C -l for any specific analysis of Symptom C-1. Further, the average of the two items assessing Symptom C -l was summed with the 5-point re sponses from the items assessing Criteria C-2 and C-3 to create a total score for Criterion C ranging from 0 to 12. The total scores for Criteria B and C were summed to create an NSI severity score ranging from 0 to 24. To address Criterion D, we did not include scab picking and nail biting as specific behaviors in Criterion A. Further, the phrase “to hurt yourself and/or cause pain” was added to the descriptions of the following NSI behaviors in Criterion A: tattooing/piercing, pulling hair/eyelashes/eyebrows, falling down, carved into skin, fighting, ingested/swallowed items, overexercised, restricted eating/purged/use laxatives, sexual behavior, and not following med ical advice or worsening medical condition. The phrase was added to clarify that the behaviors should only be endorsed if participants engaged in them for purposes o f self-injury, not for socially sanctioned purposes or purposes other than self-injury. For exam ple, the ABASI considers tattooing a form of self-injury when it is engaged in specifically to cause pain or harm (Stirn & Hinz, 2008).
34
WASHBURN, POTTHOFF, JUZWIN, AND STYER
The ABASI also provided data on the characteristics of NSI that are similar to measures already existing in the literature to identify more severe cases of NSI, such as frequency of NSI, number of methods of NSI, age of onset of NSI, and duration of NSI. Frequency of NSI was assessed by taking the average number of days of NSI in the past year across all 21 types of NSI behaviors. Number of methods of NSI was assessed by counting the number of NSI behaviors that were endorsed at least 1 day in the past year. Age of onset was assessed by asking patients to estimate the age of onset for each specific NSI behavior they engaged in during the past year, and taking the youngest age. Duration of NSI was assessed by subtracting patients’ current age from the age of onset of NSI. Global Assessment of Functioning (GAF). The GAF (Amer ican Psychiatric Association, 2000), a measure of overall symp toms and impairment with a range from 1 to 100, was obtained from licensed clinicians conducting a level of care assessment at the time of admission to the acute care treatment program. The GAF provided a measure of Criterion E, with a threshold of 50 or below as meeting the criterion. Clinical diagnoses. Clinical diagnoses obtained from nonstandardized clinical evaluations conducted by attending psychia trists at the time of discharge provided a measure of Criterion F. Individuals diagnosed with any schizophrenia spectrum disorder, delirium, substance intoxication or withdrawal, or any develop mental or intellectual disorder were identified as not meeting Criterion F. Alexian Brothers Urge to Self-Injure scale (ABUSI). The ABUSI is a five-item self-report scale assessing the degree to which an individual is motivated or impelled to self-injure. Re sponses are on a 7-point scale with higher scores reflecting more intense urges to self-injure. The ABUSI has adequate reliability and validity (Washburn, Juzwin, Styer, & Aldridge, 2010) and was included in this study as an existing measure of NSI characteris tics. Behavioral and Symptom Identification Scale (BASIS-24) total score. The BASIS-24 is a self-report inventory which as sesses symptoms of psychopathology across six domains: depression/functioning, interpersonal problems, self-harm, emotional lability, psychosis, and alcohol/drug use. The BASIS-24 has adequate reliability and validity (Eisen, Gerena, Ranganathan, Esch, & Idiculla, 2006). The BASIS-24 total score ranges from 0 to 4 (higher scores indicate greater psychopathology) and is obtained using a weighted sum derived by multiplying each item’s rating by the item’s weight and then totaling all 24 of the weighted ratings. Weights provide a linear approximation of the BASIS-24 scores obtained from item response theory (IRT) methods (McLean Hospital, 2006). The BASIS-24 total score was included in this study as a criterion measure of overall psychopathology. Behavioral and Symptom Identification Scale (BASIS-24) suicidal ideation. The BASIS-24 includes a two-item self-harm subscale, one of which specifically assesses for suicidal ideation (“Think about ending your life?”). Unweighted ratings on this item were used, ranging from 0 (None o f the time) to 4 (All o f the time). The BASIS-24 suicide ideation item was included in this study as a criterion measure of suicidal ideation. Borderline Evaluation of Severity over Time (BEST). The BEST is a 15-item self-report measure used to measure the sever
