Accepted Manuscript Title: The Trajectories of Adolescent Anxiety and Depressive Symptoms Over Course of a Transdiagnostic Treatment Author: Alexander H. Queen David H. Barlow Jill Ehrenreich-May PII: DOI: Reference:
S0887-6185(14)00069-3 http://dx.doi.org/doi:10.1016/j.janxdis.2014.05.007 ANXDIS 1606
To appear in:
Journal of Anxiety Disorders
Received date: Accepted date:
31-1-2014 12-5-2014
Please cite this article as: Queen, Alexander H., Barlow, David H., & EhrenreichMay, Jill., The Trajectories of Adolescent Anxiety and Depressive Symptoms Over Course of a Transdiagnostic Treatment.Journal of Anxiety Disorders http://dx.doi.org/10.1016/j.janxdis.2014.05.007 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
cr
ip t
Running Head: TRAJECTORIES OF ANXIETY AND DEPRESSION
The Trajectories of Adolescent Anxiety and Depressive Symptoms
an
Alexander H. Queen
us
Over the Course of a Transdiagnostic Treatment
University of Miami
M
David H. Barlow
Center for Anxiety and Related Disorders at Boston University
d
Jill Ehrenreich-May
Ac ce pt e
University of Miami
Author Note
Alexander H. Queen & Jill Ehrenreich-May, Department of Psychology, University of Miami David H. Barlow, Center for Anxiety and Related Disorders at Boston University
Correspondence concerning this manuscript should be addressed to Jill Ehrenreich-May, Department of Psychology, University of Miami, Flipse 315, 5665 Ponce de Leon Boulevard, Coral Gables, FL 33146. E-mail: [email protected]
Page 1 of 36
2 TRAJECTORIES OF ANXIETY AND DEPRESSION Abstract Anxiety and depressive disorders commonly co-occur during adolescence, share multiple vulnerability factors, and respond to similar psychosocial and pharmacological interventions.
ip t
However, anxiety and depression may also be considered distinct constructs and differ on some underlying properties. Prior research efforts on evidence-based treatments for youth have been
cr
unable to examine the concurrent trajectories of primary anxiety and depressive concerns across
us
the course of treatment. The advent of transdiagnostic approaches for these emotional disorders in youth allows for such examination Thepresent study examined the separate trajectories of
an
adolescent anxiety and depressive symptoms over the course of a transdiagnostic intervention, the Unified Protocol for the Treatment of Emotional Disorders in Adolescence (UP-A;
M
Ehrenreich et al., 2008), as well as up to six months following treatment. The sample included 59 adolescents ages 12-17 years old (M = 15.42, SD = 1.71) who completed at least eight sessions
d
of the UP-A as part of an open trial or randomized, controlled trial across two treatment sites.
Ac ce pt e
Piecewise latent growth curve analyses foundadolescent self-rated anxiety and depressive symptoms showed similar rates of improvement during treatment, but while anxiety symptoms continued to improve during follow-up, depressive symptoms showed non-significant improvementafter treatment. Parent-rated symptoms also showed similar rates of improvement for anxiety and depression during the UP-A to those observed for adolescent self-report, but little improvement after treatment across either anxiety or depressive symptoms. To a certain degree, the results mirror those observed among other evidence-based treatments for youth with anxiety and depression, though results hold implications for future iterations of transdiagnostic treatments regarding optimization of outcomes for adolescents with depressive symptoms. Keywords: anxiety, depression, adolescence, transdiagnostic, treatment
Page 2 of 36
3 TRAJECTORIES OF ANXIETY AND DEPRESSION 1. Introduction Anxiety disorders are the most prevalent psychiatric disorders in childhood and adolescence, with prevalence estimates of 10-21% in the general population in the United States
ip t
(Costello, Mustillo, Erkanli, Keeler, & Angold, 2003). Considered through the lens of DSM-IV (American Psychiatric Association, 2000), unipolar depressive disorders (i.e., major depressive
cr
disorder [MDD], dysthymic disorder) as a whole are also common mental health conditions, and
us
become more prevalent during adolescence, as compared to earlier in development (Costello et al., 2002). Anxiety and depressive disorders commonly co-occur with one another in
an
adolescence. Between 16% and 62% of youth with an anxiety disorder also meet criteria for depression, with the highest comorbidity rates found among treatment-seeking adolescents
M
(Brady & Kendall, 1992; Ollendick, Shortt, & Sander, 2005). In addition, self-report measures of youth anxiety and depressive symptoms show moderate correlations with one another (e.g., r =
d
0.34), even after controlling for overlapping items on these instruments (Stark & Laurent, 2001).
Ac ce pt e
In addition to their high comorbidity with one another, youth anxiety and depression share a number of biological, environmental, and psychological risk factors (for a more thorough review, see Garber & Weersing, 2010). For instance, behavioral inhibition in early childhood is a risk factor for the later development of both anxiety and depression (Kagan, Reznick, & Snidman, 1987), and both anxiety and depressive disorders are associated with neuroendocrine (Dahl et al., 2000; Weems, Zakem, Costa, Cannon, & Watts, 2005) and neurotransmitter dysregulation (Flores et al., 2004; Fox et al., 2005). In addition, high negative affect (NA) has shown to be a latent factor underlying all of the anxiety and depressive disorders (Brown, Chorpita, & Barlow, 1998; Trosper, Whitton, Brown, & Pincus, 2012).
