Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Jul 1.
Published in final edited form as: Behav Ther. 2024 Feb 12;55(4):898–911. doi: 10.1016/j.beth.2024.01.004

What CBT modules work best for whom? Identifying subgroups of depressed youths by their differential response to specific modules

Jennifer Frederick 1, Mei Yi Ng 1, Matthew J Valente 2, Katherine Venturo-Conerly 3, John R Weisz 3
PMCID: PMC11211639  NIHMSID: NIHMS1986842  PMID: 38937058

Abstract

Prior research suggests that the effects of specific CBT modules on symptom outcomes can be estimated. We conducted a study utilizing idiographic and nomothetic methods to clarify which CBT modules are most effective for youth depression, and for whom they are most effective. Thirty-five youths received modular CBT for depression. Interrupted time series models estimated whether the introduction of each module was associated with changes in internalizing symptoms, whereby significant symptom reduction would suggest a therapeutic response to the module. Regression models were used to explore whether participant characteristics predicted subgroups of youths based on their estimated response to certain types (e.g., cognitive) of modules, and whether group membership was associated with post-treatment outcomes. Thirty youths (86%) had at least one module associated with a significant change in internalizing symptoms from pre-module delivery to post-module delivery. The specific modules associated with these changes varied across youths. Behavioral activation was most frequently associated with symptom decreases (34% of youths). No participant characteristics predicted estimated response to module type, and group membership was not significantly associated with post-treatment outcomes. Youths display highly heterogeneous responses to treatment modules, indicating multiple pathways to symptom improvement for depressed youths.

Keywords: Major depression, child psychopathology, cognitive behavioral therapy, precision medicine, ideographic analyses

Introduction

Despite evidence that cognitive behavioral therapy (CBT) is a well-established treatment for depression in youths (Weersing et al., 2017), many youths do not demonstrate major improvements in depression symptoms after therapy (Cuijpers et al., 2020; Eckshtain et al., 2020), and the effects of psychotherapies for youth depression have evidently not improved significantly across the decades (Weisz et al., 2019). Given the high rates of depression among children and adolescents, with an estimated prevalence rate of nearly 12%, (Merikangas et al., 2010), there is a pressing need to improve the efficacy of treatments available to youths with depression.

Taking an idiographic approach and examining intensive data from individual participants may help to further research identifying effective ingredients of therapy and to inform efforts to personalize, and ultimately improve the effectiveness of, psychotherapy for youth (Ng & Weisz, 2016). However, in order to personalize treatment for youths with depression, it is necessary to know not only which treatments are effective, for whom they are the most effective, and under which circumstances (Paul, 1967)—it is also important to identify which modules in a multi-component treatment such as CBT work especially well for which youth. Therefore, it is necessary to examine both specific treatment modules, their impact on treatment outcomes, and youth characteristics associated with symptom trajectories.

There are many characteristics that may influence a given youth’s trajectory of depressive symptoms and response to treatment. Two characteristics that have strong associations with trajectories of depressive symptoms in youths are age and sex. Specifically, girls approaching adolescence are more likely to experience an increase in depressive symptoms compared to their male peers (Lewis et al., 2020). Prevalence rates for depression are similar between girls and boys in childhood but diverge in adolescence such that girls will have depression at a rate of 2–3 times that of boys of the same age (Kessler et al., 1994; Merikangas et al., 2010). Hormonal changes and the timing of puberty may contribute to age and sex differences in prevalence rates of depression (Hyde & Mezulis, 2020). Moreover, youths’ characteristics might impact treatment response. For example, higher symptom severity has been associated with poor outcomes among depressed adolescents (Courtney et al., 2022). Gender has not been consistently identified as a moderator of response to psychotherapy for depression (Courtney et al., 2022; Meyer & Curry, 2020). A recent meta-analysis examining age differences found effect sizes notably, though not significantly, larger for adolescents (0.55) than children (0.35; Cuijpers et al., 2020). Other factors such as the presence of comorbid anxiety and maternal depressive symptoms have been associated with worse treatment outcomes for depressed youths (Brent et al., 1998).

There is evidence that individuals with depression may have optimal interventions, depending in part on their personal and clinical characteristics (DeRubeis et al., 2014; Young et al., 2021). For example, youths who received an intervention for depression that aligned with type of risk for depression (e.g., CBT for youths with a negative cognitive style and interpersonal psychotherapy for youths with low peer support) had greater improvement compared to youths who did not receive the intervention that was matched to their personal vulnerabilities. (Young et al., 2021). Additionally, specific CBT components have been associated with improved outcomes. Among youths in one study with treatment-resistant depression who received medication and CBT (TORDIA), those who received problem-solving and social skills modules were more likely to respond to treatment than those who did not (Kennard et al., 2009).

Given the well-documented heterogeneity in treatment response of depressed youths, idiographic analyses such as time series models are a promising way to inform clinical practice (Piccirillo et al., 2019). Interrupted time series models can be used to examine the impact of specific CBT modules on symptoms for individual youths. Although these models cannot explain causal effects and can only produce findings specific to an individual participant, they can examine change in symptom levels that can be associated with the introduction of a therapy module. These models compare the average level of symptoms pre-module delivery (i.e., all weeks before a module was introduced) to the average level of symptoms post-module delivery (i.e., all other weeks until the end of treatment). Our prior work has used these models to demonstrate that the introduction of some modules of CBT (e.g., engagement, relaxation, cognitive reframing, behavioral activation, and psychoeducation) were associated with changes in internalizing symptoms in a sample of three depressed youths (Frederick et al., 2022). Importantly, the modules that were associated with changes in symptoms varied across youths. Moreover, some modules that were associated with a decrease in symptoms for some youths were associated with increases in symptoms, or not associated with symptom change, for others.

Objectives

To leverage the advantages of idiographic methods in capturing individual symptom change while addressing its limitations in generalizing beyond the participant sample, we utilized an innovative study design combining idiographic and nomothetic (i.e., group-level) approaches. Using a sample of 35 youths receiving modular therapy for depression, we first fitted idiographic models to examine the association of specific CBT modules with changes in internalizing symptom outcomes for each individual youth. Then we conducted group-level analyses to identify baseline characteristics that predicted individual response to specific modules, and to examine whether response to specific modules was associated with overall improvement.

We anticipated that the idiographic model findings in the present study would replicate those from the previous study we conducted with three youths drawn from a different treatment trial (Frederick et al., 2022). That is, we hypothesized that most youths would have at least one module that, when introduced in an interrupted time series model, was associated with changes in trends of youth internalizing symptoms from pre-module introduction to post-module introduction. We also expected that the specific modules associated with symptom changes would vary across youths. Although the interrupted time series models can examine whether the introduction of a therapy module was associated with a change in symptoms, within our sample of youth who received modular therapy, modules were not delivered in isolation, nor was module delivery randomized across participants. This limits the generalizability of our findings, but our results may serve as preliminary evidence of specific modules to be investigated further.

