Ecological Momentary Assessment of Nonsuicidal Self-Injury in Youth With Borderline Personality Disorder

Personality Disorders: Theory, Research, and Treatment Ecological Momentary Assessment of Nonsuicidal SelfInjury in Youth With Borderline Personality Disorder Holly E. Andrewes, Carol Hulbert, Susan M. Cotton, Jennifer Betts, and Andrew M. Chanen Online First Publication, August 8, 2016. http://dx.doi.org/10.1037/per0000205 CITATION Andrewes, H. E., Hulbert, C., Cotton, S. M., Betts, J., & Chanen, A. M. (2016, August 8). Ecological Momentary Assessment of Nonsuicidal Self-Injury in Youth With Borderline Personality Disorder. Personality Disorders: Theory, Research, and Treatment. Advance online publication. http:// dx.doi.org/10.1037/per0000205 Personality Disorders: Theory, Research, and Treatment 2016, Vol. 7, No. 4, 000 © 2016 American Psychological Association 1949-2715/16/$12.00 http://dx.doi.org/10.1037/per0000205 Ecological Momentary Assessment of Nonsuicidal Self-Injury in Youth With Borderline Personality Disorder Holly E. Andrewes Carol Hulbert The University of Melbourne and Orygen, The National Centre of Excellence in Youth Mental Health, Melbourne, Victoria, Australia The University of Melbourne This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Susan M. Cotton, Jennifer Betts, and Andrew M. Chanen Orygen, The National Centre of Excellence in Youth Mental Health, Melbourne, Victoria, Australia, and Centre for Youth Mental Health, The University of Melbourne Nonsuicidal self-injury (NSSI) is highly prevalent among individuals with borderline personality disorder (BPD). The aim of this study was to investigate the cognitive, emotional, and contextual experience of NSSI in 107 youth (aged 15–25 years) with BPD who had minimal prior exposure to treatment. Using ecological momentary assessment, participants completed a randomly prompted questionnaire about their affect, selfinjurious thoughts, and behaviors six times per day for 6 days. Twenty-four youth with BPD engaged in 52 counts of NSSI, with 56 motives identified. Open-ended questions revealed that on occasions of NSSI, a large minority of participants could identify neither their motives (27%, n ⫽ 15) nor the environmental precipitants (46%, n ⫽ 24) for NSSI. Changes in affect revealed a pattern of increasing negative and decreasing positive affect prior to NSSI, with a reduction in negative and an increase in positive affect following NSSI. These changes were absent for those who did not engage in NSSI. Initial self-injurious thoughts and changes in negative and positive affect occurred a median of 35, 15, and 10 hr prior to NSSI, respectively. These findings suggest that youth with BPD have limited capacity to reflect on their motives and environment preceding NSSI. The patterns of affect change indicate that NSSI is maintained by reward incentives as well as negative reinforcement. The time between initial self-injurious thoughts and engagement in NSSI reveals a window of opportunity for intervention. Keywords: borderline personality disorder, nonsuicidal self-injury, youth, psychiatry deaths (Chapman, Specht, & Cellucci, 2005), and high levels of treatment utilization (Bender et al., 2006). The incidence of NSSI peaks at puberty (Moran et al., 2012) in both the general population and those with BPD, with two thirds of adults with BPD reporting that they commenced NSSI prior to adulthood (Zanarini et al., 2006). Accordingly, improved understanding of NSSI early in the course of BPD is needed to guide early intervention. Background Borderline personality disorder (BPD) is a severe mental disorder that is characterized by a pervasive pattern of impulsivity and instability in emotion regulation, interpersonal relationships, and self-image (Leichsenring, Leibing, Kruse, New, & Leweke, 2011). The onset of BPD usually occurs in the period between puberty and emerging adulthood, but little is known about this disorder during its early stages (Chanen, 2015). Nonsuicidal self-injury (NSSI) is the deliberate destruction of one’s bodily tissue that is without cultural significance or lethal intent (Nock, 2009). This behavior is a core feature of BPD (Leichsenring et al., 2011) and is associated with lasting physical injury, accidental Non-Suicidal Self-Injury in BPD Retrospective research has studied the functions of NSSI with respect to motives for engagement and changes in emotions that precede and follow NSSI (Klonsky, 2007). In adults with BPD, the Career Development Fellowship (APP1061998). The content is solely the responsibility of the authors and does not necessarily represent official views of the NHMRC. A special thanks to the patients, families, and staff of the HYPE Program at Orygen Youth Health and headspace Western Melbourne. We also thank Sharnel Perera, Sinn Yuin Chong, Victoria Rayner, and Francesca Kuperman for their assistance with data collection. Correspondence concerning this article should be addressed to Andrew M. Chanen, Orygen, The National Centre of Excellence in Youth Mental Health, Locked Bag 10, Parkville, Victoria, Australia 3052. E-mail: andrew.chanen@orygen.org.au Holly E. Andrewes, Melbourne School of Psychological Sciences, University of Melbourne and Orygen, The National Centre of Excellence in Youth Mental Health, Melbourne, Victoria, Australia; Carol Hulbert, Melbourne School of Psychological Sciences, University of Melbourne; Susan M. Cotton, Jennifer Betts, and Andrew M. Chanen, Orygen, The National Centre of Excellence in Youth Mental Health and Centre for Youth Mental Health, University of Melbourne. This project was conducted within a study supported by a National Health and Medical Research Council (NHMRC) project grant (GNT0628739). Professor Sue Cotton is supported by an NHMRC 1 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 2 ANDREWES ET AL. most commonly reported motive is affect regulation (e.g., Kleindienst et al., 2008), followed by self-punishment, antidissociation, interpersonal influence (e.g., Brown, Comtois, & Linehan, 2002), and regaining control (Kleindienst et al., 2008). Motives such as sensation seeking are less commonly cited, yet have been identified as a secondary motive by 46% of adults with BPD (Kleindienst et al., 2008). To the authors’ knowledge, only one study (Sadeh et al., 2014) has investigated motives in adolescents with BPD (mean age ⫽ 16.7 years). Adolescents identified affect regulation, followed by self-punishment, as the motives most relevant to their engagement in NSSI. Emotions that precede and follow NSSI have also mostly been studied in adults with BPD using retrospective interviews, questionnaires (Chapman & Dixon-Gordon, 2007; Kleindienst et al., 2008), and laboratory studies (Reitz et al., 2012). Among these studies, most adults with BPD have retrospectively reported an increase in negative emotions, including anger (Chapman & Dixon-Gordon, 2007) and aversive tension (Kleindienst et al., 2008) prior to NSSI, with a reduction following NSSI (Chapman & Dixon-Gordon, 