Page 3 of 36
4 TRAJECTORIES OF ANXIETY AND DEPRESSION Youth anxiety and depressive disorders also demonstrate similar responses to pharmacological and psychosocial interventions. For instance, both anxiety and depression are responsive to selective serotonin reuptake inhibitor (SSRI) medications (e.g., TADS, 2004;
ip t
Walkup et al., 2008). Prior work with cognitive-behavioral therapy (CBT) trials have found “spill-over” effects onto comorbid anxiety or depressive disorders, regardless of the principal
cr
disorder. For example, anxiety-focused CBT has demonstrated positive effects on comorbid
us
depressive symptoms (Kendall, Safford, Flannery-Schroeder, & Webb, 2004), and a metaanalysis of CBT for youth depression found effect sizes in anxiety symptom reduction (ES =
an
0.39) that were only slightly less than those for depressive symptom improvement (ES = 0.57; Weisz, McCarty, & Valeri, 2006).
M
Given their frequent comorbidity, shared vulnerability factors, and similar response to treatment, some (e.g., Barlow et al., 2004) have advocated a transdiagnostic or disorder-non-
d
specific treatment approach that targets higher-order psychological factors common to the
Ac ce pt e
emotional disorders. Such an approach is hypothesized to allow for greater clinical flexibility and use with patients presenting with multiple emotional disorders, as well as reduce clinician burden in learning multiple, disorder-specific treatment manuals (McHugh & Barlow, 2010). As such, recent clinical research has focused upon the development and evaluation of transdiagnostic treatments that may effectively target anxiety and depressive disorders within a single protocol. The Unified Protocol for the Treatment of Emotional Disorders in Adolescence (UP-A; Ehrenreich et al., 2008) is a developmental adaptation of the adult Unified Protocol (UP; Barlow et al., 2010), designed for adolescents ages 12-17 years old presenting with any primary anxiety disorder, unipolar depressive disorder, or their combination. The UP-A has theoretical roots in emotion regulation, cognitive science, and behavior modification, and distills common evidence-
Page 4 of 36
5 TRAJECTORIES OF ANXIETY AND DEPRESSION based techniques that cut across disorder-specific treatment manuals for youth anxiety and depression (e.g., psychoeducation, non-judgmental awareness, cognitive reappraisal, exposure, behavioral activation, etc.) within a singular protocol. The UP-A incorporates standard evidence-
ip t
based principles within the broader function, context, and regulation of a range of positive and negative emotions (e.g., sadness, anger, fear). Therefore, it is theorized to positively effect how
cr
adolescents attend to, modulate, and experience emotions that cut across specific disorders.
us
Similar to the UP, the UP-A targets five higher-order principles thought to be latent constructs underlying lower-order specific anxiety and depressive disorders: (1) increase present-focused
an
awareness of emotions, (2) enhance cognitive flexibility, (3) prevent emotional avoidance and maladaptive emotion-driven behaviors, (4) increase acceptance of uncomfortable emotion-
M
related physiological sensations, and (5) facilitate exposure to bodily and environmental triggers of emotional experiences (Barlow et al., 2010).
d
A prior open trial of the UP-A established initial efficacy, with subjects demonstrating
Ac ce pt e
significant pre-post reductions in clinician-rated diagnostic severity across anxiety and depressive disorder diagnoses (Trosper, Buzzella, Bennett, & Ehrenreich, 2009), and an immediate treatment (TX) condition of the UP-A has found to outperform an 8-week, treatmentdelayed waitlist (WL) condition in clinician-rated diagnostic severity for the principal disorder, in a recently completed randomized controlled trial (RCT; Ehrenreich-May, Queen, Rodriguez, & Rosenfield, 2012). Analyses from this RCT also found that the presence of a depressive disorder did not moderate treatment outcomes in the UP-A (Ehrenreich-May et al., 2012), whereas many previous CBT trials for youth anxiety have found poorer outcomes for patients with comorbid depression (e.g., Berman et al., 2000; Ginsburg et al., 2011; O’Neil & Kendall, 2012).
Page 5 of 36
6 TRAJECTORIES OF ANXIETY AND DEPRESSION To summarize, youth anxiety and depression are known to commonly co-occur with one another, share multiple vulnerability factors, and may be effectively treated with a unified treatment approach. However, despite their similarities, anxiety and depression have also shown
ip t
to be distinct constructs. For instance, factor analytic studies with school-based (Crowley & Emerson, 1996) and clinical samples (Stavrakaki, Vargo, Boodoosingh, & Roberts, 1987) have
cr
found stronger support for two-factor models of anxiety and depression compared to single
us
factor models. In addition, while both anxiety and depression are characterized by high negative affect, low positive affect has shown stronger associations with depressive symptoms than with
an
anxiety symptoms (for more comprehensive reviews, see Anderson & Hope, 2008; De Bolle & De Fruyt, 2010). Given these important differences, a next step in investigating transdiagnostic
M
treatment approaches is to examine the separate trajectories of symptom change for anxiety and
rates of change.
d
depression over the course of treatment, in order to assess if they show similar or differential
Ac ce pt e
The present study examined the separate trajectories of anxiety and depressive symptoms over the course of the UP-A, and up to six months following the end of treatment, for adolescent subjects completing the UP-A as part of the open trial or RCT investigation. We used piecewise latent growth curve modeling (LGCM) to model these trajectories over two separate time periods: during the course of treatment (“treatment slope”) and up to six months after treatment ended (“follow-up slope”). Piecewise LGCM is often recommended when examining change during treatment and follow-up, given likely non-linear change (Brown, 2004).Separate models were conducted for anxiety and depressive symptoms. Separate models were also conducted for self-rated and parent-rated symptoms, in order to examine informant differences in symptom change trajectories. Therefore, a total of four piecewise LGCMs were conducted.
