Dialectical behavior therapy for adolescents (DBT-A): Outcomes among sexual minorities at high risk for suicide

Abstract

The alarming rates and pervasiveness of suicidal and self-destructive behaviors (e.g., non-suicidal self-injury) among young sexual minorities represent a major public health concern. We set out to examine whether an empirically driven treatment for suicide and self-harm, dialectical behavior therapy for adolescents (DBT-A), provides benefits for adolescents who identify as gay, lesbian, bisexual, or questioning (LGBQ). LGBQ adolescents (n = 16) were compared with non-LGBQ peers (n = 23). Psychological measures were collected before and after participation in a comprehensive DBT-A program. LGBQ participants demonstrated significant improvements in emotion regulation, depression, borderline symptoms, and coping strategies; changes were comparable to their heterosexual peers.

INTRODUCTION

Despite ample research, policy, and clinical efforts focused on detection and prevention, suicide rates have increased by 28% over the past 20 years (CDC, 2016a, 2016b; Curtin et al., 2016). Adolescents are particularly vulnerable to suicidal ideation and behaviors; indeed, suicide is the 2nd leading cause of death among 10–24-year olds (CDC, 2015; Heron, 2016) and the number of hospital encounters for suicide-related events among youth has doubled in the past 10 years (Asarnow et al., 2011). Mounting evidence highlights the vulnerability of gay, lesbian, bisexual, or questioning (LGBQ) adolescents, who are disproportionally affected by suicide. Compared with their heterosexual peers, sexual minority adolescents are 1.5–7× more likely to endorse suicidal ideation and are 2–4× more likely to have made a suicide attempt over the last year (Haas et al., 2011; King et al., 2008; Peters et al., 2020; Russell & Joyner, 2001). Furthermore, their suicide attempts tend to be more severe and are more likely to require medical treatment (CDC, 2016a, 2016b).

Sexual minority adolescents are also more likely to engage in non-suicidal self-injury (NSSI), or the intentional infliction of harm to one’s body in the absence of suicidal intent (Nock, 2009). Compared with their heterosexuals, LGBQ adolescents are 3–6x more likely to engage in NSSI (Batejan et al., 2015; Reisner et al., 2014; Liu et al., 2019; Liu & Mustanski, 2012). Importantly, suicidal thoughts and behavior often co-occur, and research consistently finds that NSSI effectively serves as a precursor to subsequent suicide attempts and completions and, as such, represents a well-documented risk factor (e.g., Hankin & Abela, 2011). In fact, NSSI may be an even stronger predictor of suicide attempts than prior attempt history (Asarnow et al., 2011; Ribeiro et al., 2016; Wilkinson et al., 2011). While suicide has increased among adolescents more broadly, overall rates of suicidal behaviors are increasing at a faster rate among sexual minorities in particular (Peter et al., 2017; Saewyc et al., 2008).

Over the past two decades, LGBQ suicide rates have received increasing attention, yet much of the research has focused on identifying risk factors and predictors. Notably, the disparity in suicide risk between LGBQ versus heterosexual youth has widened over this time. Thus, it remains unclear whether current evidence-based interventions for suicide are efficacious for this high-risk group. Several treatments have shown success in reducing suicidal ideation and NSSI behaviors among adolescents (Labelle et al., 2015; Mehlum et al., 2016; O’Brien et al., 2014), yet sexual orientation has not been reliably collected or reported on in these studies. It would be presumptuous to assume that successful outcomes from previous clinical trials, which did not operationalize sexual orientation, generalize to LGBQ youth. The first step in selecting appropriate and accessible treatment options for LGBQ adolescents is to test their response to existing, evidence-based, “gold standard” interventions.

Dialectical behavior therapy (DBT; Linehan, 1993)—initially developed to treat chronically suicidal adults diagnosed with borderline personality disorder (BPD)—has gained strong empirical support for improving outcomes beyond BPD symptomatology. DBT is categorized as a “well-established” treatment in reducing suicidal ideation, attempts, and NSSI among adults in numerous randomized controlled trials (Neacsiu et al., 2012; Robins & Chapman, 2004), which has led to the adaptation of the DBT model for adolescents and their families (DBT-A; Rathus & Miller, 2002). Clinical trials of DBT-A have documented improvements in adolescent suicidal thoughts and behavior, emotion dysregulation, psychopathology, and general psychosocial functioning (Choate, 2012; Fleischhaker et al., 2011; Hollenbaugh & Lenz, 2018; James et al., 2008; McCauley et al., 2018; Mehlum et al., 2016; Miller et al., 2000; Woodberry & Popenoe, 2008). However, none of these studies reported on sexual orientation or whether or not LGBQ youth may differentially respond to DBT.