Our expanded participant sample allowed us to group youths by their response to types of specific modules (e.g., whether they showed a decrease in internalizing symptoms after introduction of cognitive or behavioral modules), and to examine which baseline youth characteristics predicted with group membership using logistic regression. Identifying youth characteristics readily available to clinicians, such as age, sex, and baseline symptom severity, that are associated with individual treatment response to specific modules, could inform future work to develop clinical profiles to match youths with the treatment modules most likely to benefit them. Finally, we examined whether individual treatment response to specific modules had an impact on overall outcomes, using linear regression analyses to test group membership as predictors of post-treatment symptom severity. Results could indicate whether youths who responded to specific modules had differences in their overall treatment outcomes. As there has been limited research examining baseline characteristic associated with treatment response to specific modules and response to specific modules with overall treatment outcomes, there was not sufficient basis for a hypothesis and this aim was exploratory.

Methods

Participants

Participants were selected from two randomized effectiveness trials where treatment-seeking youths ages 6–15 who presented with symptoms of anxiety, depression, posttraumatic stress, conduct problems, or some combination thereof, received modular CBT (Weisz et al., 2018, 2020). The same treatment, the Modular Approach to Therapy for Children with Anxiety, Depression, Trauma, or Conduct Problems (MATCH; Chorpita & Weisz, 2009) was used in all trial conditions, which took place in community mental health clinics. The conditions across trials differed in the type of supervisor (trained clinic staff or external clinicians; Weisz et al., 2018) or the level of implementation support (with or without expert consultation; Weisz et al., 2020). Informed consent and assent were obtained from caregivers, youths, and clinicians prior to study enrollment and all study procedures were approved by institutional review boards. For the current study, we selected youths who met the following criteria: (a) had elevated self-reported or caregiver-reported internalizing symptoms at baseline as defined as a T-score ≥ 60 on either the Child Behavior Checklist (CBCL) or Youth Self Report (YSR; Achenbach & Rescorla, 2001), (b) received at least two MATCH depression modules, (c) reported symptom measures for a time span of at least 30 weeks in order to ensure there were sufficient observations for model parameters to be estimated accurately (SAS Institute Inc., 2015), and (d) had <40% missing data to ensure accurate model estimation (Velicer & Colby, 2005). Out of the 320 youths who participated in the randomized trials, 289 had elevated baseline CBCL or YSR Internalizing scores, and 42 received modules from the depression protocol and had symptom measures for a span of at least 30 weeks. Ultimately, 35 met our inclusion criteria and were included in the present study. Participant demographics, treatment duration, baseline and post-treatment symptoms for the present study sample are reported in Table 1. T-tests revealed some differences between participants drawn from the Weisz et al., 2018 and 2020 trials. Full results can be found in Table 1. Youths in the 2018 trial had significantly higher baseline CBCL Externalizing scores and had higher post-treatment YSR Internalizing and Externalizing T-scores.

Table 1.

Participant demographics and treatment characteristics (N=35)

Characteristic Mean (SD) or N (%) 2018 Trial 2020 Trial Test Statistic p-value
Sex (Male) 17 (48.57%) 5 (45.45%) 12 (50.00%) χ2 <0.01 0.99
Race/Ethnicity χ2 = 10.7 0.03
 White 19 (54.28%) 10 (90.91%) 9 (37.5%)
 Black 6 (17.15%) 0 (0%) 6 (25%)
 Latino 6 (17.15%) 0 (0%) 6 (25%)
 Asian 2 (5.71%) 0 (0%) 2 (8.33%)
 Multiracial 2 (5.71%) 1 (9.09%) 1 (4.17%)
Elevated youth-reported symptoms at baseline 21 (60.00%) 8 (72.72%) 13 (54.17%) χ2 = 0.45 0.50
Age 11.2 (2.58) 10.36 (2.16) 11.58 (2.70) t = −2.71 0.17
Number of weeks in therapy 38.18 (16.92) 43.60 (21.48) 35.70 (14.22) t = −1.12 0.28
Number of sessions attended 19.31 (9.00) 22.36 (9.87) 17.92 (8.43) t = −1.29 0.21
Baseline YSR Internalizing 61.47 (12.60) 62.80 (7.10) 60.58 (14.36) t = −0.60 0.55
Post-treatment YSR Internalizing 52.90 (16.69) 71.60 (5.42) 43.55 (11.62) t = −9.01 <0.01
Baseline YSR Externalizing 53.82 (10.60) 54.00 (8.84) 53.75 (11.43) t = −0.07 0.94
Post-treatment YSR Externalizing 51.00 (15.43) 68.2 (7.64) 42.40 (10.01) t = −7.80 <0.01
Baseline CBCL Internalizing 70.94 (5.08) 71.27 (5.26) 70.79 (5.11) t = −0.25 0.80
Post-treatment CBCL Internalizing 60.81 (9.82) 61.63 (10.29) 60.35 (9.80) t = −0.34 0.73
Baseline CBCL Externalizing 63.86 (8.34) 68.45 (7.30) 61.75 (8.05) t = −2.44 0.02
Post-treatment CBCL Externalizing 54.87 (10.77) 56.73 (13.73) 53.85 (8.99) t = −0.63 0.54
Open in a new tab

Note. CBCL = Child Behavior Checklist, YSR = Youth-Self Report. CBCL and YSR scores were baseline broadband T-scores

Modular Treatment

Participants received the Modular Approach to Therapy for Children with Anxiety, Depression, Trauma, or Conduct Problems (MATCH; Chorpita & Weisz, 2009). MATCH consists of 33 modules comprising empirically supported treatment procedures. Therapists identify the primary problem area (anxiety, depression etc.) and select appropriate modules with guidance from flowcharts. Therapists could skip, repeat, or use modules targeting any problem area. Thus, youths with the same presenting problem may not have received the same modules, same number of modules, or modules in the same order. Modules for engagement and wrap up were not specific to any protocol. Therapists in all trials used a web-based measurement feedback system to view weekly symptom ratings and to record which modules were used during a given session. Data from clinical or supervision notes are not available for these trials. Therapists in the 2018 trial were supervised by either external experts or community clinicians trained as part of the trial. Clinicians did not differ significantly on treatment fidelity or youth outcomes. Therapists in the 2020 trial received either low-cost implementation supports (such as video tutorials and downloadable materials) or the same supports plus expert consultation. Therapists who received expert consultation had stronger treatment fidelity. Trajectories of patient symptom change on all outcomes were similar across groups. For additional information about each trail see the supplemental materials and the original articles cited above.