Page 6 of 36
7 TRAJECTORIES OF ANXIETY AND DEPRESSION 2. Method 2. 1 Participants Participants were 59 adolescents (57.6% female) ages 12-17 years old (M = 15.42, SD =
ip t
1.71) who received at least eight sessions of the UP-A and completed at least one post-baseline assessment. Given the aim of the study was to examine separate trajectories of anxiety and
cr
depression symptom change for those completing the intervention, we decided to restrict
us
analyses to those receiving at least 8 sessions as this represented the minimum dosage possible to be considered a treatment completer. This subsample of 59 participants was culled from a total
an
sample of 67 participants who were enrolled in either the open trial or RCT investigation of the UP-A. Eight (11.94%) of the 67participants that did not complete at least eight sessions of the
M
intervention were part of the open trial (n = 2) or RCT (n = 6), respectively, and did not have any post-baseline assessment data. T-test and chi-square analyses revealed that those completing at
d
least eight sessions (n = 59) did not significantly differ from those who dropped out prior to eight
Ac ce pt e
sessions (n = 8) with regard to age, gender, ethnicity, severity of principal diagnosis, depressive disorder comorbidity status, or baseline levels of anxiety or depressive symptoms (child or parent report).
Participants were evenly divided between Hispanic/Latino (n = 26; 44.1%) and White, Non-Hispanic ethnicities (n = 26; 44.1%). The remaining subjects identified themselves as having Black/African-American (n = 2; 3.4%), Asian-American (n = 1; 1.7%), and “other” ethnicity (n = 4; 6.8%). The median reported annual family income was $100,000 (SD = $80,000). The majority of participants had parents who were married (n= 40; 67.8%). Remaining participants had parents who were separated/divorced (n = 13; 22.03%), widowed (n = 2; 3.4%), or never married (n = 4; 6.78%).
Page 7 of 36
8 TRAJECTORIES OF ANXIETY AND DEPRESSION Eligibility requirements for both the open trial and RCT included being between 12-17 years old (participants could be 18 if they were still in high school) and having a primary diagnosis of any DSM-IV anxiety and/or unipolar depressive disorder. Adolescents with other
ip t
psychiatric difficulties, including attention-deficit/hyperactivity disorder (ADHD), eating disorders, and non-treatment-interfering substance abuse were eligible for participation.
cr
However, the primary diagnosis had to be an anxiety or depressive disorder. Exclusion criteria
us
included bipolar disorder, treatment-interfering substance abuse, severe suicidal ideation or recent psychiatric hospitalization, treatment-interfering cognitive functioning (e.g., IQ below 80),
an
and a prior course of CBT for anxiety or depression. Participants and at least one parent were also required to read and comprehend English well enough to complete study measures. Those
M
concurrently receiving psychiatric medication were required to have been on a stable dosage of
prior to study enrollment.
d
an SSRI medication for three months and/or a stable dosage of a benzodiazepine for one month
Ac ce pt e
Adolescents completing the UP-A as part of the open trial participated at a universitybased clinic in a large urban area in the Northeastern United States (n = 15), while those in a subsequent RCT participated at a university-based clinic in a large urban area in the Southeastern United States (n = 44). There were no significant site differences with regard to gender, depressive disorder comorbidity, treatment length, or severity of anxiety and depressive symptoms. However, subjects in the RCT were slightly older (M = 15.69, SD = 1.70) than open trial participants (M = 14.63, SD = 1.55), t(57) = 2.13, p = .04, d = 0.64. In addition, as expected based upon demographic characteristics of the surrounding communities, RCT participants were more likely to be Hispanic/Latino (61.36%), whereas all of the open trial participants were White, Non-Hispanic, χ2 (3) = 25.53, p .20. Participants in the WL arm then began the UP-A and completed assessments at the same
M
intervals as those in the open trial and TX arm of the RCT.
2.2.1 Treatment structure and content.The UP-A follows a principle-based, flexible
d
treatment structure. The intervention can be delivered over a varying length of time (between 8-
Ac ce pt e
21 weekly sessions administered in an individual format). There are five required modules, which correspond to the five core principles underlining the UP/UP-A (Barlow et al., 2010). Therapists are allowed to devote more sessions to a required module dependent upon patient needs, but ideally all modules are completed during the course of treatment. Therefore, while treatment length may differ among patients, all subjects receive the same five core modules and relevant skills (e.g., psychoeducation, non-judgmental awareness/mindfulness, cognitive reappraisal, exposure, behavioral activation). See Table 2 for a display of the five required modules, suggested session length for each module, including treatment skills delivered within each module and the corresponding UP/UP-A core principles targeted. Importantly, one of the unique features of the UP models is the ability to target emotion regulation deficits across a
Page 12 of 36
13 TRAJECTORIES OF ANXIETY AND DEPRESSION broader range of positive and negative emotions, compared to traditional, disorder-specific treatment manuals. In addition to the five required modules, three optional modules are available to the
ip t
therapist as needed to address parenting skills, motivational enhancement, and clinical deterioration, and can be used at any point in treatment. At least one parent is required to be
cr
actively involved in treatment. While some variability is allowable, typically a parent is involved
us
in the last 10-15 minutes of every session to review content and the assigned homework. Individual sessions with the parent may be used as part of an optional module if the therapist
an
feels that certain parenting behaviors (e.g., overprotection, high criticism, parent-teen conflict, etc.) are counterproductive to treatment progress or as needed for exposure planning.