In addition to the robust empirical support for successfully treating suicidality among adolescents, a strong theoretical justification exists for applying DBT-A to high-risk sexual minorities. DBT is grounded in the Biosocial Model, which contends that emotion dysregulation manifests from the transaction of environmental invalidation and the individual’s biological temperament (Linehan, 1993). Difficulties in emotion regulation (ER) are implicated in virtually all forms of psychopathology (Aldao et al., 2016) and maladaptive behaviors, including suicide and NSSI (Anestis et al., 2011; Carpenter & Trull, 2013). Disparities between sexual minorities and heterosexual peers in ER deficits were shown to emerge as early as middle school (Hatzenbuehler et al., 2008). Hatzenbuehler (2009) posited that LGBQ people are more prone to deficits in ER skills as a result of experiencing chronic stress and stigma related to their sexual orientation, referred to as minority stress (Meyer, 2003). This additional, identity-based stress burden is hypothesized to account for their disparities in negative mental health outcomes. In alignment with the Biosocial Model, sexual minorities likely evidence higher levels of emotion and behavioral dysregulation—even in the absence of a strong biological vulnerability—due to their exposure to persistent cultural, social, and intrapersonal experiences of invalidation, discrimination, and rejection (Meyer, 2003).

Consistent with Meyer’s minority stress theory (Meyer, 2003), sexual minority youth disproportionally experience invalidation, including devaluation by parents and peers. Compared with their heterosexual peers, sexual minorities are 2–4× more likely to be bullied on school property, cyber-bullied, threatened or injured with a weapon on school property, and injured in a physical fight and to experience dating violence (Johns et al., 2020). A meta-analysis demonstrated that sexual minority youth were at increased risk for peer victimization as well as parental abuse (Friedman et al., 2011), both of which constitute extreme forms of invalidation. These identity-based experiences of invalidation continue into adulthood; 55% of sexual minority adults have been verbally harassed and/or threatened, 28% have been physically assaulted, and nearly 10% have been spit on because of their sexual orientation (Katz-Wise & Hyde, 2012).

Importantly, several studies have linked identity-based minority stress experiences—from physical victimization to microaggressions—with higher rates of suicidal thoughts and behaviors (including attempts) and NSSI engagement among LGBQ adolescents and adults alike (Almeida et al., 2009; Bouris et al., 2016; Hamilton & Mahalik, 2009; Hill & Pettit, 2012; Nadal et al., 2011; Smith & Perrin, 2017). While adolescence is associated with the increased salience of peer relationships, the family environment remains particularly vital for healthy socioemotional development (Repetti et al., 2002); therefore, familial acceptance (or lack thereof) may have a significant bearing on the deleterious mental health outcomes observed among LGBQ youth compared with responses from others in adolescents’ social milieu. One study found that LGBQ young adults are 8x more likely to make a suicide attempt if they reported experiencing high levels of family rejection during adolescence due to their sexual minority status (Ryan et al., 2009). DBT-A provides a protocol for developing more adaptive ER strategies and for increasing validating interactions, with caregivers learning skills alongside their teens.

The current study

Efforts to develop treatments for LGBQ suicidal teens are already underway (Budge et al., 2017; Diamond et al., 2012); however, the need for this population to access treatment presently is critical. DBT-A can likely meet the needs of this unique population due to its sensitivity to invalidation and high level of family education and involvement. However, the same factors that contribute to and maintain vulnerabilities to the alarming rates of death by suicide among LGBQ individuals (i.e., lack of family acceptance) could also potentially interfere with treatment effectiveness. Thus, it seems logical to investigate the efficacy of DBT-A for suicidal LGBQ youth to better inform clinicians in making suitable treatment decisions and recommendations. The overarching aim of the current study is to test whether LGBQ adolescents enrolled in a comprehensive DBT-A program demonstrate improvements in psychosocial risk factors associated with suicide and NSSI (e.g., emotion dysregulation; borderline, depressive, and anxiety symptoms) from baseline to posttreatment. We also evaluate how changes in LGBQ participants’ outcomes compare to heterosexual peers. Overall, we predict that both LGBQ and non-LGBQ participants will show significant reductions across all outcomes, including psychological symptoms and risk factors, after participating in DBT-A for 18 weeks.