2007; Kleindienst et al., 2008; Reitz et al., 2012). An increase in positive affect has been posited to follow NSSI in BPD participants (Kleindienst et al., 2008), yet few studies have investigated this. Relying solely on these retrospective and laboratory findings is problematic as people with BPD often retrospectively overestimate levels of negative affect (EbnerPriemer et al., 2006) and are inaccurate in their descriptions of mood fluctuations over time (Solhan, Trull, Jahng, & Wood, 2009). Findings from laboratory designs, which rely on proxies for self-injury, such as an incision made by the investigator in the arm of participants (Reitz et al., 2012), might have limited generalizability. Furthermore, both retrospective reports and laboratory studies fail to capture important details, such as variability in affect intensity, and timing of cognitions and affective changes prior to engagement in NSSI. Ecological Momentary Assessment To address these limitations, ecological momentary assessment (EMA) has become the preferred method for elucidating the phenomenology of NSSI (Armey, 2012). EMA captures people’s internal states and cognitions as they happen in their natural environment (Shiffman, Stone, & Hufford, 2008). To the authors’ knowledge, no study has employed EMA to identify motives for NSSI or assess emotional changes surrounding NSSI among treatment-seeking patients with BPD. One study that recruited from the community through newspaper advertisements (Snir, Rafaeli, Gadassi, Berenson, & Downey, 2015) found that adults with BPD (n ⫽ 18, mean age ⫽ 30.9 years) most frequently identified emotional relief, feeling generation, and selfpunishment from a provided list of reasons for NSSI. However, a major methodological shortcoming of this study and previously mentioned retrospective studies is their use of checklists, rather than open-ended questions, to prompt participant selection of motives. This is problematic as people with BPD show impaired metacognitive capacities (Dimaggio & Lysaker, 2015), along with a compromised recall of autobiographical memories, which have been found to be less specific (Maurex et al., 2010), less coherent, and more disorganized (Jørgensen et al., 2012) than among people with obsessive– compulsive disorder and nonclinical community samples. Thus, in moments of distress, people with BPD might have limited capacity to spontaneously identify their motives for NSSI. It is probable that prior studies that rely on checklists of motives have masked these cognitive difficulties. Snir and colleagues (2015) also identified inferred motives from the emotional changes occurring pre- and post-NSSI. Their findings showed an increase in dissociation, as well as perceived rejection/isolation prior to NSSI and a reduction following NSSI, approximating a quadratic curve. Interestingly, participants did not exhibit any reliable changes to indicate a quadratic trajectory in negative affect surrounding NSSI. This contrasts with results from retrospective and laboratory studies in adults with BPD in which participants reported an increase in negative affect prior to and a reduction following NSSI (Chapman & Dixon-Gordon, 2007; Kleindienst et al., 2008; Reitz et al., 2012). It also fails to support affect regulation motives for engaging in NSSI (e.g., Kleindienst et al., 2008). Furthermore, changes in positive affect were not assessed in Snir and colleagues’ (2015) study, and therefore, the real-time role of positive reinforcement as a maintaining factor for NSSI in BPD remains unclear. In addition, findings from studies among adults with BPD might be confounded by duration of illness effects that include polypharmacy, other treatments, and recurrent co-occurring mental state disorders that, over time, might influence the motives, cognitive changes, and affect associated with NSSI (Chanen, 2015). Changes in affect pre- and post-NSSI have been assessed in other clinical and undergraduate samples recruited from the campus and community, but these findings are contradictory. An increase in negative affect prior to NSSI and a reduction following were demonstrated with 17 adolescent undergraduates (mean age ⫽ 18.7 years; Armey, Crowther, & Miller, 2011). In contrast, negative affect increased prior to NSSI in 19 adults with bulimia nervosa (mean age ⫽ 25.3 years; Muehlenkamp et al., 2009) and did not reduce following NSSI. Changes in positive affect were also investigated, with the changes experienced by undergraduate youth failing to show reliable changes before and after NSSI. In contrast, adults with bulimia nervosa experienced a reduction in positive affect prior to NSSI and an increase following NSSI. A particular advantage of EMA is that it also enables the assessment of the time between initial changes in affect, selfinjurious thoughts, and engagement in NSSI. Among undergraduate students, Armey and colleagues (2011) found that there was a mean of 8 hr between a change in negative affect and engagement in NSSI. However, as youth with BPD have higher levels of impulsivity and a lower capacity for emotion regulation than healthy adolescents in the community (Brunner et al., 2007), the potential window for intervention, between change in affect and engagement in NSSI, might be shorter. Using retrospective reporting within an EMA paradigm and via self-report respectively, Armey and colleagues (2011) and Snir and colleagues (2015) also found that the majority of participants reported that they spent less than an hour preparing for NSSI. However, it is unclear whether retrospective reports of preparation for NSSI are akin to the duration of time between an initial self-injurious thought and engagement in NSSI. NONSUICIDAL SELF-INJURY IN YOUTH WITH BPD This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Current Investigation In summary, real-time motives for NSSI reported by youth with BPD remain unclear due to employment of largely retrospective paradigms that rely on checklists rather than open-ended questions to elicit motives. Elucidation of the changes in affect that occur pre- and post-NSSI are also required due to conflicting findings from undergraduate and clinical community samples that are not generalizable to acutely unwell treatment-seeking “youth” (aged 15–25 years) with BPD. This study aimed to (a) use both open- and closed-ended questions in experience sampling to delineate the real-time motives and changes in affect that trigger and maintain NSSI in youth with BPD and (b) establish the temporal course of changes in affect, initial self-injurious thoughts, and engagement in NSSI. It was hypothesized that (a) although the most commonly reported functions of NSSI will be affect regulation and selfpunishment (Sadeh et al., 2014), the use of open-ended inquiry will reveal a group of participants who are unable to identify their motives; (b) negative affect will increase prior to and then decrease following NSSI, whereas positive affect will decrease prior to and increase following NSSI, with both approximating a quadratic curve; (c) in those who do not engage in NSSI, affect before and after a random point will not approximate a quadratic curve; and (d) youth with BPD will experience an interval of less than 8 hr between changes in positive and negative affect and NSSI. Method Participants were recruited as part of a larger randomized control trial of early intervention for youth with first-presentation BPD (for details, see Chanen et al., 2015). In total, 113 help-seeking youth, aged 15–25 years, with a diagnosis of BPD were recruited through the triage systems of two government-funded youth mental health services in western metropolitan Melbourne, Australia. 