Measures

Caregiver and Youth Reported Symptoms

Youths and caregivers across both trials completed the YSR and CBCL, parallel youth- and caregiver-rated inventories of youth psychopathology, at baseline and post-treatment. The broadband Internalizing and Externalizing symptoms scales on the YSR and CBCL have well-documented reliability and validity (Achenbach & Rescorla, 2001). Youths were included in our sample if they had either youth-reported or caregiver-reported elevated (T-score ≥ 60) internalizing symptoms at baseline. Youth and caregiver reports of depressive symptoms have low to moderate levels of agreement, yet both offer valid reports of symptoms (Cohen et al., 2019). Therefore, when fitting each youth’s time series model, we selected weekly symptom reports by the respondent who reported elevated symptoms at baseline. If a participant had both elevated youth-reported and caregiver-reported symptoms, youth-reported symptoms were used. All youths had elevated baseline caregiver-reported internalizing symptoms, and 21 youths had elevated baseline youth-reported internalizing symptoms. One youth had elevated caregiver and youth-reported symptoms at baseline, but had insufficient youth-reported data for analyses, so caregiver-report was used for their time series model.

Weekly Symptom Reports

To measure caregiver and youth-reported symptoms on a weekly basis, participants from the Weisz et al., 2018 trial used the Brief Problem Checklist (BPC; Chorpita et al., 2010) and participants from the Weisz et al., 2020 trial used the Brief Problem Monitor (BPM; Achenbach et al., 2011). Both measures have a subscale measuring internalizing symptoms that consist of the same 6 items drawn from the CBCL and YSR (Chorpita et al., 2010; Piper et al., 2014). Both the BPM and BPC have strong reliability and validity and have been shown to be highly correlated with CBCL and YSR scores (Chorpita et al., 2010; Piper et al., 2014).

Data Analysis

Binning of Weekly Symptom Reports

Treatment providers generally aimed for weekly sessions, however there was considerable variability in the length of time between sessions among participants in both trials (M = 14.59 days between sessions, SD = 11.30 days, range = 0–106 days), reflecting the conditions of community-based youth mental health services. There was also variability in the length of time between weekly symptom ratings (M = 7.75 days between sessions, SD = 3.50 days, range = 1–52 days). Because equally spaced observations are recommended for time series analyses (Nelson, 1998), we sought to equalize the units of time between observations. Dates of treatment sessions and reports of symptoms were organized into week-long bins. We maximized the number of weekly bins containing a session that was paired with a symptom report in the subsequent bin to facilitate the estimation of treatment component effects on symptoms. Analyses were conducted in R 4.1.3 and SAS 9.4.

Missing data and model selection

Missing weekly symptom report data were imputed using R package imputeTS (Moritz & Bartz-Beielstein, 2017). Methods included in this package account for time dependencies, which are preferred for univariate time series models. Model fit was compared for each participant across the imputation methods available in the imputeTS package, and an option that uses Kalman Smoothing fitted by maximum likelihood yielded the best fit across participants.

Separate interrupted time series models were used to evaluate the association of the introduction of specific modules on internalizing symptoms for each participant. Autocorrelation function (ACF) plots were examined to aid the determination of which time series model would be best suited for our data and a χ2 statistic was estimated to test the null hypothesis that none of the autocorrelations in the series were significantly different from zero. ACF plots and χ2 tests revealed that for all but 6 youth, autoregressions decayed gradually over time, suggesting that autoregressive (AR) models would be appropriate. For the remaining 6 youth, we took a conservative approach of accounting for lagged dependencies, as these χ2 tests may have been underpowered due to our relatively low number of observations (Box et al., 2013) We used an AR model with a lag of 1, where an autocorrelation coefficient is used to represent correlation among symptom scores for a given lag in time (e.g., Yt and Yt−1). As our data were organized into weekly bins, a lag of 1 reflects the correlation between two scores at adjacent time points (a period of 7 days). Models were estimated using PROC ARIMA in SAS.

Each model contained a dummy-coded variable that represented the introduction of a module during a given participant’s course of treatment. In the event that multiple modules were introduced in the same week, one dummy-coded variable was used to represent all modules. If that dummy-coded variable resulted in an association with significant changes in symptoms, all modules associated with that week were included in our tally. Model parameters estimated by maximum likelihood were examined to determine the association of the introduction of a given module on symptoms. Parameter estimates quantified the change in level of average internalizing symptoms from pre-module delivery to post-module delivery that could be attributed to the introduction of that module. After estimating each model, a χ2 statistic was calculated to test whether the residuals from each model were uncorrelated to further evaluate model fit.

Group Level Analyses

Results from idiographic models were used to group youths based on their responses to specific module types. We constructed two binary (i.e., dummy-coded) variables representing (1) the Cognitive Decrease Group – youths for whom cognitive modules were v. were not associated with decreases in symptoms and (2) the Behavioral Decrease Group – youths for whom behavioral modules were v. were not associated with decreases. These groups were mutually exclusive. As only one youth each had a behavioral module and a cognitive module associated with increases in symptoms, these two youths were excluded from these analyses, as our sample size was not large enough to estimate accurate predictions for group memberships as small as a single individual. Participant #2, a 10-year-old male in the 2018 trial who had elevated YSR Internalizing and CBCL Internalizing and Externalizing scores at baseline, was excluded from the analyses examining predictors of the Cognitive Decrease Group. Participant #15, a 13-year-old female in the 2018 trial who had elevated YSR and CBCL Internalizing and Externalizing scores at baseline, was excluded from the analyses examining predictors of the Behavioral Decrease Group. We examined whether participant characteristics (i.e., age, sex, baseline Internalizing and Externalizing symptoms on the CBCL and YSR, respondent of weekly symptom reports) were associated with the two group memberships described above using separate logistic regression modes. We also included a variable to represent the trial from which each participant’s data was drawn in order to account for the differences between trials.

We used linear regression models to examine whether group membership predicted post-treatment Internalizing scores in the CBCL and YSR or number of sessions youths attended. We ran three separate models, (for CBCL symptoms, for YSR symptoms, and for number of sessions attended), examining the two groups as predictors. Participant #2 and #15 were excluded from these analyses. For models where CBCL or YSR post-treatment Internalizing symptoms were the outcome of interest, we controlled for corresponding report of baseline internalizing symptoms.

Results

Frequency of modules associated with changes in symptoms across youths

A total of 30 out of 35 (85.71%) youths had specific modules that were associated with significant change in the level of average internalizing symptoms pre-module delivery to post-module delivery. Fourteen youths (40.00%) had modules that were associated with decreases in symptoms, 3 youths (8.57%) had modules that were associated with increases in symptoms, and 13 (37.14%) had modules associated with either increases or decreases in symptoms. For those who had modules associated with any change in symptoms, the number of modules associated with a change ranged from 1 to 10 modules. T-tests revealed there were significant differences in youth- and caregiver-reported internalizing symptoms from baseline (youth M = 61.23, SD = 12.60, caregiver M = 70.94, SD = 5.08) to post-treatment (youth M = 60.81, SD = 16.69, t = 2.23, p = 0.03; caregiver M = 52.9, SD = 12.60, t = 5.17, p > 0.01) within our sample, indicating that treatment was effective for youths.