ity of thoughts, emotions, and behaviors typical of borderline personality disorder (BPD). The BEST has adequate reliability and validity (Pfohl et al., 2009). The BEST total score was used in this study, which sums the two pathology subscales minus the positive behavior subscale, with higher scores reflecting greater BPD pa thology. The BEST total score was included in this study as a criterion measure of borderline personality traits. Quality of Life Enjoyment and Satisfaction scale—short form (QLES). The QLES (Endicott, Nee, Harrison, & Blumenthal, 1993) is a self-report scale designed to assess levels of enjoyment and satisfaction experienced over the past week in the following areas of functioning: physical health, mood, school/ learning, interpersonal relationships, economic status, and ability to complete work/hobbies. Responses are rated on a 5-point scale ranging from Very poor to Very good. The short form of the QLES, which comprises the first 14 items assessing the aforementioned various areas of daily functioning, were summed to compute a maximum score that is a proportion (0%-100%) of the maximum possible score of 70. The QLES has adequate reliability and validity (Endicott, Nee, Yang, & Wohlberg, 2006; Rapaport, Clary, Fayyad, & Endicott, 2005). The QLES maximum score was included in this study as a criterion measure of general quality of life. Work and Social Adjustment Scale (WSAS). The WSAS is a five-item self-report measure designed to assess functional im pairment associated with an identified disorder or problem. Re sponses are rated on an 8-point scale ranging from No impairment to Very severe impairment. The WSAS has adequate reliability and validity (Mataix-Cols et al., 2005; Mundt, Marks, Shear, & Greist, 2002). The WSAS total score, ranging from 0 to 40, was included in this study as a criterion measure of functional impairment. D ata A nalysis To examine the factor structure of the ABASI, we submitted a polychoric correlation matrix of the seven ABASI items assessing the core NSI disorder criteria (Criteria B and C) to an exploratory factor analysis using maximum likelihood factoring and oblique (promax) rotation. In addition to Kaiser’s eigenvalue of 1 criterion and examination of the scree plot, Horn’s parallel analysis with Glorfeld’s Monte Carlo extension (Dinno, 2009; Glorfeld, 1995; Horn, 1965) was used to identify the optimal number of factors to retain for the final factor solution. Through generation of a large number of random data sets based on the number of subjects and variables of the target data set, parallel analysis compares the eigenvalues from the random data sets with the target data set to determine the optimal number of factors to retain. For the current parallel analysis, the critical eigenvalue was set at the 95th per centile and 1,000 random permutation data sets were generated by matching distributions from the raw data of the seven variables of the ABASI. Test-retest reliability of NSI disorder was examined using kappa coefficients. For NSI severity and the subscales of NSI disorder, interclass correlation coefficients (ICC) using a two-way random effects model with absolute agreement were used to ex amine test-retest reliability. To examine the utility of NSI disorder in identifying a more severe presentation of NSI, we examined group differences on measures of psychopathology and impair ment using t test statistics and effect size (Cohen’s d) estimates,
NONSUICIDAL SELF-INJURY DISORDER
with effect size confidence intervals estimated using 1,000 boot strap replications. Hierarchical multiple regressions were used to examine the incremental validity of NSI disorder when compared to existing measures of NSI characteristics (i.e., frequency of NSI, number of methods of NSI, urge to self-injure, age of onset of NSI, and duration of NSI). To explore the benefit of the ABASI as a dimensional measure of NSI severity, we entered the NSI severity score, an ordinal scale representing the summation of the seven ABASI items assessing the Criteria B and C of NSI disorder, into the hierarchical multiple regressions instead of NSI disorder. Vari ables created from a summation of Likert-scale responses were assumed to approximate an interval scale and were treated as such in analyses. For suicidal ideation, an ordinal regression was also performed to determine if results differed substantially when treat ing the dependent measure as an ordinal-level variable; no sub stantive differences in results were observed, and the original results were retained. The alpha level was adjusted to .01 to account for multiple comparisons. All analyses were conducted using Stata Statistical software (Release 13; StataCorp, 2013).