3 The United Nations and World Health Organization define youth as persons between ages 15 and 24 years (United Nations, 2001), and the term is also used by the International Association of Youth Mental Health to describe young people aged 12–25 years. The use of this term reflects studies from sociology, developmental psychology, and developmental neuroscience (Arnett, 2000; Nelson, Leibenluft, McClure, & Pine, 2005; Paus, 2005; Steinberg, 2005) that all point to an extended but coherent period of development from puberty to around 25 years of age in economically developed societies. Recently, we have recommended that, for personality disorder research and treatment, more natural developmental periods would be childhood, youth (adolescents and emerging adulthood), adulthood, and old age (Chanen, Tackett, & Thompson, in press). Six participants were excluded due to technical failure in the EMA safety alert system and/or failure to complete any of the EMA questionnaires. The final sample comprised 107 young people with BPD, 94% (n ⫽ 99) of whom had engaged in NSSI in the past 3 months. Table 1 outlines the sample characteristics. Measures Psychopathology. The Structured Clinical Interview for DSM–IV (SCID) Axis I–Patient Edition (SCID-I/P; First, Spitzer, Gibbon, & Williams, 2002) and SCID-II (First, Gibbon, Spitzer, Williams, & Benjamin, 1997) were used as diagnostic tools. The SCID-I/P and SCID-II have good interrater reliability for (DSM– IV–TR; American Psychiatric Association, 2000) clinical diagnoses, with Cohen’s ␬ ranging between 0.64 and 0.87 for the SCIDI/P (Lobbestael, Leurgans, Arntz, & Wiley, 2011), and a Cohen’s ␬ of between 0.65 and 0.95 using the SCID-II (Lobbestael et al., 2011). History of NSSI. The 13-item Parasuicidal Behaviors subscale of the Borderline Personality Disorder Severity Index (BPDSI; Arntz et al., 2003) captured engagement in NSSI over the Table 1 Demographic Characteristics of the Total Cohort Characteristic Total (N ⫽ 107) NSSI⫹ (n ⫽ 24) NSSI⫺(n ⫽ 83) Female (%) Age (years), M (SD) High socioeconomic disadvantage (%)a Completed final year of school (%)b Unemployed (%) Taken leave from school/work (%) Employed part-time (%) Employed full-time (%) Mental state disorder (%) Mood disorders Anxiety disorder Eating disorders Personality disorders (%) Antisocialc Paranoid Avoidant Narcissistic Histrionic Dependent 83.2 18.1 (2.7) 45.8ⴱ 38.8 12.0 49.0 11.0 7.5 87.5 18.2 (2.9) 29.2 75.0 20.8 45.8 8.3 8.3 81.9 19.1 (2.7) 50.6 31.7 9.6 49.4 12.0 7.2 83.2 71.3 8.4 83.3 75.0 12.5 83.2 70 7.2 31.8 20.6 23.4 3.7 3.7 3.0 37.5 29.2 16.7 0 4.2 4.2 30.1 18.1 25.3 4.8 3.6 2.4 Note. NSSI ⫽ nonsuicidal self-injury. a Rated according to the participants’ residential postcode (Vinson, 2007). b Participants over 18 years. agnosis made ignoring Criterion B that requires ⬎ 18 years of age (Chanen et al., 2007). ⴱ p ⬍ .05 (indicating a significant difference between those who did and did not engage in NSSI). c Di- This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 4 ANDREWES ET AL. preceding 3 months. This subscale exhibits good interrater reliability, with an intraclass correlation of 0.93 and a median Cronbach’s alpha (␣) of 0.69 (Arntz et al., 2003). Current positive and negative affect. This was measured by the 10-item short-form of the Positive and Negative Affect Scale (PANAS; Kercher, 1992), delivered through the Mobiletype EMA program (Reid et al., 2009). The PANAS comprised the following question: “How [insert positive or negative affect] are you feeling right now?” with a rating scale of 1–5 (1 ⫽ very slightly or not at all to 5 ⫽ extremely). The total scores for positive affect and negative affect were calculated separately by summing the ratings on each scale. The structural characteristics of the 10-item PANAS are stable across age groups and gender (Mackinnon et al., 1999). The reliability of both the positive and negative affect scales is also high, with a Cronbach’s alpha of 0.78 and 0.87, respectively (Mackinnon et al., 1999). NSSI and self-injurious thoughts. This was captured by a set of items delivered by the Mobiletype EMA program (Reid et al., 2009), which comprised the closed questions with yes/no answers (“Since the last signal have you thought about deliberately hurting yourself?” “Since the last signal, have you actually hurt yourself?”) and open questions (“What did you do to hurt yourself?” “What was going on, just before you thought about hurting yourself?” “Why did you hurt yourself?”). These questions identified the occurrence of engagement in a self-injurious thought and/or NSSI, the methods used to engage in NSSI, and the context and motive for engaging in NSSI, respectively. Motives for NSSI were used to distinguish NSSI with and without suicidal intent. Descriptive motives were coded by two raters according to the seven most common reasons for NSSI reported in Klonsky’s (2007) review. The interrater reliability indicated almost perfect agreement, with a Cohen’s ␬ of 0.95, p ⬍ .001. A third rater coded the few conflicting ratings, with final categorization according to agreement with one of the raters. Procedure Ethical approval was obtained from the Melbourne Health Human Research Ethics Committee. Written informed consent was obtained from participants and from a parent or guardian if participants were minors. Upon entry to the study, participants completed the SCID-I/P, SCID-II, and the Parasuicidal Behaviors subscale of the BPD-SI and were issued with a mobile phone that contained an electronic diary program (Mobiletype; Reid et al., 2009). This program prompted participants to complete a survey identifying their affect, experiences, and behaviors six times per day for 6 days. Response prompts were randomized within 2-hr time blocks between 10 a.m. and 10 p.m. Participants were given 15 min from the prompt to begin completing the survey, so as to ensure that participant responses were completed contemporaneously. A total of 8 min was provided for the participant to complete the survey, after which point they were locked out. If participants did not complete at least one survey per day, the phone was left with them for a second week to ensure sufficient data collection. Participants were reimbursed $40 for their participation, regardless of their adherence to the Mobiletype protocol. Statistical Procedures An assessment of