M
Motivational interviewing (Miller & Rollnick, 2002) techniques can be used in sessions as part of a second optional module for adolescents who appear reluctant to engage in treatment. Finally,
d
sessions within the third optional module can be used to develop a safety plan with the
Ac ce pt e
adolescent and his or her parent in the event the patient experiences impulses for self-harm. However, adolescents who show immediate risk for suicide or other pronounced clinical deterioration could also be discontinued from the UP-A studies and referred for more intensive services, if appropriate. 2.3 Measures
Anxiety Disorders Interview Schedule for the DSM-IV-Child/Parent Version(ADIS-IVC/P; Silverman & Albano, 1996). The ADIS-IV-C/P was completed at the adolescent’s baseline assessment to determine diagnostic eligibility. The ADIS-IV-C/P is a semi-structured diagnostic interview for children and adolescents ages 7-17 years that assesses all DSM-IV anxiety disorders. The ADIS-IV-C/P also assesses for unipolar depressive disorders and externalizing
Page 13 of 36
14 TRAJECTORIES OF ANXIETY AND DEPRESSION disorders (i.e., ADHD, oppositional-defiant disorder, conduct disorder), and screens for substance abuse, schizophrenia, pervasive developmental disorders, eating disorders, and somatoform disorders. In addition, participants were screened for bipolar disorder. Inter-rater
ip t
reliability for the primary diagnosis within this sample was very good (κ = .82, p .05), CFI > .95, RMSEA < .06, and SRMR < .08 (Kline, 2011). Significant tests for parameter estimates were conducted with the z-statistic using a two-tailed test. LGCM analyses were conducted using MPLUS, Fifth Version (Muthen & Muthen, 2005). 3. Results
Page 15 of 36
16 TRAJECTORIES OF ANXIETY AND DEPRESSION The data were first screened for normality and to determine if there were any statistical outliers. Skewness and kurtosis levels were within acceptable ranges (Kline, 2011). No significant outliers (z ≥ 3) were identified at any time points. All subjects had complete data for
ip t
Time 1 as well as at least one other post-baseline assessment. Fifty-five subjects (93.22%) had data at the 8-week assessment (Time 2) and 48 participants (81.36%) had data at the end of
cr
treatment (M = 15.58 sessions) assessment (Time 3). However, there was considerable missing
us
data at the two follow-up time points, with 29 subjects (49.15%) having available data three months after treatment (Time 4), and 24 participants (40.68%) having data six months after
an
treatment (Time 5), although 62.71% of subjects had available data for at least one follow-up time point. Attrition analyses revealed no significant differences between those with complete
M
data and those with missing data on any indicators of symptom severity at baseline or any demographic variables (e.g., age, gender, ethnicity). Furthermore, those with and without follow-
Ac ce pt e
3.1 Descriptive Statistics
d
up data did not show significant differences on any variables at the post-treatment assessment.
See Table 3 for observed means and standard deviations of RCADS/RCADS-P Total Anxiety Tscores and RCADS/RCADS-P MDD T-scores at each time point. In addition, the percentage of participants with T-scores 65 and T-scores 70 are presented. Mean T-scores on all four measures reduced to below 65 by Time 3, indicating that mean levels of both self-rated and parent-rated anxiety and depressive symptoms were within the normative range by the end of treatment. Analysis of percentages of participants with T-scores 65 and T-scores 70 also revealed reductions in those with borderline elevated and clinically elevated anxiety and depressive symptoms from baseline to the end of treatment, which were largely maintained during follow-up.