Results Psychometrics Results of the parallel analysis indicated that the eigenvalues for Factor 1 (2.78) and Factor 2 (1.01) were significantly higher than those expected by chance. Based on the parallel analysis, Kaiser’s criterion, and examination of the scree plot, a two-factor solution was retained, explaining 89.6% of the variance. As shown in Table 2, all three items assessing Criterion B loaded on Factor 1 and all four items assessing Criterion C loaded on Factor 2. Criterion B and Criterion C factors were significantly correlated, r(509) = .32, p < .001. Internal consistency, as measured by Cronbach’s alpha, was a - .75 for all seven items, with a = .79 for the Criterion B subscale and a = .65 for the Criterion C subscale. Despite ade quate internal consistency, corrected item-total correlations show that the ABASI item assessing Symptom B-3 is only weakly associated with the NSI severity score (see Table 3). Test-retest reliability using kappa coefficients and ICCs were calculated for 95 patients who were re-administered the ABASI between 1 and 14 days (M = 6.8, SD = 3.1) due to transfer between levels of care (e.g., from partial hospitalization to inpa tient). Test-retest reliability was moderate for the full NSI disorder
35
( k = .432), and good specifically for Criterion A ( k = .784) and Criterion C ( k = .662), but poor for Criterion B ( k = .216). Test-retest reliability was good for the ABASI NSI severity score (ICC = .606) and was moderate for the Criterion B (ICC = .484) and Criterion C (ICC = .517) subscales of the ABASI.
NSI Disorder Criteria Most, but not all, patients met Criterion A, with 85.5% of patients self-injuring 5 or more days in the past year with at least one specific type of NSI behavior. Across all types NSI, patients self-injured a mean of 76.8 (SD = 102.7) days in the past year, with a mean of 9.7 (SD = 14.3) separate incidents for each day injured. Patients used a mean of 4.0 (SD = 3.0) different types of NSI in the past year. Mean age of onset was 12.5 years old (SD = 4.0), with 96.6% reporting onset of NSI prior to age 18 years. Among the 14.5% (n = 74) of patients who reported fewer than 5 days for a specific type of NSI behavior in the past year, the mean number of days of NSI was 1.9 (SD = 1.3), with 2.7% of the full sample reporting no NSI in the last year. Given the ABASI likely underreported total NSI in the past year, the total number of NSI days reported across all types of NSI was summed to deter mine the maximum range of possible patients meeting Criterion A. At most, 21 additional patients were eligible for Criterion A, making the uppermost estimate of Criterion A 89.6% of the sam ple. As shown in Table 4, cutting was the most common type of NSI, followed by scratching, skin carving, food restriction, binging/purging, and hitting oneself. Because the ABASI in cludes 10 behaviors that may not necessarily result in damage to the surface of the body, it is possible that the ABASI overestimated the rate for Criterion A. To address this potential, the following 10 behaviors were removed from the algorithm creating Criterion A for NSI disorder: tattooing/piercing, selfchocking, pulling hair/eyebrow/eyelash, falling down, fighting, ingesting/swallowing objects, overexercising, binging/purging/ restricting food intake, sexual behaviors, and ignoring medical advice or making medical conditions worse. Removing these 10 behaviors reduced the proportion of those meeting Criterion A by 2.5%, making the lowermost estimate of Criterion A 83.0% of the sample. Using the uppermost estimate, which assumes different types of NSI behavior occurred on different days, and the lowermost estimate, which provides a conservative listing of only 11 NSI behaviors, the true rate of Criterion A is likely
Table 2 Rotated Factor Loadings fo r Criteria A and B o f the Alexian Brothers Assessment o f Self-Injury (ABASI) ABASI item (DSM criterion)
Factor 1
Factor 2
Relief from negative feelings or thoughts (B-l) Fix or resolve problems with other people (B-2) Create or increase positive feelings (B-3) Negative feelings or thoughts immediately before self-injuring (C-l) Difficulties or problems with other people immediately before self-injuring (C-l) Strong desire or urge to hurt yourself that was difficult to resist before self-injuring (C-2) Think about hurting yourself (C-3)
0.14 -0.06 -0.06 0.74 0.54
0.63 0.91 0.51 -0.03 0.14
0.89 0.74
-0.02 -0.04
Note.