normality, linearity, homoscedasticity, or univariate and multivariate outliers was performed for all demographic variables, Parasuicidal Behaviors subscale of the BPD-SI, and EMA data, including missing data, timing of assessment, and timing between the change in affect, cognitions, and NSSI. Where nonnormality was cited, the median was presented as the best reportable measure of central tendency (Tabachnick & Fidell, 2013). To compare the changes in affect pre- and post-NSSI (NSSI⫹; Hypothesis 2), the two consecutive data points immediately preceding the first NSSI occurrence and the two available time points immediately following this event were extracted (Armey et al., 2011). Data points selected were all within 20-hr pre- and postNSSI. As EMA studies have not previously assessed youth with BPD, a longer time period was selected than has been used in prior EMA studies to ensure that initial affect changes were captured. Time was centered on the recording of the NSSI event (T0) with the times prior to and following the event respectively added and subtracted from this point and identified as negative and positive time values. Multiple NSSI events occurring over a 24-hr period were excluded, as per Snir and colleagues’ (2015) procedure. For individuals who did not engage in NSSI (NSSI–; Hypothesis 3), the data from five consecutive time points were identified by randomly selecting a time point (T0) and two time points immediately before and after. Multilevel modeling (MLM) was employed to analyze these data because it allows for the assessment of individual changes in affect overtime (Level 1) and a comparison of changes in affect between individuals (Level 2; Singer & Willett, 2003). MLM also handles large amounts of missing data without excluding whole cases. This analysis was conducted using maximum likelihood estimation with robust standard errors to provide parameter estimates that facilitate the fit of a model with missing data (Yuan & Bentler, 2000) and deviations from normality (Maas & Hox, 2004). As time points differed between participants, a variable that represents time (in hours) before and after T0 was created, and intercepts, linear slopes, and quadratic slopes were allowed to vary using random effects. The fit statistics of a linear, quadratic, and cubic model were initially compared before adding a dummycoded variable to distinguish the NSSI⫹ group (coded 1) from the NSSI– group (coded 0; Singer & Willett, 2003). To adjust for small growth variances in the quadratic MLM, time was divided by 24, with changes in affect interpreted as occurring over a day, rather than an hour. The standard equation for this quadratic MLM is as follows: Yij ⫽ ␥00 ⫹ ␥10TIMEij ⫹ ␥20TIMEij2 ⫹ ␥01NSSI ⫹ ␥11共NSSIj TIMEij兲 ⫹ ␥21共NSSIj TIMEij2兲 ⫹ ␨0j ⫹ ␨1jTIMEij ⫹ ␨2jTIMEij2 ⫹ εij Timing of the initial change in affect reported prior to NSSI (Hypothesis 4) over the 6 days of data collection was identified as the highest or lowest rating in positive or negative affect, respectively, prior to a reduction or increase leading to NSSI. Timing of the first self-injurious thought reported over the 6 days of data collection and the first self-injurious thought following a change in affect but prior to engagement in NSSI was also recorded. Addi- This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. NONSUICIDAL SELF-INJURY IN YOUTH WITH BPD tional criteria for data point selection of both thoughts and affect included (a) if 24 hr of missing data were identified prior to NSSI, then the participant was excluded from this analysis, and (b) for individuals who engaged in more than one NSSI event, only changes in affect or thoughts occurring prior to the first NSSI were included. Measures of effect were calculated for all statistics. Partial eta squared (␩2) and Pearson coefficient (r) were calculated for t tests and correlations, respectively. Local effect sizes for MLM were identified by the proportional reduction in the variance of the quadratic model after NSSI was added as a predictor (Peugh, 2010). Multilevel modeling was completed using MPLUS version 7.2. All other analysis was complete using SPSS version 21. Results Characteristics of the Sample NSSI⫹ and NSSI– participants were compared on age, gender, comorbidity, country of birth, education, employment, and social disadvantage rank. A significant correlation was found for social disadvantage rank (1 ⫽ low disadvantage, 3 ⫽ high disadvantage), with lower social disadvantage rankings found for participants who engaged in NSSI, ␹2(1, N ⫽ 107) ⫽ 8.81, p ⫽ .003. Characteristics of the EMA Data Participants completed a total of 1,986 diary entries. Twentytwo participants (21%) were given the phone for a second week because they completed an insufficient number of data entries in the first week. An average of 18.56 (SD ⫽ 10.0) of a possible 36 diary entries were completed over 6 days of data collection. This represents an average of 51.56% of the total available random samples. The total number of diary entries completed by the NSSI⫹ group (M ⫽ 21.17, SD ⫽ 8.48) did not differ significantly from the NSSI– group (M ⫽ 17.42, SD ⫽ 10.37), t(105) ⫽ ⫺1.62, p ⫽ .108. Two acts of self-injury with suicidal intent were removed from the sample and excluded from further analysis. In the final sample, 24 participants engaged in 52 counts of NSSI over the 6-day period, which equates to an average of 2.2 per person (range: 1– 8). A significantly higher number of selfinjurious thoughts were reported by the NSSI⫹ group (n ⫽ 201, 5 M ⫽ 8.33, SD ⫽ 6.78) compared with the NSSI– group (n ⫽ 131, M ⫽ 1.47, SD ⫽ 2.35), t(24.6) ⫽ ⫺4.87, p ⬍ .001. The magnitude of the difference was very large (␩2 ⫽ 0.23; Cohen, 1988). Method, Context, and Function of NSSI Cutting was the most common method of NSSI. When participants were asked about the context in which NSSI occurred, 46% stated that they were unaware of what was happening just prior to engaging in this behavior (Mage ⫽ 17.74 years; range ⫽ 15.14 – 22.93 years). Participants most commonly reported engaging in NSSI to reduce their distress. Twenty-seven percent of the reported motives were unable to be categorized (Mage ⫽ 18.93 years; range ⫽ 15.19 –24.65 years), with participants acknowledging that they did not know why they had engaged in NSSI in the large majority of cases. One third (33%) of participants who were unable to identify their motives for engagement were also unable to identify the context in which NSSI occurred (see Table 2). Change in Affect Before and After NSSI The log-likelihood deviance statistic (Singer & Willett, 2003) for negative and positive affect showed that the quadratic model was a better fit than the linear model (negative, ⌬␹2(4) ⫽ 28.10 ⫽ 4, p ⬍ .001; positive, ⌬␹2(4) ⫽ 46.05, p ⬍ .001), or the cubic model, which failed to identify. This finding supported the mean plots for positive and negative affect (see Figure 1). The fixed effects reflect a typical person in each group (Bolger & Laurenceau, 2013). These revealed that when NSSI was