The specific modules associated with these changes varied across youths (see Table 2). A total of 23 modules (76.67%) out of the 30 different modules received by youths in this sample were associated with any change in symptoms. Out of those 23 modules, 9 modules (39.13%) were associated only with decreases in symptoms. The remaining 14 modules were associated with increases or decreases across youths. Behavioral activation stands out as the module most frequently associated with decreases in symptoms—one-third (10/29) of the youths who received the module, had reductions in their average level of internalizing symptoms after behavioral activation was introduced. This module was associated with decreases in symptoms for one youth. Engagement was the module most frequently associated with increases in symptoms (6/35 youths) but was also associated with decreases in symptoms for 6 youths and had no significant associations for 23 youths. For 28/30 modules associated with a change in symptoms, there were more youths who received the module and did not have any significant association between the introduction of the module and changes in symptoms than there were youths for whom the module was associated with a change in symptoms. These results are shown in Figure 1. Out of all the weeks where modules were introduced and were associated with a change in symptoms across participants, 36.20% of the weeks where modules were introduced and associated with changes in symptoms occurred within the first quarter of weeks of treatment, 18.97% occurred in the second quarter, 22.41% occurred in the third, and 22.41% occurred in the final quarter. A Fisher’s exact test of a contingency table of youth who had or did not have modules associated with change each quarter revealed there was no significant difference between quarters. While there was no evidence for significant differences in the proportion of modules associated with change across the four quarters, this does not imply there was evidence that the proportion of modules associated with change across the four quarters was equal. Looking at specific modules found to be associated with changes in symptoms, there was a wide range of weeks where those modules were associated with changes. For example, across our sample behavioral activation was associated with changes in symptoms across a range of 6 to 28 weeks and problem solving was associated with changes in symptoms across a range of 6 to 46 weeks.

Table 2.

Summary of logistic regression analyses examining participant characteristics as predictors of group membership based on which types of modules were associated with decreases in symptoms for youths.

Outcome Cognitive Decrease Group Behavioral Decrease Group
Number of youths in group N = 6 N = 10
b SE p-value b SE p-value
Intercept −0.8 9.94 0.94 −6.41 15.84 0.69
Sex 2.22 1.37 0.11 −2.49 2.25 0.27
Age 0.35 0.28 0.21 −0.73 0.55 0.19
CBCL Internalizing 0.01 0.16 0.93 −0.21 0.21 0.31
YSR Internalizing 0.0004 0.09 0.99 −0.17 0.19 0.36
CBCL Externalizing −0.16 0.14 0.24 0.37 0.26 0.15
YSR Externalizing 0.01 0.14 0.92 0.22 0.19 0.25
Trial 1.16 1.48 0.43 5.31 3.03 0.08
Respondent −0.72 1.76 0.68 2.36 3.37 0.48
Open in a new tab

Note. CBCL = Child Behavior Checklist, YSR = Youth Self Report. CBCL and YSR scores were baseline broadband T-scores

Figure 1.

Figure 1.

Open in a new tab

Number of youths (N=35) for whom specific modules were associated with increases, associated with decreases, or not associated with changes in internalizing symptoms.

As our analyses were idiographic, we detail the sequence of modules and results for one exemplar participant and provide contextual information below to illustrate the interpretation of results. Figure 2 illustrates the introduction of modules over the course of treatment associated with a significant change in level of average internalizing symptoms for this participant. Model parameters for all participants can be found in the supplemental materials.

Figure 2.

Figure 2.

Open in a new tab

Introduction of specific MATCH modules in relation to internalizing symptoms over time

Note. Modules associated with a significant change in average internalizing symptom severity

(change in the level of average youth-reported internalizing symptoms from pre-module delivery to post-module delivery) are indicated by vertical lines, with blue dashed lines indicating a decrease in symptoms.

Course of treatment and specific module associations for one participant

Participant #19 was an 11-year-old boy who was part of the 2020 trial. He had baseline self-reported Internalizing and Externalizing T-scores of 67 and 56 on the YSR. He had caregiver-reported Internalizing and Externalizing T-scores of 70 and 65 on the CBCL. Youth-reported symptoms were used for his interrupted time series model. His self-identified top problems were not being able to handle his anger and feeling like he did not belong with peers at school. Caregiver-reported top problems were feeling like he does not fit in, bottling up emotions, and threatening to self-harm when upset. He was introduced to 11 different modules over 29 weeks following the depression protocol. He completed two weeks of symptom reports before starting therapy where engagement and child-focused psychoeducation were introduced in the first two sessions. Parent-focused psychoeducation and problem solving were introduced during weeks 15 and 17 of assessment. Behavioral activation was introduced during week 22 of assessment and was associated with a decrease in his level of average internalizing symptoms from pre- to post-module delivery, B = −1.05, SE = 0.53, p = 0.05. Relaxation was introduced during the next session. During week 30, the therapist covered both relaxation and social skills, and these modules were associated with a decrease in symptoms, B = −1.85, SE = 0.40, p = <0.01. His last two sessions continued to focus on social skills, introduced cognitive reframing, and wrapped up treatment. He had post-treatment self-reported Internalizing and Externalizing T-scores of 51 and 42 on the YSR and caregiver-reported T-scores of 63 and 60 on the CBCL.

Predictors of group membership based on associations of modules with symptom changes

We examined participant sex, age, baseline CBCL and YSR Internalizing and Externalizing T-scores, trial, and respondent as predictors of group membership based on which modules were associated with group membership based on decreases or no association with changes in symptoms for youths (i.e., Cognitive Decrease and Behavioral Decrease Groups). The results of these analyses are detailed in Tables 2. No participant characteristics predicted Behavioral or Cognitive group membership. In addition, group membership did not predict number of sessions attended, or youth- and caregiver-reported post-treatment internalizing symptoms. Results can be found in Table 3.

Table 3.

Summary of linear regression analyses examining group membership as predictors of youth- or caregiver-reported post-treatment internalizing symptoms or the number of sessions attended.

Outcome Post-Treatment YSR Internalizing T-score Post-Treatment CBCL Internalizing T-score Number of Sessions Attended
b SE p-value b SE p-value b SE p-value
Intercept 55.29 19.12 0.01 −8.38 22.72 0.72 55.65 24.58 0.03
Cognitive Decrease Group −8.9 8.83 0.32 3.83 4.67 0.42 2.52 4.02 0.54
Behavioral Decrease Group −14.86 7.31 0.05 4.4 3.54 0.23 −4.52 3.48 0.21
Baseline YSR Internalizing T-score 0.06 0.29 0.83 −0.48 0.32 0.15
Baseline CBCL Internalizing T-score 0.96 0.32 0.01 −0.01 0.14 0.93
Open in a new tab

Note. CBCL = Child Behavior Checklist, YSR = Youth Self Report. Baseline YSR Internalizing T-scores were included as a covariate in the model examining post-treatment YSR Internalizing T-scores as an outcome. Baseline CBCL scores were included in the caregiver report model. Both youth- and caregiver-reported baseline internalizing symptoms were included as covariates in the model predicting the number of sessions attended.