METHODS

Participants and procedure

Mindful Teen is an 18-week comprehensive DBT-A outpatient program offered to adolescents between the ages of 13–18 within a children’s psychiatric hospital setting in the Northeast. Inclusion criteria included a documented history of suicidality, NSSI, and/or other self-destructive behaviors, placing them at risk for needing a higher than outpatient-based level of care. DBT-A was delivered with fidelity to the standard model proposed by Miller et al. (2007), including a weekly multifamily skills training group, individual therapy, 24/7 phone coaching, and a therapist consultation team. Individual treatment was delivered by doctoral-level providers intensively trained in DBT and highly experienced with multiproblem youth. At least one guardian was required to participate alongside their adolescent in skills training. To ensure efficacy and promote commitment to the treatment, participants were required to not miss more than four groups or individual sessions.

Participants were recruited to complete an optional treatment effectiveness study after they enrolled in the Mindful Teen program, and routine informed parental consent and child assent procedures were followed. All adolescents receiving treatment were eligible to participate in this study; thus, inclusion criteria were broad and participants qualified if they met age restrictions, failed at lower levels of treatment, and were able to reliably attend all treatment components. Those included in the study were asked to complete a battery of self and parent report questionnaires at pretreatment, posttreatment, and 6-month follow-up. To optimize sample size, current analyses focused on pre- and posttreatment self-report measures. Surveys were primarily delivered via REDCap, a secure online data collection application, with paper-and-pencil options offered if electronic means were not accessible (Harris et al., 2009).

The current study included 16 participants who self-identified as LGBQ and 23 who identified as heterosexual who completed baseline and posttreatment measures. Sexual orientation was assessed at baseline by prompting the selection of one of four options including straight, gay/lesbian, bisexual, or questioning/other. The final sample included 39 adolescents (age M = 15.21 years, SD = 1.65), most of whom identified as female (86.8%). Race/ethnicity was representative of the community, with 71.1% non-Hispanic White, 22.9% Hispanic, 13.1% bi- or multi-racial, and 7.9% Asian, African-American, or other. All participants completed the entire 18-week DBT-A program along with their guardian(s). There were no significant demographic differences between LGBQ and non-LGBQ groups.

Measures

Difficulties in Emotion Regulation Scale (DERS; Gratz & Roemer, 2004)

The DERS is a 36-item questionnaire used to measure an individual’s ability to regulate their emotions consisting of six subscales that reflect different dimensions and approaches to understand emotion regulation. Total scores range from 36 to 180, such that higher scores reflect greater difficulties with effective emotion regulation. The DERS has demonstrated high internal consistency, test–retest reliability, and adequate criterion-related validity in standardized (Gratz & Roemer, 2004) and clinical samples (Osbourne et al., 2017).

Beck Depression Inventory, 2nd edition (BDI-II; Beck et al., 1996)

The BDI-II is a widely used self-report inventory used to assess for general symptoms of depression. The 21 items are each rated using a 0 to 3 scale, with higher scores indicating more severe depressive symptoms.

Beck Anxiety Inventory (Beck & Steer, 1993)

The BAI consists of 21 items and was constructed to measure severity of anxiety symptoms. The items ask about common physical symptoms associated with anxiety, and higher scores represent greater anxiety severity.

The dialectical behavior therapy ways of coping checklist (DBT-WCCL; Neacsiu et al., 2010)

The DBT-WCCL was developed to assess acquisition and application of skills specific to DBT by using lay language. It contains 59 items and is represented by two factors: adaptive skill use and dysfunctional coping. Previous reports indicate good to excellent psychometric properties (Neacsiu et al., 2010).