recorded (T0), the mean ratings of negative and positive affect were significantly higher and lower, respectively, than those of the NSSI– group at a random point. Changes in both positive and negative affect approximated a quadratic curve in the NSSI⫹ group, which was not found in the NSSI– group (see Table 3). Local effect sizes revealed that after adding NSSI as a predictor variable to the model, the variance in the intercept reduced by 13.21% and 11.57% for positive and negative affect, respectively. The variance in the quadratic slope also reduced by 85.22% and 66.52% for positive and negative affect, respectively. Local effect sizes revealed a reduction in Level 1 variance by 13.21% and 13.88% in negative and positive affect, respectively, when NSSI was added as a predictor to the model. Table 2 Method of NSSI, Context Surrounding Engagement, and Motives Method of NSSI n (%) Cutting/carving Scratching skin Hitting object/self Biting self Burning self Refused to state Strangling self 37 (70) 8 (15) 3 (6) 1 (2) 1 (2) 2 (4) 1 (2) Total methods 53 (100)a Context of NSSI n (%) Categorization n (%) Nothing Interpersonal conflict Memories of past Rumination Stressful event Suicidal thoughts Eating Bored Total contexts 24 (46) 11 (21) 5 (10) 4 (8) 3 (6) 3 (6) 1 (2) 1 (2) 52 (100) Affect regulation Self-punishment Sensation seeking Antidissociation Antisuicide No category “I don’t know” “I had the urge” Total motives 25 (45) 5 (9) 4 (7) 3 (5) 1 (2) 15 (27) 14 (25) 1 (2) 56 (100)b Note. NSSI ⫽ nonsuicidal self-injury. a Fifty-three total methods were cited because two methods were endorsed in one event. b and 56 motives were cited because two motives were identified in four separate events. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Mean Negative Affect (PANASNA) 25 23 21 19 17 15 13 11 9 Mean Positive Affect (PANAS-PA) ANDREWES ET AL. 6 NSSINSSI+ 25 23 21 19 17 15 13 11 9 NSSINSSI+ Figure 1. Mean of total negative (left) and positive affect (right) ratings with standard error bars before and after NSSI (NSSI⫹) or a random point for the group who did not engage in NSSI (NSSI–). Possible range 5–25. NSSI ⫽ nonsuicidal self-injury; PANAS-NA ⫽ Positive and Negative Affect Scale - Negative Affect; PANAS-PA ⫽ Positive and Negative Affect Scale - Positive Affect. literature. First, young people with BPD who were assessed in real time most commonly reported affect regulation as their motive for engaging in NSSI. This finding supports a recent EMA study in a community sample of adults with BPD who responded to advertisements (Snir et al., 2015), as well as retrospective studies among treatment-seeking adolescents (Sadeh et al., 2014) and adult patients with BPD (Kleindienst et al., 2008). However, only 9% of participants in the current study cited self-punishment as a reason for NSSI, which contrasts with a prior retrospective study among adolescents with BPD who rated this as the second most relevant motivation for engaging in NSSI (Sadeh et al., 2014). Second, 27% of participants were unable to identify their motives for engaging in NSSI when given a free choice answer format, which is consistent with research indicating that people with BPD have impaired metacognitive capacities (Dimaggio & Lysaker, 2015). This finding gives rise to concerns about the validity of motives reported in studies using checklists to prompt participants’ responses (e.g., Sadeh et al., 2014). Furthermore, for almost half of the NSSI events, participants were unable to identify Timing in Changes of Affect and Cognitions Changes in negative and positive affect occurred a median of 15.18 hr (range, 2.33–38.44) and 10.04 hr (range, 2.14 –38.55), respectively, prior to engagement NSSI. The time between the first self-injurious thought in the 6-day assessment period and engagement in NSSI was a median of 34.86 hr but ranged between 0 and 122 hr (5 days), indicating high levels of variability. Following a change in negative affect, additional self-injurious thoughts occurred a median of 2.5 hr before NSSI, with a range of 0 –21.79 hr. A strong positive relationship was found between the timing of initial thoughts about NSSI and the change in negative affect (r ⫽ .47, n ⫽ 18, p ⫽ .050) and positive affect (r ⫽ .73, n ⫽ 20, p ⫽ .001). Discussion This study is the first to investigate the real-time cognitions and emotional changes pre- and post-NSSI in treatment-seeking patients with BPD. It contributes four major findings to the current Table 3 Parameter Estimates for a Conditional Quadratic Growth Model of Negative and Positive Emotions as a Function of Engagement in NSSI Negative affect Variable Fixed effects Intercept (NSSI–) NSSI (intercept) TIME NSSIj ⫻ Timeij TIME2 NSSIj ⫻ TIMEIJ2 Random effectsb Level 2: intercept variance Level 1: residual variance Positive affect a Parameter Estimate (SE) z 95% CI Estimate (SE) z 95% CIa ␥00 ␥01 ␥10 ␥11 ␥20 ␥21 9.09 (.70) 3.24 (1.20) ⫺1.15 (1.21) 1.44 (1.21) .28 (1.27) ⫺6.14 (1.65) 12.98ⴱⴱⴱ 2.68ⴱⴱ .70 1.19 .23 ⫺3.71ⴱⴱⴱ [7.72, 10.46] [.87, 5.61] [⫺2.46, 7.12] [⫺9.39, 3.80] [⫺8.27, 5.93] [⫺9.39, 2.89] 9.45 (.44) ⫺2.30 (.87) 1.70 (1.66) ⫺.07 (1.25) ⫺2.14 (3.88) 3.77 (1.61) 21.54ⴱⴱⴱ ⫺2.65ⴱ 1.02 ⫺.06 ⫺.55 2.35ⴱⴱ [8.60, 10.32] [⫺4.01, ⫺.60] [⫺1.55, 4.95] [⫺2.53, 2.39] [⫺9.74, 5.46] [.62, 6.93] ␨0i εi 12.28 (3.01) 11.04 (.55) 4.10ⴱⴱⴱ 19.89ⴱⴱⴱ [6.44, 18.22] [9.96, 12.13] 5.35 (1.82) 11.89 (.73) 2.94ⴱⴱ 16.30ⴱⴱⴱ [1.78, 8.91] [10.46, 13.32] Note. NSSI ⫽ nonsuicidal self-injury. a Bootstrapped confidence intervals. b Due to the discrepancy between group sizes, the within-group variance in the NSSI– group likely obscured the within-person variance in the NSSI⫹ group. For this reason, caution is required when interpreting these results, and only significant random effects were reported. ⴱ p ⱕ .05. ⴱⴱ p ⱕ .005. ⴱⴱⴱ p ⱕ .001. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. NONSUICIDAL SELF-INJURY IN YOUTH WITH BPD what was happening just prior to engagement, suggesting that youth with BPD might have limited external awareness in states of high distress. The age of participants was not associated with the ability to identify their motives or the context in which NSSI occurred, indicating this deficit might be unrelated to developmental maturity in youth with BPD. This is clinically relevant because reduced insight into motives and the environmental precipitants for NSSI might prevent an individual from intervening with appropriate cognitive strategies aimed at reducing NSSI. Reduced awareness of their experiences might also contribute to enhanced distress, which may increase the chance of engaging in NSSI. The latter prediction offers an area for future investigation, which is supported by prior studies showing an association between distress (or NSSI) and reduced emotional clarity (Ebner-Priemer et al., 2008; Zaki, Coifman, Rafaeli, Berenson, & Downey, 2013). Third, changes in negative affect significantly increased prior to NSSI and reduced following NSSI, whereas positive affect significantly reduced prior to NSSI and increased