Discussion

Our aim was to investigate whether the introduction of specific CBT modules was associated with changes in youth internalizing symptoms. In line with our hypothesis, the majority of our sample (85.71%) had at least one module that was significantly associated with changes in internalizing symptoms after module delivery. Overall, behavioral activation, problem solving, and engagement were the modules that had the greatest number of significant associations with changes in symptoms. However, nearly all of the modules youths received that were associated with changes in symptoms for some youths were not associated with changes for other youths. Within the context of these analyses, a module lacking an association with significant change does not mean that the module did not contribute to the participant’s overall trajectory during their course of treatment, the lack of association indicates that the introduction of the module was not associated with a change in symptoms above and beyond other modules.

Our preliminary results suggest that only a few CBT modules may be particularly beneficial or iatrogenic for each youth. Across youths, the number of modules that were associated with increases or decreases in symptoms varied, as did the content of the modules. Our findings replicate our previous work (Frederick et al., 2022), which demonstrated that specific modules have measurable effects for youths, and that the specific modules associated with increases and decreases in symptoms varies across youths. Interestingly, modules in this study that were associated with changes in symptoms occurred across a wide range of time (weeks 1–65). However, without multiple baseline observations to establish a trend, we cannot rule out regression to the mean as a possible explanation for symptom change over time.

Specific modules associated with increases or decreases in symptoms

Behavioral activation was associated with decreases in symptoms for a third of the youths who received the module. Behavioral activation is one of the core components of CBT for depression—it facilitates engagement in rewarding activities youths avoid in order to improve their mood. There is a growing body of evidence that the extent to which adolescents and adults receiving psychotherapy for depression engage in behavioral activation is associated with greater improvements in symptoms (Webb et al., 2016, 2019). One study of adolescent girls comparing the use of cognitive and behavioral skill use found that the extent to which adolescents engage with behavioral, but not cognitive skills, predicted symptom improvement from session to session (Webb et al., 2019). Behavioral activation may be particularly useful for youths because developmentally appropriate activities can be scheduled for younger children, and parental involvement can be easily incorporated (Dimidjian & Mccauley, 2016; Martin & Oliver, 2019).

Taken together, the two cognitive restructuring modules for depression were associated with decreases in symptoms for 6 youths and associated with an increase in symptoms for 1 youth. It is notable amongst our findings that there was a relatively low number of youths for whom cognitive restructuring modules were associated with decreases in symptoms, especially as cognitive restructuring is a core component of CBT. Cognitive skills require higher-order cognitive abilities, such as metacognition and abstract thinking (Webb et al., 2019) that may exceed the developmental level of younger adolescents or children (Coffman et al., 2007). Depressed adults with high levels of rumination have been shown to benefit from cognitive skills (Jones et al., 2008), and adults with more severe depression benefit more from behavioral, not cognitive skills (Coffman et al., 2007; Webb et al., 2016). Additional work with younger samples is needed to identify subgroups for whom cognitive skills are most effective.

We found that problem solving was associated with decreases in symptoms for 8 youths and associated with increases for 5. Our finding partially aligns with emerging support for problem solving as an effective stand-alone intervention for depression in youths (Michelson et al., 2022), as well as with evidence indicating that problem solving may be beneficial for a subset of depressed youth. Improvement in problem solving was associated with remission of depressive symptoms in the adolescents whose mothers had lower levels of depressive symptoms (Dietz, 2014). A possible explanation for these seemingly inconsistent findings, is that some youths may encounter greater barriers to using problem solving skills – particularly those who have more negative cognitive distortions. Cognitive distortions (e.g., negative thoughts about oneself or the world) have been found to be associated with negative problem orientation (e.g., “nothing will work”) in youths (Wilson et al., 2014), and trying but failing to effectively problem solve could intensify these negative thoughts. Perhaps problem solving can be more effective for these youths if they first learn reframing strategies to target cognitive disruptions before undertaking problem solving exercises, selected by the therapist to increase the probability of mastery experiences (Michelson et al., 2022).

Engagement was evenly split, associated with decreases in symptoms for 6 youths and associated with increases for 6 youths. Engagement is critical for building therapeutic alliance, which explains a proportion of the variance in clinical outcomes for youths (McLeod, 2011). Although alliance was not examined as a part of this study, challenges to establish alliance during early treatment sessions may have contributed to increases in symptoms, whereas successful engagement and an agreement between a therapist’s explanation of treatment and youth’s conceptualization of their symptoms may have contributed to a reduction in symptoms (Cuijpers et al., 2019; Hawley & Weisz, 2003). Given that our sample was made up of youths who all had elevated caregiver-reported internalizing symptoms at baseline, but not all youths had self-reported elevated symptoms, some youths may not have been motivated to seek treatment and may have found it challenging to discuss their symptoms with a therapist. One of the benefits of modular therapy is that a therapist can adjust the content and course of treatment based on how their patient is responding to treatment. Without examining patient change processes and change mechanisms, we cannot determine how much a participant’s responsiveness to treatment may have influenced the sequence and order of modules that they received (Doss, 2004). This may be particularly salient with engagement, as therapists may have devoted more time to building rapport with patients who appeared to be disengaged.

Lack of association between modules and changes in symptoms

Although the majority of youths in our sample had at least one module that was associated with increases or decreases in symptoms, it is notable that for 28/30 modules associated with a change in symptoms, there were more youths who received the module and did not have any significant association than there were youths for whom the module was associated with changes in symptoms. In these analyses, a lack of association indicates that the introduction of the module was not associated with a change in symptoms above and beyond other modules. Interestingly, when we examined the distribution of modules that were associated with changes in symptoms for a subsample of youths who showed substantial (≥25%) pre-post improvement in internalizing symptoms, we found a similar pattern of results compared to our entire sample (see Supplemental Tables 1 and 2). A stable trajectory of symptom improvement may be the norm for some youth. For example, many youths do not experience any sudden gains (Dour et al., 2013) or sudden regressions (Durland et al., 2018) during therapy. Therefore, it may be possible that youths may benefit from the skills taught in therapy gradually and a sudden improvement in symptoms may not be necessary for improvement in symptoms overall.