Borderline symptoms list (BSL; Bohus et al., 2007)

The BSL measures borderline-typical symptomatology. The 35 items are based on the DSM symptoms for BPD, with higher scores indicating greater impairment. Psychometric properties include high internal consistency, test–retest reliability, and convergent validity (Bohus et al., 2007).

Data analysis

First, repeated-measures bootstrapped t-tests were conducted to examine changes in pre- and posttreatment scores following 18 weeks of treatment for LGBQ participants only. Two-tailed 0.05 p-values were applied for all analyses. In the event that one item was missing on a self-report scale, the mean score of the remaining items was used in place of that item. A 2 × 2 mixed-model ANOVA was performed to compare treatment effects across LGBQ and non-LGBQ groups. Specifically, interactions were examined to test group (LGBQ and non-LGBQ) by time (pre-and post-) effects on each of the six outcome measures. Considering the small sample size and exploratory nature of the analyses, partial eta squared was used to assess the size of the interaction effect (η² = 0.01 small; η² = 0.06 medium; and η² = 0.14 large). Cohen’s d was also reported for treatment effects (d = 0.20 small; d = 0.50 medium; and d = 0.80 large).

RESULTS

DBT-A resulted in significant posttreatment improvements in participants who identified as LGBQ on all outcomes assessed, with the exception of anxiety (Table 1). Specifically, LGBQ participants demonstrated significant reductions in ER problems as well as depressive and borderline symptoms. In terms of DBT skills, repeated-measures t-tests indicated that LGBQ adolescents successfully acquired healthy coping skills and decreased dysfunctional coping as measured by the DBT-WCCL. Effect sizes were all considered to be in the large range (Table 1).

TABLE 1.

Paired-sample tests for pre-and posttreatment scores in LGBQ participants

	Pretreatment		Posttreatment
Measures	Mean	SD	Mean	SD	t Test	Cohen’s d
DERS	122.35	16.16	81.41	23.52	6.78***	1.90
BDI-II	28.64	14.61	15.77	13.52	4.68**	1.05
BAI	26.70	12.31	20.77	12.80	1.78	0.57
WCCL-Skill Use	1.40	0.52	1.83	0.47	−2.71*	0.67
WCCL-D ysfunctional Coping	1.72	0.50	1.28	0.55	2.70*	0.89
BSL	45.15	21.60	23.46	18.05	2.94**	0.95

^*p < 0.05

^**p < 0.01

^***p < 0.001.

Between-subjects main effects on the ANOVA showed that using both time points, LGBQ participants rated significantly higher on borderline symptom (F(1, 34) = 18.13, p < 0.01) measures. There were no significant differences between groups on the BDI-II, BAI, DERS, or DBT-WCCL. The mixed-model ANOVA did not produce significant group × time interaction effects on any of the outcomes, indicating that changes over time did not differ between LGBQ and non-LGBQ participants. In other words, DBT-A participation led to similar rates of improvement among LGBQ adolescents as their heterosexual peers. While statistically nonsignificant, interaction effect sizes on the DERS, BDI-II, and WCCL-Skill Use were small to medium, suggesting that sexual minorities may benefit slightly more with respect to ER, depression, and effective skill use compared with their heterosexual peers. Interactions are illustrated in Figure 1, and a complete report of the interaction statistics is in Table 2.

FIGURE 1.

Interactions of group (LGBQ, non-LGBQ) by time (pre, post) on six outcome measures

TABLE 2.

Main and Interaction effects between LGBQ and non-LGBQ participants across time

	Factor: time	LGB status × Time
Measures	F	Partial eta squared	F	Partial eta squared
DERS	46.71***	0.57	0.65	0.02
BDI-II	10.65**	0.23	0.85	0.02
BAI	2.30	0.06	0.44	0.01
WCCL-Skill Use	27.12***	0.43	1.35	0.04
WCCL-Dysfunctional Coping	14.05**	0.28	0.05	<0.01
BSL	18.13***	0.35	<0.01	0.00

^*p < 0.05

^**p < 0.01

^***p < 0.001.

η² = 0.01 small; η² = 0.06 medium; η² = 0.14 large.