DSM = Diagnostic and Statistical Manual o f Mental Disorders.
WASHBURN, POTTHOFF, JUZWIN, AND STYER
36 Table 3
Corrected Item-Total Correlations o f Nonsuicidal Self-Injury Severity Score (NSI) NSI Criteria Criteria severity B C
ABASI item (DSM criterion) Relief from negative feelings or thoughts (B-l) Fix or resolve problems with other people (B-2) Create or increase positive feelings (B-3) Negative feelings or thoughts immediately before self-injuring (C-l) Difficulties or problems with other people immediately before self-injuring (C-l) Strong desire or urge to hurt yourself that was difficult to resist before self-injuring (C-2) Think about hurting yourself (C-3)
0.48 0.46 0.26 0.51 0.48
0.49 0.54 0.39 0.62 0.51
0.61 0.52
0.71 0.57
Note. ABASI = Alexian Brothers Assessment of Self-Injury; DSM = Diagnostic and Statistical Manual of Mental Disorders.
between 83.0% and 89.6%. For the remaining estimates and analyses, the original estimate of 85.5% for Criterion A was used because it represents a reasonable mid-point between the lowermost and uppermost estimates. Sensitivity analyses were conducted on all subsequent analyses; no differences were found in the results when either the lowermost or uppermost estimates were used in place of the original estimate of Crite rion A (analyses available in the online supplementary materi als). Most, but not all, patients met Criterion B, with 87.7% indicating at least one expectation of NSI. Specifically, most patients expected that NSI would bring relief from either a negative feeling or cognitive state (82.0%) or resolution of an interpersonal problem (57.1%). Slightly more than one third of the sample expected that NSI would create or enhance a posi tive feeling state (34.8%). The majority of patients (61.8%)
reported two or more expectations of NSI, with nearly one quarter (24.5%) reporting all three expectations of NSI. Nearly all of the patients met Criterion C, with 92.8% of patients indicating either a precipitant to NSI or thoughts about NSI. Specifically, nearly all patients reported interpersonal problems or negative thoughts or feelings prior to NSI (91.3%), whereas most reported a preoccupation or urge to self-injure prior to NSI (71.6%) and thinking about harming themselves (72.8%). The majority of patients (80.2%) reported two or more symptoms within Criterion C, with 62.4% reporting all three symptoms. Nearly all patients met Criteria E and F, with 98.2% of patients receiving GAF scores at 50 or below, and only four patients (0.8%) having a clinical diagnosis than could exclude them from NSI disorder (three patients had an autism spectrum disorder diagnosis, and one had a drug withdrawal diagnosis).