Predictors of Group Membership and Associations of Group Membership with Symptoms

Out of the participant characteristics (age, sex, baseline YSR and CBCL Internalizing and Externalizing scores, respondent, and trial) we examined, none were found to be predictors of group membership based on the types of modules that were associated with changes in symptoms. Additionally, group membership did not predict youth- or caregiver-reported post-treatment internalizing symptoms or the number of sessions that youth attended. If replicated, the results of these group-level analyses might indicate equifinality: learning different skills from different modules may all lead to the reduction of depressive symptoms. Some individuals with depression may not have an optimal treatment and may nonetheless benefit similarly from multiple options. For example, efforts to identify optimal treatments for adults with depression found that 37% of patients were not predicted to have a clinically meaningful advantage in one therapy over another (Huibers et al., 2015). It is important to identify not only youths who may benefit from a specific treatment, but also youths who may benefit equally from multiple options for treatment, as they may be able to access effective treatment more easily.

Limitations

Our study has several limitations. Modules were not delivered to participants in a random order, and participant response to treatment may have influenced therapists’ module selection therefore we cannot interpret the observed associations as causal effects and the idiographic modules are specific to individual participants. Time series accord more power to modules delivered earlier in treatment. However, sequence effects are unlikely to account for all our findings given the wide range of weeks (1–65) in which modules with measurable associations were introduced. Without multiple baseline assessments to establish a trend of symptom change, we cannot rule out regression to the mean as a potential explanation for symptom change. There are many unmeasured factors that could have contributed to changes in symptom trajectories. Our analyses did not include factors such as therapeutic alliance, external events outside of therapy, supervision quality, levels of implementation support available to therapists, therapist competency, or the degree content delivered during sessions was understood by youths. Finally, given our sample size of 35, our regression analyses were underpowered to detect participant characteristics as predictors of group membership based on the types of modules that were associated with decreases in symptoms for youths or associations between group membership and treatment outcomes. Additional work with larger sample sizes will be needed to build on this preliminary work to investigate other clinically relevant characteristics (e.g., risk profiles) as predictors of subgroups for whom certain modules are especially beneficial or detrimental.

Future Directions

Behavioral activation was the module most frequently associated with decreases in symptoms for depressed youths within this sample. Youths in this sample varied in the number of times that their therapist covered behavioral activation principles during sessions and further work is needed to determine how frequently behavioral activation should be used during therapy. Additional factors, such as how well a youth understands the content conveyed during a therapy session (e.g., how to apply the principles of behavioral activation), how often and how successfully they use the skill outside of sessions may also explain why certain modules benefit youths or not (Hundt et al., 2013). Additional consideration should be given to the sequencing of active therapeutic ingredients. Based on our results, we would focus on behavioral activation, cognitive restructuring, and problem solving. Youths with more severe symptoms may benefit from starting with behavioral activation before advancing to other skills. We have also proposed that some youths may have more success with problem solving after they have begun to challenge negative cognitions. Group designs such as sequential multiple assignment randomized trials or factorial designs could be used to investigate whether module sequence has an impact on outcomes. Finally, this work highlights why ongoing monitoring during treatment is vital. This component of evidence-based practice can inform treatment decision making, and frequent symptom monitoring can be used to address questions about how to improve treatment. Additional research is needed to identify the change processes and mechanisms by which these youths achieved symptom improvement, as understanding the active therapeutic ingredients and mechanisms of successful treatment for youths with internalizing disorders may lead to the development of more efficient and effective interventions (Ng et al., 2020).

Conclusions

For the majority of youths, specific modules of CBT were associated with increases or decreases in symptoms. Our findings show preliminary evidence that there is considerable heterogeneity across youths in which modules are associated with decreases or increases in symptoms or appear to lack any association with symptom change. If replicated in larger samples, the lack of association between the types of modules associated with changes in symptoms for youths and youths’ post-treatment internalizing symptoms might indicate that there are multiple pathways to symptom improvement for youth with depression (Ng et al., in press). However, many youths appear to benefit from behavioral activation, suggesting that it may be fruitful for researchers to examine how to enhance the effects of this module. Our results contribute to a growing body of evidence (Ng & Weisz, 2016) that can be used to inform treatment personalization for depressed youths.

Supplementary Material

1

Highlights.

  • Most depressed youths had changes in symptoms associated with specific CBT modules

  • The types of CBT modules associated with changes varied across youths

  • Behavioral activation was most frequently associated with decreases in symptoms

  • Baseline characteristics did not predict the type of modules that were impactful

  • There may be many pathways to symptom improvement for youths with depression

Funding:

This study was not supported by any specific grant from funding agencies in the public commercial, or not-for-profit sector; but the original parent study that provided the data was supported by grants (to John R. Weisz) from the Annie E. Casey Foundation, the Norlien Foundation, the John D. and Catherine T. MacArthur Foundation and the Child Health and Human Development Institute. Mei Yi Ng is supported by a National Institute of Mental Health grant (1R21MH126394-01). Katherine Venturo-Conerly reports working for the Shamiri Institute, a 501(c)3 non-profit organization and reports funding from the Templeton World Charity Foundation, the Harvard University Dean’s Competitive Fund, the U.S. National Institute of Mental Health (F31MH134555), the Agency Fund, and the Fund for Innovation in Development.

Footnotes

Publisher's Disclaimer: 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.

Declaration of interests: The authors declare the following potential competing interests: John R. Weisz is a co-author of the MATCH manual used in treatment of the youths in this study, and receives income from sales of the manual.