DISCUSSION

Suicide is a leading cause of death among adolescents, and rates of ideation and attempts are disproportionately high among adolescent sexual minority populations. This study was set up to preliminarily examine whether DBT-A would be specifically effective for youth who identify as LGB and/or who are actively questioning their sexuality. Results indicate that DBT-A had desirable effects on ER, depressive, and borderline symptoms, in addition to increases in skill acquisition and decreases in dysfunctional coping from pre- to posttreatment comparable to their heterosexual counterparts. These findings are consistent with the growing literature on DBT-A as an efficacious and reliable treatment for adolescents (Choate, 2012; Hollenbaugh, & Lenz, 2018; McCauley et al., 2018; Mehlum et al., 2016) and has additional implications for clinicians working with this vulnerable population. Accordingly, it may be the case that LGBQ adolescents who demonstrate self-destructive behaviors can benefit from DBT-A in its current form, although this requires further attention given the small sample of the current study.

The finding that LGBQ participants reported significantly elevated borderline symptoms compared with non-LGBQ individuals, when averaged across time, is consistent with prior research showing higher rates of psychopathology among sexual minorities (Bostwick et al., 2010; King et al., 2008). Increased risk for suicide and NSSI may even be explained by these higher prevalence rates (Fergusson et al., 1999). However, recent research on psychiatric samples of adults indicates that sexual minorities are almost 2.5 times more likely to receive a diagnosis of borderline personality disorder even after taking into account robust clinical correlates of BPD (e.g., age, gender, PTSD, and maladaptive personality domains; Rodriguez-Seijas et al., 2020). This phenomenon has yet to be examined among adolescents, and it is unclear whether this diagnostic disparity reflects clinician bias and/or the conflation of BPD symptoms with normative exploration of identity concerns and experiences of minority stress. Nevertheless, our data suggest that affective symptoms are more pronounced among LGBQ adolescents at initial presentation; however, their higher levels of pathology do not impede their treatment gains. Therefore, it may be appropriate for clinicians to approach the specific needs of LGBQ adolescents by implementing the DBT model.

Presently, there is a large gap in empirically driven treatments geared toward suicidal and self-harming LGBQ individuals. While efforts to address this gap are already underway, capitalizing on extant gold standard, evidence-supported treatments that reduce suicide risk may increase accessibility more rapidly. Previously, we were unable to assume that what has worked for heteronormative populations would confer the same favorable benefits for sexual minorities. Our study showed clinically significant changes across a range of behavioral and symptom-based outcomes among LGBQ adolescents enrolled in a heterogeneous DBT-A program. Effect sizes suggest that DBT-A may benefit LGBQ participants in specific domains to a greater extent than their heterosexual peers. The results of this study provide preliminary—albeit promising—support for the use of DBT to decrease suffering among sexual minority youth. Future studies should seek to conduct more outcome and dissemination research on DBT-A within this and related minority populations. Clinical scientists should routinely and comprehensively assess both sexual orientation and gender identity; indeed, a wealth of research suggests that transgender and other gender-expansive minorities experience increased risk of suicidal thoughts and behaviors across the lifespan (Janakiraman et al., 2020). New adaptations to address the unique experiences of sexual and/or gender minorities related to suicide could be developed and tested to see whether larger effects are feasible across multiple settings (Carmel et al., 2014; Dimeff et al., 2015). Our results speak to the possible benefits of incorporating DBT into traditional training methods and may be effective in preparing a myriad of mental health workers for high-risk, diverse cases.

While our results are encouraging that LGBQ youth may benefit from DBT-A, they should still be considered as preliminary. Due to our relatively low N, our analyses were underpowered, resulting in a lower likelihood of detecting true effects and a greater risk of producing unreliable (or overestimated) effects. Therefore, the results of the present exploratory study should be interpreted with caution in light of this important issue. Furthermore, future studies with larger sample sizes should be designed in order to increase statistical power and allow for the examination of potential moderators (e.g., family acceptance and LGBQ community connectedness). Our large effect sizes certainly stand out when considering these limitations, and the fact that our participants had unsuccessfully utilized many higher levels of care.

To our knowledge, this is the first study to investigate whether DBT-A is an effective treatment for LGBQ adolescents with suicide and/or NSSI behaviors. Our results are promising, such that LGBQ participants significantly benefitted from treatment and demonstrated comparable changes to non-LGBQ participants. Thus, we urge clinicians and researchers alike to further investigate DBT-A within this high-risk population and help LGBQ youth build lives worth living.