Page 16 of 36
17 TRAJECTORIES OF ANXIETY AND DEPRESSION 3.2 LGCMs for Adolescent Self-Rated Symptoms 3.2.1 Anxiety symptoms.Self-rated anxiety and depressive symptom trajectories (observed and estimated) are displayed in Figure 2. The piecewise LGCM for self-rated anxiety symptoms was
ip t
modeled by specifying three latent factors: an intercept (baseline levels of anxiety symptoms), a treatment slope factor (change during the UP-A), and a follow-up slope factor (change during
cr
follow-up). All loadings from the intercept to observed variables were fixed at 1. Loadings for
us
the treatment slope factor were fixed at 0, -1, -2, -2, and -2. Loadings for the follow-up slope factor were fixed at 0, 0, 0, -1.5, and -3. These loadings can be interpreted as a period of eight
an
weeks per one unit change. The final model imposing equality constraints for the residual variances of the first three indicators and last two indicators demonstrated acceptable model fit
M
by most indices, χ2 (9) = 13.63, p = 0.14, CFI = 0.97, RMSEA = 0.09, SRMR = 0.19. There was a significant mean intercept (Mi = 53.68, SE = 1.65, p
S0887-6185(14)00069-3 http://dx.doi.org/doi:10.1016/j.janxdis.2014.05.007 ANXDIS 1606
To appear in:
Journal of Anxiety Disorders
Received date: Accepted date:
31-1-2014 12-5-2014
Please cite this article as: Queen, Alexander H., Barlow, David H., & EhrenreichMay, Jill., The Trajectories of Adolescent Anxiety and Depressive Symptoms Over Course of a Transdiagnostic Treatment.Journal of Anxiety Disorders http://dx.doi.org/10.1016/j.janxdis.2014.05.007 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
cr
ip t
Running Head: TRAJECTORIES OF ANXIETY AND DEPRESSION
The Trajectories of Adolescent Anxiety and Depressive Symptoms
an
Alexander H. Queen
us
Over the Course of a Transdiagnostic Treatment
University of Miami
M
David H. Barlow
Center for Anxiety and Related Disorders at Boston University
d
Jill Ehrenreich-May
Ac ce pt e
University of Miami
Author Note
Alexander H. Queen & Jill Ehrenreich-May, Department of Psychology, University of Miami David H. Barlow, Center for Anxiety and Related Disorders at Boston University
Correspondence concerning this manuscript should be addressed to Jill Ehrenreich-May, Department of Psychology, University of Miami, Flipse 315, 5665 Ponce de Leon Boulevard, Coral Gables, FL 33146. E-mail: [email protected]
Page 1 of 36
2 TRAJECTORIES OF ANXIETY AND DEPRESSION Abstract Anxiety and depressive disorders commonly co-occur during adolescence, share multiple vulnerability factors, and respond to similar psychosocial and pharmacological interventions.
ip t
However, anxiety and depression may also be considered distinct constructs and differ on some underlying properties. Prior research efforts on evidence-based treatments for youth have been
cr
unable to examine the concurrent trajectories of primary anxiety and depressive concerns across
us
the course of treatment. The advent of transdiagnostic approaches for these emotional disorders in youth allows for such examination Thepresent study examined the separate trajectories of
an
adolescent anxiety and depressive symptoms over the course of a transdiagnostic intervention, the Unified Protocol for the Treatment of Emotional Disorders in Adolescence (UP-A;
M
Ehrenreich et al., 2008), as well as up to six months following treatment. The sample included 59 adolescents ages 12-17 years old (M = 15.42, SD = 1.71) who completed at least eight sessions
d
of the UP-A as part of an open trial or randomized, controlled trial across two treatment sites.
Ac ce pt e
Piecewise latent growth curve analyses foundadolescent self-rated anxiety and depressive symptoms showed similar rates of improvement during treatment, but while anxiety symptoms continued to improve during follow-up, depressive symptoms showed non-significant improvementafter treatment. Parent-rated symptoms also showed similar rates of improvement for anxiety and depression during the UP-A to those observed for adolescent self-report, but little improvement after treatment across either anxiety or depressive symptoms. To a certain degree, the results mirror those observed among other evidence-based treatments for youth with anxiety and depression, though results hold implications for future iterations of transdiagnostic treatments regarding optimization of outcomes for adolescents with depressive symptoms. Keywords: anxiety, depression, adolescence, transdiagnostic, treatment
Page 2 of 36
3 TRAJECTORIES OF ANXIETY AND DEPRESSION 1. Introduction Anxiety disorders are the most prevalent psychiatric disorders in childhood and adolescence, with prevalence estimates of 10-21% in the general population in the United States
ip t
(Costello, Mustillo, Erkanli, Keeler, & Angold, 2003). Considered through the lens of DSM-IV (American Psychiatric Association, 2000), unipolar depressive disorders (i.e., major depressive
cr
disorder [MDD], dysthymic disorder) as a whole are also common mental health conditions, and
us
become more prevalent during adolescence, as compared to earlier in development (Costello et al., 2002). Anxiety and depressive disorders commonly co-occur with one another in
an
adolescence. Between 16% and 62% of youth with an anxiety disorder also meet criteria for depression, with the highest comorbidity rates found among treatment-seeking adolescents
M
(Brady & Kendall, 1992; Ollendick, Shortt, & Sander, 2005). In addition, self-report measures of youth anxiety and depressive symptoms show moderate correlations with one another (e.g., r =
d
0.34), even after controlling for overlapping items on these instruments (Stark & Laurent, 2001).
Ac ce pt e
In addition to their high comorbidity with one another, youth anxiety and depression share a number of biological, environmental, and psychological risk factors (for a more thorough review, see Garber & Weersing, 2010). For instance, behavioral inhibition in early childhood is a risk factor for the later development of both anxiety and depression (Kagan, Reznick, & Snidman, 1987), and both anxiety and depressive disorders are associated with neuroendocrine (Dahl et al., 2000; Weems, Zakem, Costa, Cannon, & Watts, 2005) and neurotransmitter dysregulation (Flores et al., 2004; Fox et al., 2005). In addition, high negative affect (NA) has shown to be a latent factor underlying all of the anxiety and depressive disorders (Brown, Chorpita, & Barlow, 1998; Trosper, Whitton, Brown, & Pincus, 2012).