References

  1. Achenbach TM, McConaughy SH, Ivanova MY, & Rescorla LA (2011). Manual for the ASEBA Brief Problem Monitor (BPM) (Vol. 33). ASEBA. [Google Scholar]
  2. Achenbach TM, & Rescorla LA (2001). Manual for the ASEBA school-age forms & profiles: an integrated system of multi-informant assessment. University of Vermont, Research Center for Children, Youth, & Families. [Google Scholar]
  3. Box GEP, Jenkins GM, & Reinsel GC (2013). Time series analysis: Forecasting and control: Fourth edition. Time Series Analysis: Forecasting and Control: Fourth Edition, 1–746. 10.1002/9781118619193 [DOI] [Google Scholar]
  4. Brent DA, Kolko DJ, Birmaher B, Baugher M, Bridge J, Roth C, & Holder D (1998). Predictors of Treatment Efficacy in a Clinical Trial of Three Psychosocial Treatments for Adolescent Depression. Journal of the American Academy of Child & Adolescent Psychiatry, 37(9), 906–914. 10.1097/00004583-199809000-00010 [DOI] [PubMed] [Google Scholar]
  5. Chorpita BF, Reise S, Weisz JR, Grubbs K, Becker KD, & Krull JL (2010). Evaluation of the brief problem checklist: Child and caregiver interviews to measure clinical progress. Journal of Consulting and Clinical Psychology, 78(4), 526–536. 10.1037/a0019602 [DOI] [PubMed] [Google Scholar]
  6. Chorpita B, & Weisz J (2009). MATCH-ADTC: Modular Approach to Therapy for Children (MATCH) with anxiety, depression, trauma, or conduct problems.
  7. Coffman SJ, Martell CR, Dimidjian S, Gallop R, & Hollon SD (2007). Extreme Nonresponse in Cognitive Therapy: Can Behavioral Activation Succeed Where Cognitive Therapy Fails? Journal of Consulting and Clinical Psychology, 75(4), 531–541. 10.1037/0022-006X.75.4.531 [DOI] [PubMed] [Google Scholar]
  8. Cohen JR, So FK, Young JF, Hankin BL, & Lee BA (2019). Youth Depression Screening with Parent and Self-Reports: Assessing Current and Prospective Depression Risk. Child Psychiatry and Human Development, 50(4), 647. 10.1007/S10578-019-00869-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Courtney DB, Watson P, Krause KR, Chan BWC, Bennett K, Gunlicks-Stoessel M, Rodak T, Neprily K, Zentner T, & Szatmari P (2022). Predictors, Moderators, and Mediators Associated With Treatment Outcome in Randomized Clinical Trials Among Adolescents With Depression: A Scoping Review. JAMA Network Open, 5(2), e2146331–e2146331. 10.1001/JAMANETWORKOPEN.2021.46331 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cuijpers P, Karyotaki E, Eckshtain D, Ng MY, Corteselli KA, Noma H, Quero S, & Weisz JR (2020). Psychotherapy for Depression across Different Age Groups: A Systematic Review and Meta-analysis. In JAMA Psychiatry (Vol. 77, Issue 7, pp. 694–702). American Medical Association. 10.1001/jamapsychiatry.2020.0164 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cuijpers P, Reijnders M, & Huibers MJH (2019). The Role of Common Factors in Psychotherapy Outcomes. Annual Review of Clinical Psychology, 15(1), 207–231. 10.1146/annurev-clinpsy-050718-095424 [DOI] [PubMed] [Google Scholar]
  12. DeRubeis RJ, Cohen ZD, Forand NR, Fournier JC, Gelfand LA, & Lorenzo-Luaces L (2014). The Personalized Advantage Index: Translating Research on Prediction into Individualized Treatment Recommendations. A Demonstration. PLoS ONE, 9(1), e83875. 10.1371/journal.pone.0083875 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dietz LJ (2014). Social problem solving among depressed adolescents is enhanced by structured psychotherapies. Journal of Consulting and Clinical Psychology, 82(2), 202. 10.1037/A0035718 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dimidjian S, & Mccauley E (2016). Modular, Scalable, and Personalized: Priorities for Behavioral Interventions for Adolescent Depression. Clinical Psychology: Science and Practice, 23(1), 58–61. 10.1111/CPSP.12142 [DOI] [Google Scholar]
  15. Doss BD (2004). Changing the way we study change in psychotherapy. Clinical Psychology: Science and Practice, 11(4), 368–386. 10.1093/clipsy/bph094 [DOI] [Google Scholar]
  16. Dour HJ, Chorpita BF, Lee S, Weisz JR, Chorpita B, Garland A, Gibbons R, Glisson C, Green EP, Hoagwood K, Kelleher K, Landsverk J, Mayberg S, Miranda J, Palinkas L, & Schoenwald S (2013). Sudden gains as a long-term predictor of treatment improvement among children in community mental health organizations. Behaviour Research and Therapy, 51(9), 564–572. 10.1016/j.brat.2013.05.012 [DOI] [PubMed] [Google Scholar]
  17. Durland PH, Wyszynski CM, & Chu BC (2018). Predictors and Outcomes of Sudden Gains and Sudden Regressions in Cognitive Behavioral Therapy for Youth Anxiety. Behavior Therapy, 49(5), 823–835. 10.1016/J.BETH.2018.01.006 [DOI] [PubMed] [Google Scholar]
  18. Eckshtain D, Kuppens S, Ugueto A, Ng MY, Vaughn-Coaxum R, Corteselli K, & Weisz JR (2020). Meta-Analysis: 13-Year Follow-up of Psychotherapy Effects on Youth Depression. Journal of the American Academy of Child and Adolescent Psychiatry, 59(1), 45–63. 10.1016/j.jaac.2019.04.002 [DOI] [PubMed] [Google Scholar]
  19. Frederick J, Ng MY, Valente MJ, Chorpita BF, & Weisz JR (2022). Do specific modules of cognitive behavioral therapy for depression have measurable effects on youth internalizing symptoms? An idiographic analysis. Psychotherapy Research, 1–17. 10.1080/10503307.2022.2131475 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hawley KM, & Weisz JR (2003). Child, parent, and therapist (dis)agreement on target problems in outpatient therapy: The therapist’s dilemma and its implications. Journal of Consulting and Clinical Psychology, 71(1), 62–70. 10.1037/0022-006X.71.1.62 [DOI] [PubMed] [Google Scholar]
  21. Huibers MJH, Cohen ZD, Lemmens LHJM, Arntz A, Peeters FPML, Cuijpers P, & DeRubeis RJ (2015). Predicting Optimal Outcomes in Cognitive Therapy or Interpersonal Psychotherapy for Depressed Individuals Using the Personalized Advantage Index Approach. PloS One, 10(11). 10.1371/JOURNAL.PONE.0140771 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hundt NE, Mignogna J, Underhill C, & Cully JA (2013). The Relationship Between Use of CBT Skills and Depression Treatment Outcome: A Theoretical and Methodological Review of the Literature. Behavior Therapy, 44(1), 12–26. 10.1016/j.beth.2012.10.001 [DOI] [PubMed] [Google Scholar]
  23. Hyde JS, & Mezulis AH (2020). Gender Differences in Depression: Biological, Affective, Cognitive, and Sociocultural Factors. Harvard Review of Psychiatry, 28(1), 4–13. 10.1097/HRP.0000000000000230 [DOI] [PubMed] [Google Scholar]