Table 4
Characteristics o f Nonsuicidal Self-Injury (NSI) Behaviors Assessed by the Alexian Brothers Assessment of Self-Injury (ABASI) No. of days
Times per day
Age of onset
NSI behaviors
n
%
M
SD
M
SD
M
SD
Banged head Binged/purged/starved for pain/harm8 Broke limbs Burned skin Carved skin for pain/harm Choked self® Cut skin Drawn blood Fell down for pain/harm8 Fought for pain/harma Gouged skin Hit self Ingested objects for pain/harma Inserted/embedded objects Over-exercised for pain/harm8 Prevented healing of injuries Pulled hair/eyelashes/eyebrows for pain/harm0 Scratched skin Sexual behavior for pain/harm8 Tattooed for pain/harm8 Worsened medical condition for pain/harm0
511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511 511
23.5 29.9 1.2 26.4 32.5 14.9 93.0 9.4 2.5 3.5 1.4 29.7 2.7 5.5 7.4 26.4 13.7 56.0 5.1 11.9 4.3
5.27 24.54 0.02 3.10 6.62 0.75 52.13 4.25 0.17 0.34 0.12 8.07 0.11 1.59 3.08 13.84 6.23 25.22 1.06 0.50 0.72
24.90 70.06 0.20 15.70 31.99 3.38 77.99 27.38 3.11 4.55 1.84 30.74 1.02 17.90 21.83 49.49 35.06 65.32 16.08 2.19 6.44
0.58 0.71 0.01 0.38 0.43 0.18 2.86 0.17 0.03 0.04 0.02 0.85 0.04 0.14 0.11 0.84 0.42 1.52 0.09 0.20 0.07
2.03 1.49 0.11 0.98 0.89 0.52 6.30 0.82 0.16 0.20 0.14 2.85 0.27 0.91 0.55 4.78 1.76 3.63 0.59 1.17 0.40
14.54 13.70 15.60 14.21 14.02 14.01 13.27 12.87 13.25 13.00 13.67 13.79 20.31 12.93 13.70 12.66 13.23 13.51 17.77 14.97 15.43
5.13 2.90 3.58 2.85 3.88 4.21 3.27 1.78 1.76 1.71 2.16 4.75 12.85 2.66 2.12 2.44 3.54 4.87 8.32 3.82 3.60
Note. ABASI = Alexian Brothers Assessment of Self-Injury. a Behaviors considered to be self-injurious in the ABASI, but may not result in damage to the surface of the body.
NONSUICIDAL SELF-INJURY DISORDER
37
Rates of NSI Disorder
Incremental Criterion Validity of NSI Disorder
Nearly three quarters (74.0%) of the sample met all criteria for NSI disorder, as described in the DSM-5. If the uppermost esti mate of 89.9% for Criterion A is used to identify NSI disorder, the rate of NSI disorder increases to 76.9%. In contrast, if the lower most estimate of 83.0% for Criterion A is used, the rate of NSI decreases to 71.8%. No significant differences were found in the proportion of patients meeting NSI disorder criteria by age, gen der, or ethnicity. As shown in Table 5, regardless of NSI disorder status, the ABASI item assessing C-l was the most commonly endorsed symptom, with all of the patients meeting criteria for NSI disorder endorsing this item. In contrast, the item assessing Symptom B-3 was the least commonly endorsed symptom. The item assessing Symptom B-3 was also the least useful in meeting criteria for NSI disorder; only seven patients with NSI disorder met Criterion B through the item assessing B-3 alone. It was also rare for patients to meet Criterion B without also meeting Criterion C. Across the entire sample, only five patients who met Criterion B did not meet Criterion C. In contrast, 62 patients who met Criterion C did not meet Criterion B. Among patients who did not meet criteria for NSI disorder, 82.7% met all but one of the NSI disorder criteria, most commonly failing to meet either Criterion A (48.2%) or Criterion B (40.0%); very few of these patients failed to meet Criterion C (1.8%).