Page 3 of 36
4 TRAJECTORIES OF ANXIETY AND DEPRESSION Youth anxiety and depressive disorders also demonstrate similar responses to pharmacological and psychosocial interventions. For instance, both anxiety and depression are responsive to selective serotonin reuptake inhibitor (SSRI) medications (e.g., TADS, 2004;
ip t
Walkup et al., 2008). Prior work with cognitive-behavioral therapy (CBT) trials have found “spill-over” effects onto comorbid anxiety or depressive disorders, regardless of the principal
cr
disorder. For example, anxiety-focused CBT has demonstrated positive effects on comorbid
us
depressive symptoms (Kendall, Safford, Flannery-Schroeder, & Webb, 2004), and a metaanalysis of CBT for youth depression found effect sizes in anxiety symptom reduction (ES =
an
0.39) that were only slightly less than those for depressive symptom improvement (ES = 0.57; Weisz, McCarty, & Valeri, 2006).
M
Given their frequent comorbidity, shared vulnerability factors, and similar response to treatment, some (e.g., Barlow et al., 2004) have advocated a transdiagnostic or disorder-non-
d
specific treatment approach that targets higher-order psychological factors common to the
Ac ce pt e
emotional disorders. Such an approach is hypothesized to allow for greater clinical flexibility and use with patients presenting with multiple emotional disorders, as well as reduce clinician burden in learning multiple, disorder-specific treatment manuals (McHugh & Barlow, 2010). As such, recent clinical research has focused upon the development and evaluation of transdiagnostic treatments that may effectively target anxiety and depressive disorders within a single protocol. The Unified Protocol for the Treatment of Emotional Disorders in Adolescence (UP-A; Ehrenreich et al., 2008) is a developmental adaptation of the adult Unified Protocol (UP; Barlow et al., 2010), designed for adolescents ages 12-17 years old presenting with any primary anxiety disorder, unipolar depressive disorder, or their combination. The UP-A has theoretical roots in emotion regulation, cognitive science, and behavior modification, and distills common evidence-
Page 4 of 36
5 TRAJECTORIES OF ANXIETY AND DEPRESSION based techniques that cut across disorder-specific treatment manuals for youth anxiety and depression (e.g., psychoeducation, non-judgmental awareness, cognitive reappraisal, exposure, behavioral activation, etc.) within a singular protocol. The UP-A incorporates standard evidence-
ip t
based principles within the broader function, context, and regulation of a range of positive and negative emotions (e.g., sadness, anger, fear). Therefore, it is theorized to positively effect how
cr
adolescents attend to, modulate, and experience emotions that cut across specific disorders.
us
Similar to the UP, the UP-A targets five higher-order principles thought to be latent constructs underlying lower-order specific anxiety and depressive disorders: (1) increase present-focused
an
awareness of emotions, (2) enhance cognitive flexibility, (3) prevent emotional avoidance and maladaptive emotion-driven behaviors, (4) increase acceptance of uncomfortable emotion-
M
related physiological sensations, and (5) facilitate exposure to bodily and environmental triggers of emotional experiences (Barlow et al., 2010).
d
A prior open trial of the UP-A established initial efficacy, with subjects demonstrating
Ac ce pt e
significant pre-post reductions in clinician-rated diagnostic severity across anxiety and depressive disorder diagnoses (Trosper, Buzzella, Bennett, & Ehrenreich, 2009), and an immediate treatment (TX) condition of the UP-A has found to outperform an 8-week, treatmentdelayed waitlist (WL) condition in clinician-rated diagnostic severity for the principal disorder, in a recently completed randomized controlled trial (RCT; Ehrenreich-May, Queen, Rodriguez, & Rosenfield, 2012). Analyses from this RCT also found that the presence of a depressive disorder did not moderate treatment outcomes in the UP-A (Ehrenreich-May et al., 2012), whereas many previous CBT trials for youth anxiety have found poorer outcomes for patients with comorbid depression (e.g., Berman et al., 2000; Ginsburg et al., 2011; O’Neil & Kendall, 2012).
Page 5 of 36
6 TRAJECTORIES OF ANXIETY AND DEPRESSION To summarize, youth anxiety and depression are known to commonly co-occur with one another, share multiple vulnerability factors, and may be effectively treated with a unified treatment approach. However, despite their similarities, anxiety and depression have also shown
ip t
to be distinct constructs. For instance, factor analytic studies with school-based (Crowley & Emerson, 1996) and clinical samples (Stavrakaki, Vargo, Boodoosingh, & Roberts, 1987) have
cr
found stronger support for two-factor models of anxiety and depression compared to single
us
factor models. In addition, while both anxiety and depression are characterized by high negative affect, low positive affect has shown stronger associations with depressive symptoms than with
an
anxiety symptoms (for more comprehensive reviews, see Anderson & Hope, 2008; De Bolle & De Fruyt, 2010). Given these important differences, a next step in investigating transdiagnostic
M
treatment approaches is to examine the separate trajectories of symptom change for anxiety and
rates of change.
d
depression over the course of treatment, in order to assess if they show similar or differential
Ac ce pt e
The present study examined the separate trajectories of anxiety and depressive symptoms over the course of the UP-A, and up to six months following the end of treatment, for adolescent subjects completing the UP-A as part of the open trial or RCT investigation. We used piecewise latent growth curve modeling (LGCM) to model these trajectories over two separate time periods: during the course of treatment (“treatment slope”) and up to six months after treatment ended (“follow-up slope”). Piecewise LGCM is often recommended when examining change during treatment and follow-up, given likely non-linear change (Brown, 2004).Separate models were conducted for anxiety and depressive symptoms. Separate models were also conducted for self-rated and parent-rated symptoms, in order to examine informant differences in symptom change trajectories. Therefore, a total of four piecewise LGCMs were conducted.