following NSSI, both approximating a quadratic curve. These changes were not found among participants who did not engage in this behavior. These findings indicate that treatment-seeking youth with BPD experience both positive and negative reinforcement, unlike undergraduate adolescents or community samples of adults with BPD or bulimia (Armey et al., 2011; Muehlenkamp et al., 2009; Snir et al., 2015). These differences might be partially explained by the clinical acuity of this sample, unlike the community samples in prior EMA studies. In comparison with Snir and colleagues (2015), who assessed changes in affect 10 hours pre- and post-NSSI, the current findings might also reflect the longer time frame chosen in this study to capture changes in negative affect. Given negative affect was found to increase a median of 15 hr prior to NSSI in the current study, it is possible that initial changes in negative affect were not captured by Snir and colleagues (2015), which prevented the changes in negative affect from fitting a quadratic model. Changes in positive and negative affect in the current study were supported by explicit motives, with 41% of participants engaging in NSSI to reduce their negative emotions. However, as only 12% of participants endorsed positively reinforcing motives (“feel something,” “relieve an urge/to feel calmer/because I liked it”; Selby, Nock, & Kranzler, 2014), it is possible that reward incentives arising from an increase in positive affect might be a secondary, rather than a primary maintaining factor for NSSI (Kleindienst et al., 2008). This finding suggests that an important step for therapeutic change might be to help clients to acknowledge not only the relief experienced from NSSI but also the rewarding aspects of this behavior. Fourth, the length of time between initial self-injurious thoughts, changes in affect, and engagement in NSSI was highly variable but also longer than expected. Specifically, the median duration between initial changes in positive and negative affect and engagement in NSSI was 10 and 15 hr, respectively. The participants in the current study exhibited a longer time between changes in affect and engagement in NSSI than the 8 hr found in a community sample of youth (Armey et al., 2011). A median of 35 hr also elapsed between initial self-injurious thoughts and engagement in NSSI. It is not possible to determine whether this represents the time spent consciously planning for NSSI or a period of conscious or unconscious rumination about NSSI, which eventually led to engagement in the act. This finding does, how- 7 ever, challenge the notion that NSSI is always an impulsive act and contrasts with evidence that youth with BPD (n ⫽ 30) rate themselves as more impulsive and exhibit a poorer ability to delay gratification on a computerized delayed discounting task compared with healthy controls (Lawrence, Allen, & Chanen, 2010). Instead, the current study suggests that youth with BPD often experience lower levels of urgency to engage in NSSI in the face of worsening affect. It also indicates that they engage in some premeditation and planning, as shown by intermittent thoughts about NSSI identified hours prior to engagement. It is possible that a self-injurious thought itself increased participants’ levels of distress tolerance via the conscious or unconscious knowledge that this self-regulatory behavior was available. This might have reduced the intensity and speed at which negative affect increased, thereby lowering the levels of impulsivity associated with this behavior. Consequent self-injurious thoughts that follow changes in negative affect occurred a median of 2.5 hr prior to NSSI, which indicates that once affect has changed, youth with BPD engage in NSSI relatively quickly after experiencing an additional selfinjurious thought. The duration of time between the self-injurious thought that follows a change in affect and NSSI might be reflective of the conscious preparation time for NSSI reported by adolescents and young adults in Armey and colleagues’ (2011) study and retrospectively by adults with BPD (Snir et al., 2015). There was also a strong association between the length of time from initial self-injurious thoughts to engaging in NSSI, as well as the time from change in affect to engaging in NSSI. This suggests that those who experienced self-injurious thoughts many hours prior to engaging in NSSI also experienced a longer duration of worsening affect than those who acted more quickly on their thoughts about NSSI. Strengths and Limitations A major strength of this study was the recruitment of acutely unwell, help-seeking youth with BPD who had never received evidence-based BPD treatment. This provided the opportunity to assess NSSI early in the course of the disorder, with minimal confounding by treatment and by duration of illness effects. The sample size of 107 participants with BPD is also larger than any previous retrospective, laboratory, or EMA study investigating motives and affect changes related to NSSI in individuals with BPD. Furthermore, this study complements previous research by identifying motives for NSSI using open-ended questions, capturing and analyzing changes in both negative and positive emotions using MLM, and examining a broader window of time to assess the temporal relationship between initial self-injurious thoughts, changes in affect, and engagement in NSSI. A limitation to this study was the small number of participants who engaged in NSSI (n ⫽ 24; 22%). This might be attributable to the relatively short period of data collection in this study compared with other EMA studies (Muehlenkamp et al., 2009; Nock, Prinstein, & Sterba, 2009; Snir et al., 2015; Zaki et al., 2013) that have collected data over a 2- and 3-week period. Although the small sample engaging in NSSI limits the generalizability of the findings, it is likely that extending the window of data collection would have resulted in lowered adherence rates due to the level of clinical acuity in this sample. A second limitation was that the prompts in this study occurred randomly within 2-hr time blocks, which might ANDREWES ET AL. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 8 have prevented some changes in affect from being captured. To determine if additional changes in affect occurred just prior to NSSI, which indicate greater impulsivity, future studies might prompt participants more intensively, yet over a shorter period, so as to not increase participant fatigue. A third limitation was that the time between actual engagement in NSSI and completion of the questionnaire could not be verified because prompts to complete questionnaires occurred randomly. In an attempt to address this limitation in EMA methodology, Snir and colleagues (2015) centered time at the midpoint between reported engagement in NSSI and the prior entry. To better overcome this limitation, future studies could combine “event contingent” assessments, in which participants can initiate questionnaires immediately following an NSSI event, with a question in the survey that asks the time that has elapsed since engaging in NSSI. Summary and Clinical Implications Findings from the current study suggest that both relief from negative affect and reward incentives were found to maintain NSSI, which might explain why NSSI is so prevalent among the BPD population (Wedig et al., 2012). A failure to identify motives and environmental precipitants for NSSI is common among youth with BPD and should be addressed through techniques such as behavioral chain analysis (Linehan, 1993) to facilitate the introduction of cognitive strategies. Finally, the considerable period of time available from the initial self-injurious thoughts and changes in affect to engagement in NSSI suggests that this behavior is not as “impulsive” as might be commonly believed. Rather, this finding should encourage clinicians that there is sufficient time for the introduction of cognitive and behavioral strategies to avert NSSI. References American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text revision). Washington, DC: Author. Armey, M. F. (2012). Ecological momentary assessment and intervention in nonsuicidal self-injury: A novel approach to treatment. Journal of Cognitive Psychotherapy, 26, 299 –317. http://dx.doi.org/10.1891/08898391.26.4.299 Armey, M. F., Crowther, J. H., & Miller, I. W. (2011). Changes in ecological momentary assessment reported affect associated with episodes of nonsuicidal self-injury. Behavior Therapy, 42, 579 –588. http:// dx.doi.org/10.1016/j.beth.2011.01.002 Arnett, J. J. (2000). Emerging adulthood. A theory of development from the late teens through the twenties. American Psychologist, 55, 469 – 480. http://dx.doi.org/10.1037/0003-066X.55.5.469 Arntz, A., van den Hoorn, M., Cornelis, J., Verheul, R., van den Bosch, W. M. C., & de Bie, A. J. H. T. (2003). Reliability and validity of the borderline personality disorder severity index. Journal of Personality Disorders, 17, 45–59. http://dx.doi.org/10.1521/pedi.17.1.45.24053 Bender, D. S., Skodol, A. E., Pagano, M. E., Dyck, I. R., Grilo, C. M., Shea, M. T., . . . Gunderson, J. G. (2006). Prospective assessment of treatment use by patients with personality disorders. Psychiatric Services, 57, 254 –257. http://dx.doi.org/10.1176/appi.ps.57.2.254 Bolger, N., & Laurenceau, P. J. (2013). Intensive longitudinal methods: An introduction to diary and experience sampling research. New York, NY: Guilford. Brown, M. Z., Comtois, K. A., & Linehan, M. M. (2002). Reasons for suicide attempts and nonsuicidal self-injury in women with borderline personality disorder. Journal of Abnormal Psychology, 111, 198 –202. http://dx.doi.org/10.1037/0021-843X.111.1.198 Brunner, R., Parzer, P., Haffner, J., Steen, R., Roos, J., Klett, M., & Resch, F. (2007). Prevalence and psychological correlates of occasional and repetitive deliberate self-harm in adolescents. Archives of Pediatrics & Adolescent Medicine, 161, 641– 649. http://dx.doi.org/10.1001/archpedi .161.7.641 Chanen, A. M. (2015). Borderline personality disorder in young people: Are we there yet? Journal of Clinical Psychology, 71, 778 –791. http:// dx.doi.org/10.1002/jclp.22205 Chanen, A., Jackson, H., Cotton, S. M., Gleeson, J., Davey, C. G., Betts, J., . . . McCutcheon, L. (2015). Comparing three forms of early intervention for youth with borderline personality disorder (the MOBY study): Study protocol for a randomised controlled trial. Trials, 16, 476. http://dx.doi.org/10.1186/s13063-015-1001-x Chanen, A. M., Tackett, J. L., & Thompson, K. (in press). Personality pathology and disorder in children and youth. In J. Livesley & R. Larstone (Ed.), Handbook of personality disorders: Theory, research, and treatment (2nd ed.). New York, NY: Guilford Press Chapman, A. L., & Dixon-Gordon, K. L. (2007). Emotional antecedents and consequences of deliberate self-harm and suicide attempts. Suicide and Life-Threatening Behavior, 37, 543–552. http://dx.doi.org/10.1521/ suli.2007.37.5.543 Chapman, A. L., Specht, M. W., & Cellucci, T. (2005). Borderline personality disorder and deliberate self-harm: Does experiential avoidance play a role? Suicide and Life-Threatening Behavior, 35, 388 –399. http:// dx.doi.org/10.1521/suli.2005.35.4.388 Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Dimaggio, G., & Lysaker, P. H. (2015). Metacognition and mentalizing in the psychotherapy of patients with psychosis and personality disorders. Journal of Clinical Psychology, 71, 117–124. http://dx.doi.org/10.1002/ jclp.22147 Ebner-Priemer, U. W., Kuo, J., Schlotz, W., Kleindienst, N., Rosenthal, M. Z., Detterer, L., . . . Bohus, M. (2008). Distress and affective dysregulation in patients with borderline personality disorder: A psychophysiological ambulatory monitoring study. Journal of Nervous and Mental Disease, 196, 314 –320. http://dx.doi.org/10.1097/NMD .0b013e31816a493f Ebner-Priemer, U. W., Kuo, J., Welch, S. S., Thielgen, T., Witte, S., Bohus, M., & Linehan, M. M. (2006). A valence-dependent groupspecific recall bias of retrospective self-reports: A study of borderline personality disorder in everyday life. Journal of Nervous and Mental Disease, 194, 774 –779. http://dx.doi.org/10.1097/01.nmd.0000239900 .46595.72 First, M., Gibbon, M., Spitzer, R., Williams, J., & Benjamin, L. (1997). Structured Clinical Interview for DSM–IV Axis II Personality Disorders (SCID-II). Washington, DC: American Psychiatric Press. First, M., Spitzer, R., Gibbon, M., & Williams, J. (2002). Structured Clinical Interview for DSM–IV–TR Axis I Disorders, Research Version, Patient Edition (SCID-I/P). New York: Biometrics Research, New York State Psychiatric Institute. Jørgensen, C. R., Berntsen, D., Bech, M., Kjølbye, M., Bennedsen, B. E., & Ramsgaard, S. B. (2012). Identity-related autobiographical memories and cultural life scripts in patients with borderline personality disorder. Consciousness and Cognition: An International Journal, 21, 788 –798. http://dx.doi.org/10.1016/j.concog.2012.01.010 Kercher, K. (1992). Assessing subjective well-being in the old-old: The PANAS as a measure of orthogonal dimensions of positive and negative affect. Research on Aging, 14, 131–168. http://dx.doi.org/10.1177/ 0164027592142001 Kleindienst, N., Bohus, M., Ludäscher, P., Limberger, M. F., Kuenkele, K., Ebner-Priemer, U. W., . . . Schmahl, C. (2008). Motives for nonsuicidal self-injury among women with borderline personality disorder. Journal of Nervous and Mental Disease, 196, 230 –236. http://dx.doi.org/10 .1097/NMD.0b013e3181663026 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. NONSUICIDAL SELF-INJURY IN YOUTH WITH BPD Klonsky, E. D. (2007). The functions of deliberate self-injury: A review of the evidence. Clinical Psychology Review, 27, 226 –239. http://dx.doi .org/10.1016/j.cpr.2006.08.002 Lawrence, K. A., Allen, J. S., & Chanen, A. M. (2010). Impulsivity in borderline personality disorder: Reward-based decision-making and its relationship to emotional distress. Journal of Personality Disorders, 24, 