As shown in Table 7, commonly assessed characteristics of NSI explained between 17% and 49% of the variance across the crite rion measures of psychopathology and impairment. Among the characteristics of NSI, urge to self-injure was significantly asso ciated with all the criterion measures. The number of NSI methods was also significantly associated with overall psychopathology, borderline personality traits, and functional impairment. In con trast, frequency of NSI was only associated with borderline per sonality traits, and neither age of onset of NSI nor duration of NSI was associated with any of the criterion measures. Adding NSI disorder as a dichotomous variable to the model (see Step 2a in Table 7) failed to account for any additional variance in the criterion variables. In contrast, a model including NSI severity instead of NSI disorder in Step 2 accounted for additional variance beyond the characteristics of NSI for three of the criterion measures (see Step 2b in Table 7). Specifically, NSI severity accounted for an additional 2% of the variance for overall psychopathology, an additional 1% of variance for borderline personality traits, and an additional 1% of variance for quality of life. A model including the items assessing the specific criteria of NSI disorder instead of either NSI Disorder or NSI severity in Step 2 was also examined. The item assessing Symptom C-l was significantly associated with all of the criterion measures of psy chopathology and impairment, except for suicidal ideation. For suicidal ideation, only the item assessing C-3 remained signifi cantly associated after accounting for the characteristics of NSI. Given the conceptual and empirical overlap between the ABUSI and the item assessing C-2 (/is = .55, p < .001), we repeated the analyses with the ABUSI total score removed from Step 1. The item assessing C-2 was not found to account for any significant variance across any of the criterion measure, even with the ABUSI total score removed from the model.
Correlates of NSI Disorder Table 6 summarizes the differences between patients with and without NSI disorder on common characteristics of NSI, as well as criterion measures of psychopathology and impairment. In contrast to those without NSI disorder, those with NSI disorder experienced significantly greater urges to self-injure, more days of self-injury in the past year, and a greater number of methods used to selfinjure in the past year. No differences between those with and without NSI disorder were found on age of onset or duration of NSI. NSI disorder was also associated with significantly worse overall psychopathology, suicidal ideation, traits of borderline personality disorder, quality of life, and functional impairment.
Table 5 Endorsement o f Specific NSI Disorder Criteria by NSI Disorder Status NSI disorder Criteria Criterion Criterion Criterion Criterion Criterion Criterion Criterion Criterion Criterion Note.
A B1 B2 B3 Cl C2 C3 E F
Positive for (%)
Negative for (%)
100.0 93.4 67.5 41.3 100.0 96.8 98.4 100.0 100.0
44.4 49.6 27.8 16.5 92.5 74.4 86.5 93.2 97.0
NSI = Nonsuicidal Self-Injury.
Discussion The ABASI adds to a limited but growing body of literature examining the utility of a criterion-based approach to identifying clinically significant NSI. This study demonstrates that NSI dis order can be reliably identified among people who self-injure using a brief self-report instrument. The findings also indicate that NSI disorder, as identified by the ABASI, represents a clinically significant level of NSI. NSI disorder characterizes nearly three quarters of adolescents and adults being treated for NSI at an acute level of care and is indicative of a more severe case presentation when compared with simply identifying self-injurious behavior as a clinical concern. Specifically, NSI disorder is associated with significantly higher frequency of NSI, a greater number of NSI methods, a stronger urge to self-injure, more severe psychopathol ogy, greater suicidal ideation, more borderline personality traits, and more impaired quality of life and functioning than found in people who self-injure but do not meet criteria for NSI Disorder. As a dichotomous diagnosis, however, NSI disorder appears to have limited utility above and beyond NSI characteristics that are commonly assessed in the literature. Results of the hierarchical regression analyses indicate that NSI disorder fails to account for any additional variance in criterion measures of psychopathology and impairment beyond the urge to self-injure, number of NSI methods, and frequency of NSI. In contrast, an ordinal level NSI
WASHBURN, POTTHOFF, JUZWIN, AND STYER
38
Table 6 Correlates o f Nonsuicidal Self-Injury Disorder (NSI) NSI disorder Positive for Criterion NSI characteristics Frequency Methods ABUSI total score Onset Duration Psychopathology BASIS total Suicidal ideation BEST total Impairment QLES WSAS
Negative for
M
SD
M
SD
Cohen’s d
88.72 4.29 21.06 12.52 4.72
104.80 2.78 7.86 3.53 6.23
42.91 3.21 16.83 13.07 3.89
88.31 3.40 8.62 4.21 3.87
0.45 0.37 0.52 0.15 0.14
[.27, [.14, [.32, [.08, [.02,
.64] .60] .73] .37] .31]