Page 6 of 36
7 TRAJECTORIES OF ANXIETY AND DEPRESSION 2. Method 2. 1 Participants Participants were 59 adolescents (57.6% female) ages 12-17 years old (M = 15.42, SD =
ip t
1.71) who received at least eight sessions of the UP-A and completed at least one post-baseline assessment. Given the aim of the study was to examine separate trajectories of anxiety and
cr
depression symptom change for those completing the intervention, we decided to restrict
us
analyses to those receiving at least 8 sessions as this represented the minimum dosage possible to be considered a treatment completer. This subsample of 59 participants was culled from a total
an
sample of 67 participants who were enrolled in either the open trial or RCT investigation of the UP-A. Eight (11.94%) of the 67participants that did not complete at least eight sessions of the
M
intervention were part of the open trial (n = 2) or RCT (n = 6), respectively, and did not have any post-baseline assessment data. T-test and chi-square analyses revealed that those completing at
d
least eight sessions (n = 59) did not significantly differ from those who dropped out prior to eight
Ac ce pt e
sessions (n = 8) with regard to age, gender, ethnicity, severity of principal diagnosis, depressive disorder comorbidity status, or baseline levels of anxiety or depressive symptoms (child or parent report).
Participants were evenly divided between Hispanic/Latino (n = 26; 44.1%) and White, Non-Hispanic ethnicities (n = 26; 44.1%). The remaining subjects identified themselves as having Black/African-American (n = 2; 3.4%), Asian-American (n = 1; 1.7%), and “other” ethnicity (n = 4; 6.8%). The median reported annual family income was $100,000 (SD = $80,000). The majority of participants had parents who were married (n= 40; 67.8%). Remaining participants had parents who were separated/divorced (n = 13; 22.03%), widowed (n = 2; 3.4%), or never married (n = 4; 6.78%).
Page 7 of 36
8 TRAJECTORIES OF ANXIETY AND DEPRESSION Eligibility requirements for both the open trial and RCT included being between 12-17 years old (participants could be 18 if they were still in high school) and having a primary diagnosis of any DSM-IV anxiety and/or unipolar depressive disorder. Adolescents with other
ip t
psychiatric difficulties, including attention-deficit/hyperactivity disorder (ADHD), eating disorders, and non-treatment-interfering substance abuse were eligible for participation.
cr
However, the primary diagnosis had to be an anxiety or depressive disorder. Exclusion criteria
us
included bipolar disorder, treatment-interfering substance abuse, severe suicidal ideation or recent psychiatric hospitalization, treatment-interfering cognitive functioning (e.g., IQ below 80),
an
and a prior course of CBT for anxiety or depression. Participants and at least one parent were also required to read and comprehend English well enough to complete study measures. Those
M
concurrently receiving psychiatric medication were required to have been on a stable dosage of
prior to study enrollment.
d
an SSRI medication for three months and/or a stable dosage of a benzodiazepine for one month
Ac ce pt e
Adolescents completing the UP-A as part of the open trial participated at a universitybased clinic in a large urban area in the Northeastern United States (n = 15), while those in a subsequent RCT participated at a university-based clinic in a large urban area in the Southeastern United States (n = 44). There were no significant site differences with regard to gender, depressive disorder comorbidity, treatment length, or severity of anxiety and depressive symptoms. However, subjects in the RCT were slightly older (M = 15.69, SD = 1.70) than open trial participants (M = 14.63, SD = 1.55), t(57) = 2.13, p = .04, d = 0.64. In addition, as expected based upon demographic characteristics of the surrounding communities, RCT participants were more likely to be Hispanic/Latino (61.36%), whereas all of the open trial participants were White, Non-Hispanic, χ2 (3) = 25.53, p .20. Participants in the WL arm then began the UP-A and completed assessments at the same
M
intervals as those in the open trial and TX arm of the RCT.
2.2.1 Treatment structure and content.The UP-A follows a principle-based, flexible
d
treatment structure. The intervention can be delivered over a varying length of time (between 8-
Ac ce pt e
21 weekly sessions administered in an individual format). There are five required modules, which correspond to the five core principles underlining the UP/UP-A (Barlow et al., 2010). Therapists are allowed to devote more sessions to a required module dependent upon patient needs, but ideally all modules are completed during the course of treatment. Therefore, while treatment length may differ among patients, all subjects receive the same five core modules and relevant skills (e.g., psychoeducation, non-judgmental awareness/mindfulness, cognitive reappraisal, exposure, behavioral activation). See Table 2 for a display of the five required modules, suggested session length for each module, including treatment skills delivered within each module and the corresponding UP/UP-A core principles targeted. Importantly, one of the unique features of the UP models is the ability to target emotion regulation deficits across a
Page 12 of 36
13 TRAJECTORIES OF ANXIETY AND DEPRESSION broader range of positive and negative emotions, compared to traditional, disorder-specific treatment manuals. In addition to the five required modules, three optional modules are available to the
ip t
therapist as needed to address parenting skills, motivational enhancement, and clinical deterioration, and can be used at any point in treatment. At least one parent is required to be
cr
actively involved in treatment. While some variability is allowable, typically a parent is involved
us
in the last 10-15 minutes of every session to review content and the assigned homework. Individual sessions with the parent may be used as part of an optional module if the therapist
an
feels that certain parenting behaviors (e.g., overprotection, high criticism, parent-teen conflict, etc.) are counterproductive to treatment progress or as needed for exposure planning.