785–799. http://dx.doi.org/10.1521/pedi.2010.24.6.785 Leichsenring, F., Leibing, E., Kruse, J., New, A. S., & Leweke, F. (2011). Borderline personality disorder. The Lancet, 377, 74 – 84. http://dx.doi .org/10.1016/S0140-6736(10)61422-5 Linehan, M. M. (1993). Cognitive behavioural treatment of borderline personality disorder. New York, NY: Guilford. Lobbestael, J., Leurgans, M., Arntz, A., & Wiley, J. (2011). Inter-rater reliability of the Structured Clinical Interview for DSM–IV Axis I Disorders (SCID I) and Axis II Disorders (SCID II). Clinical Psychology & Psychotherapy, 18, 75–79. http://dx.doi.org/10.1002/cpp.693 Maas, C. J. M., & Hox, J. J. (2004). The influence of violations of assumptions on multilevel parameter estimates and their standard errors. Computational Statistics & Data Analysis, 46, 427– 440. http://dx.doi .org/10.1016/j.csda.2003.08.006 Mackinnon, A., Jorm, A., Christensen, H., Korten, A., Jacomb, P., & Rodgers, B. (1999). A short form of the Positive and Negative Affect Schedule: Evaluation of factorial validity and invariance across demographic variables in a community sample. Personality and Individual Differences, 27, 405– 416. http://dx.doi.org/10.1016/S01918869(98)00251-7 Maurex, L., Lekander, M., Nilsonne, A., Andersson, E. E., Asberg, M., & Ohman, A. (2010). Social problem solving, autobiographical memory, trauma, and depression in women with borderline personality disorder and a history of suicide attempts. The British Journal of Clinical Psychology/the British Psychological Society, 49, 327–342. http://dx.doi .org/10.1348/014466509X454831 Moran, P., Coffey, C., Romaniuk, H., Olsson, C., Borschmann, R., Carlin, J. B., & Patton, G. C. (2012). The natural history of self-harm from adolescence to young adulthood: A population-based cohort study. The Lancet, 379, 236 –243. http://dx.doi.org/10.1016/S01406736(11)61141-0 Muehlenkamp, J. J., Engel, S. G., Wadeson, A., Crosby, R. D., Wonderlich, S. A., Simonich, H., & Mitchell, J. E. (2009). Emotional states preceding and following acts of non-suicidal self-injury in bulimia nervosa patients. Behaviour Research and Therapy, 47, 83– 87. http:// dx.doi.org/10.1016/j.brat.2008.10.011 Nelson, E. E., Leibenluft, E., McClure, E. B., & Pine, D. S. (2005). The social re-orientation of adolescence: A neuroscience perspective on the process and its relation to psychopathology. Psychological Medicine, 35, 163–174. http://dx.doi.org/10.1017/S0033291704003915 Nock, M. K. (2009). Why do people hurt themselves? New insights into the nature and functions of self-injury. Current Directions in Psychological Science, 18, 78 – 83. http://dx.doi.org/10.1111/j.1467-8721.2009 .01613.x Nock, M. K., Prinstein, M. J., & Sterba, S. K. (2009). Revealing the form and function of self-injurious thoughts and behaviors: A real-time ecological assessment study among adolescents and young adults. Journal of Abnormal Psychology, 118, 816 – 827. http://dx.doi.org/10.1037/ a0016948 Paus, T. (2005). Mapping brain maturation and cognitive development during adolescence. Trends in Cognitive Sciences, 9, 60 – 68. http://dx .doi.org/10.1016/j.tics.2004.12.008 Peugh, J. L. (2010). A practical guide to multilevel modeling. Journal of School Psychology, 48, 85–112. http://dx.doi.org/10.1016/j.jsp.2009.09 .002 9 Reid, S. C., Kauer, S. D., Dudgeon, P., Sanci, L. A., Shrier, L. A., & Patton, G. C. (2009). A mobile phone program to track young people’s experiences of mood, stress and coping: Development and testing of the mobiletype program. Social Psychiatry and Psychiatric Epidemiology, 44, 501–507. http://dx.doi.org/10.1007/s00127-008-0455-5 Reitz, S., Krause-Utz, A., Pogatzki-Zahn, E. M., Ebner-Priemer, U., Bohus, M., & Schmahl, C. (2012). Stress regulation and incision in borderline personality disorder—A pilot study modeling cutting behavior. Journal of Personality Disorders, 26, 605– 615. http://dx.doi.org/10 .1521/pedi.2012.26.4.605 Sadeh, N., Londahl-Shaller, E. A., Piatigorsky, A., Fordwood, S., Stuart, B. K., McNiel, D. E., . . . Yaeger, A. M. (2014). Functions of nonsuicidal self-injury in adolescents and young adults with borderline personality disorder symptoms. Psychiatry Research, 216, 217–222. http://dx.doi.org/10.1016/j.psychres.2014.02.018 Selby, E. A., Nock, M. K., & Kranzler, A. (2014). How does self-injury feel? Examining automatic positive reinforcement in adolescent selfinjurers with experience sampling. Psychiatry Research, 215, 417– 423. http://dx.doi.org/10.1016/j.psychres.2013.12.005 Shiffman, S., Stone, A. A., & Hufford, M. R. (2008). Ecological momentary assessment. Annual Review of Clinical Psychology, 4, 1–32. http:// dx.doi.org/10.1146/annurev.clinpsy.3.022806.091415 Singer, J., & Willett, J. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York, NY: Oxford University Press. http://dx.doi.org/10.1093/acprof:oso/9780195152968.001.0001 Snir, A., Rafaeli, E., Gadassi, R., Berenson, K., & Downey, G. (2015). Explicit and inferred motives for nonsuicidal self-injurious acts and urges in borderline and avoidant personality disorders. Personality Disorders: Theory, Research, and Treatment, 6, 267–277. http://dx.doi.org/ 10.1037/per0000104 Solhan, M. B., Trull, T. J., Jahng, S., & Wood, P. K. (2009). Clinical assessment of affective instability: Comparing EMA indices, questionnaire reports, and retrospective recall. Psychological Assessment, 21, 425– 436. http://dx.doi.org/10.1037/a0016869 Steinberg, L. (2005). Cognitive and affective development in adolescence. Trends in Cognitive Sciences, 9, 69 –74. http://dx.doi.org/10.1016/j.tics .2004.12.005 Tabachnick, B., & Fidell, L. (2013). Using multivariate statistics (6th ed.). Boston, MA: Pearson. United Nations. (2001). Implementation of the world programme of action for youth to the year 2000 and beyond. Retrieved May 25, 2016, from http://www.youthpolicy.org/basics/2001_WPAY_Imple mentation_Report.pdf Vinson, T. (2007). Dropping off the edge: The distribution of disadvantage in Australia. Sydney, Australia: Jesuit Social Services. Wedig, M. M., Silverman, M. H., Frankenburg, F. R., Reich, D. B., Fitzmaurice, G., & Zanarini, M. C. (2012). Predictors of suicide attempts in patients with borderline personality disorder over 16 years of prospective follow-up. Psychological Medicine, 42, 2395–2404. http://dx .doi.org/10.1017/S0033291712000517 Yuan, K., & Bentler, P. M. (2000). Three likelihood-based methods for mean and covariance structure analysis with nonnormal missing data. Sociological Methodology, 5, 165–200. Zaki, L. F., Coifman, K. G., Rafaeli, E., Berenson, K. R., & Downey, G. (2013). Emotion differentiation as a protective factor against nonsuicidal self-injury in borderline personality disorder. Behavior Therapy, 44, 529 –540. http://dx.doi.org/10.1016/j.beth.2013.04.008 Zanarini, M. C., Frankenburg, F. R., Ridolfi, M. E., Jager-Hyman, S., Hennen, J., & Gunderson, J. G. (2006). Reported childhood onset of self-mutilation among borderline patients. Journal of Personality Disorders, 20, 9 –15. http://dx.doi.org/10.1521/pedi.2006.20.1.9