M
Motivational interviewing (Miller & Rollnick, 2002) techniques can be used in sessions as part of a second optional module for adolescents who appear reluctant to engage in treatment. Finally,
d
sessions within the third optional module can be used to develop a safety plan with the
Ac ce pt e
adolescent and his or her parent in the event the patient experiences impulses for self-harm. However, adolescents who show immediate risk for suicide or other pronounced clinical deterioration could also be discontinued from the UP-A studies and referred for more intensive services, if appropriate. 2.3 Measures
Anxiety Disorders Interview Schedule for the DSM-IV-Child/Parent Version(ADIS-IVC/P; Silverman & Albano, 1996). The ADIS-IV-C/P was completed at the adolescent’s baseline assessment to determine diagnostic eligibility. The ADIS-IV-C/P is a semi-structured diagnostic interview for children and adolescents ages 7-17 years that assesses all DSM-IV anxiety disorders. The ADIS-IV-C/P also assesses for unipolar depressive disorders and externalizing
Page 13 of 36
14 TRAJECTORIES OF ANXIETY AND DEPRESSION disorders (i.e., ADHD, oppositional-defiant disorder, conduct disorder), and screens for substance abuse, schizophrenia, pervasive developmental disorders, eating disorders, and somatoform disorders. In addition, participants were screened for bipolar disorder. Inter-rater
ip t
reliability for the primary diagnosis within this sample was very good (κ = .82, p .05), CFI > .95, RMSEA < .06, and SRMR < .08 (Kline, 2011). Significant tests for parameter estimates were conducted with the z-statistic using a two-tailed test. LGCM analyses were conducted using MPLUS, Fifth Version (Muthen & Muthen, 2005). 3. Results
Page 15 of 36
16 TRAJECTORIES OF ANXIETY AND DEPRESSION The data were first screened for normality and to determine if there were any statistical outliers. Skewness and kurtosis levels were within acceptable ranges (Kline, 2011). No significant outliers (z ≥ 3) were identified at any time points. All subjects had complete data for
ip t
Time 1 as well as at least one other post-baseline assessment. Fifty-five subjects (93.22%) had data at the 8-week assessment (Time 2) and 48 participants (81.36%) had data at the end of
cr
treatment (M = 15.58 sessions) assessment (Time 3). However, there was considerable missing
us
data at the two follow-up time points, with 29 subjects (49.15%) having available data three months after treatment (Time 4), and 24 participants (40.68%) having data six months after
an
treatment (Time 5), although 62.71% of subjects had available data for at least one follow-up time point. Attrition analyses revealed no significant differences between those with complete
M
data and those with missing data on any indicators of symptom severity at baseline or any demographic variables (e.g., age, gender, ethnicity). Furthermore, those with and without follow-
Ac ce pt e
3.1 Descriptive Statistics
d
up data did not show significant differences on any variables at the post-treatment assessment.
See Table 3 for observed means and standard deviations of RCADS/RCADS-P Total Anxiety Tscores and RCADS/RCADS-P MDD T-scores at each time point. In addition, the percentage of participants with T-scores 65 and T-scores 70 are presented. Mean T-scores on all four measures reduced to below 65 by Time 3, indicating that mean levels of both self-rated and parent-rated anxiety and depressive symptoms were within the normative range by the end of treatment. Analysis of percentages of participants with T-scores 65 and T-scores 70 also revealed reductions in those with borderline elevated and clinically elevated anxiety and depressive symptoms from baseline to the end of treatment, which were largely maintained during follow-up.
Page 16 of 36
17 TRAJECTORIES OF ANXIETY AND DEPRESSION 3.2 LGCMs for Adolescent Self-Rated Symptoms 3.2.1 Anxiety symptoms.Self-rated anxiety and depressive symptom trajectories (observed and estimated) are displayed in Figure 2. The piecewise LGCM for self-rated anxiety symptoms was
ip t
modeled by specifying three latent factors: an intercept (baseline levels of anxiety symptoms), a treatment slope factor (change during the UP-A), and a follow-up slope factor (change during
cr
follow-up). All loadings from the intercept to observed variables were fixed at 1. Loadings for
us
the treatment slope factor were fixed at 0, -1, -2, -2, and -2. Loadings for the follow-up slope factor were fixed at 0, 0, 0, -1.5, and -3. These loadings can be interpreted as a period of eight
an
weeks per one unit change. The final model imposing equality constraints for the residual variances of the first three indicators and last two indicators demonstrated acceptable model fit
M
by most indices, χ2 (9) = 13.63, p = 0.14, CFI = 0.97, RMSEA = 0.09, SRMR = 0.19. There was a significant mean intercept (Mi = 53.68, SE = 1.65, p