  24. Jones NP, Siegle GJ, & Thase ME (2008). Effects of rumination and initial severity on remission to cognitive therapy for depression. Cognitive Therapy and Research, 32(4), 591–604. 10.1007/S10608-008-9191-0/TABLES/1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kennard BD, Clarke GN, Weersing VR, Asarnow JR, Shamseddeen W, Porta G, Berk M, Hughes JL, Spirito A, Emslie GJ, Keller MB, Wagner KD, & Brent DA (2009). Effective Components of TORDIA Cognitive–Behavioral Therapy for Adolescent Depression: Preliminary Findings. Journal of Consulting and Clinical Psychology, 77(6), 1033. 10.1037/A0017411 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, Wittchen HU, & Kendler KS (1994). Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Results from the National Comorbidity Survey. Archives of General Psychiatry, 51(1), 8–19. 10.1001/ARCHPSYC.1994.03950010008002 [DOI] [PubMed] [Google Scholar]
  27. Lewis AJ, Sae-Koew JH, Toumbourou JW, & Rowland B (2020). Gender differences in trajectories of depressive symptoms across childhood and adolescence: A multi-group growth mixture model. Journal of Affective Disorders, 260, 463–472. 10.1016/J.JAD.2019.09.027 [DOI] [PubMed] [Google Scholar]
  28. Martin F, & Oliver T (2019). Behavioral activation for children and adolescents: a systematic review of progress and promise. European Child and Adolescent Psychiatry, 28(4), 427–441. 10.1007/S00787-018-1126-Z/FIGURES/2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. McLeod BD (2011). Relation of the alliance with outcomes in youth psychotherapy: A meta-analysis. Clinical Psychology Review, 31(4), 603–616. 10.1016/J.CPR.2011.02.001 [DOI] [PubMed] [Google Scholar]
  30. Merikangas KR, He JP, Burstein M, Swanson SA, Avenevoli S, Cui L, Benjet C, Georgiades K, & Swendsen J (2010). Lifetime Prevalence of Mental Disorders in US Adolescents: Results from the National Comorbidity Study-Adolescent Supplement (NCS-A). Journal of the American Academy of Child and Adolescent Psychiatry, 49(10), 980. 10.1016/J.JAAC.2010.05.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Meyer AE, & Curry JF (2020). Moderators of Treatment for Adolescent Depression. Journal of Clinical Child & Adolescent Psychology, 00(00), 1–12. 10.1080/15374416.2020.1796683 [DOI] [PubMed] [Google Scholar]
  32. Michelson D, Hodgson E, Bernstein A, Chorpita BF, & Patel V (2022). Problem Solving as an Active Ingredient in Indicated Prevention and Treatment of Youth Depression and Anxiety: An Integrative Review. Journal of Adolescent Health, 71(4), 390–405. 10.1016/J.JADOHEALTH.2022.05.005 [DOI] [PubMed] [Google Scholar]
  33. Moritz S, & Bartz-Beielstein T (2017). imputeTS: Time series missing value imputation in R. R Journal, 9(1), 207–218. 10.32614/rj-2017-009 [DOI] [Google Scholar]
  34. Ng MY, DiVasto KA, Cootner S, Gonzalez N, & Weisz JR (2020). What do 30 years of randomized trials tell us about how psychotherapy improves youth depression? A systematic review of candidate mediators. Clinical Psychology: Science and Practice. 10.1111/cpsp.12367 [DOI] [Google Scholar]
  35. Ng MY, DiVasto KA, Gonzalez N, Cootner S, Lipsey MW, & Weisz JR (In Press). How do cognitive behavioral therapy and interpersonal psychotherapy improve youth depression? Applying meta-analytic structural equation modeling to three decades of randomized trials. Psychological Bulletin. [DOI] [PubMed] [Google Scholar]
  36. Ng MY, & Weisz JR (2016). Building a science of personalized intervention for youth mental health. Journal of Child Psychology and Psychiatry, 57 (3), 216–236. DOI: 10.1111/jcpp.12470 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Paul GL (1967). Strategy of Outcome Research In Psychotherapy. Journal of Consulting Psychology, 31(2), 109–118. 10.1037/h0024436 [DOI] [PubMed] [Google Scholar]
  38. Piccirillo ML, Beck ED, & Rodebaugh TL (2019). A Clinician’s Primer for Idiographic Research: Considerations and Recommendations. Behavior Therapy, 50(5), 938–951. 10.1016/J.BETH.2019.02.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Piper BJ, Gray HM, Raber J, & Birkett MA (2014). Reliability and Validity the Brief Problem Monitor, an Abbreviated Form of the Child Behavior Checklist. Psychiatry and Clinical Neurosciences, 68(10), 759. 10.1111/PCN.12188 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. SAS Institute Inc. (2015). SAS/STAT 14.1 User’s Guide. SAS Institute Inc. [Google Scholar]
  41. Velicer WF, & Colby SM (2005). A Comparison of Missing-Data Procedures for Arima Time-Series Analysis. Educational and Psychological Measurement, 65(4), 596–615. 10.1177/0013164404272502 [DOI] [Google Scholar]
  42. Webb CA, Beard C, Kertz SJ, Hsu KJ, & Björgvinsson T (2016). Differential role of CBT skills, DBT skills and psychological flexibility in predicting depressive versus anxiety symptom improvement. Behaviour Research and Therapy, 81, 12–20. 10.1016/j.brat.2016.03.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Webb CA, Stanton CH, Bondy E, Singleton P, Pizzagalli DA, & Auerbach RP (2019). Cognitive Versus Behavioral Skills in CBT for Depressed Adolescents: Disaggregating Within-Patient Versus Between-Patient Effects on Symptom Change. Journal of Consulting and Clinical Psychology, 87(5), 484–490. 10.1037/ccp0000393 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Weersing VR, Jeffreys M, Do MCT, Schwartz KTG, & Bolano C (2017). Evidence Base Update of Psychosocial Treatments for Child and Adolescent Depression. Journal of Clinical Child and Adolescent Psychology, 46(1), 11–43. 10.1080/15374416.2016.1220310 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Weisz JR, Kuppens S, Ng MY, Vaughn-Coaxum RA, Ugueto AM, Eckshtain D, & Corteselli KA (2019). Are Psychotherapies for Young People Growing Stronger? Tracking Trends Over Time for Youth Anxiety, Depression, Attention-Deficit/Hyperactivity Disorder, and Conduct Problems. Perspectives on Psychological Science: 14(2), 216–237. 10.1177/1745691618805436 [DOI] [PubMed] [Google Scholar]
  46. Weisz JR, Thomassin K, Hersh J, Santucci LC, MacPherson HA, Rodriguez GM, Bearman SK, Lang JM, Vanderploeg JJ, Marshall TM, Lu JJ, Jensen-Doss A, & Evans SC (2020). Clinician training, then what? Randomized clinical trial of child STEPs psychotherapy using lower-cost implementation supports with versus without expert consultation. Journal of Consulting and Clinical Psychology, 88(12), 1065–1078. 10.1037/ccp0000536 [DOI] [PubMed] [Google Scholar]
  47. Weisz JR, Ugueto AM, Herren J, Marchette LK, Bearman SK, Lee EH, Thomassin K, Alleyne A, Cheron DM, Lindsey Tweed J, Hersh J, Raftery-Helmer JN, Weissman AS, & Jensen-Doss A (2018). When the torch is passed, does the flame still burn? Testing a “train the supervisor” model for the child STEPs treatment program. Journal of Consulting and Clinical Psychology, 86(9), 726–737. 10.1037/ccp0000331 [DOI] [PubMed] [Google Scholar]
  48. Wilson CJ, Bushnell JA, Rickwood DJ, Caputi P, & Thomas SJ (2014). The role of problem orientation and cognitive distortions in depression and anxiety interventions for young adults. Http://Dx.Doi.Org/10.5172/Jamh.2011.10.1.52, 10(1), 52–61. 10.5172/JAMH.2011.10.1.52 [DOI] [Google Scholar]
  49. Young JF, Jones JD, Gallop R, Benas JS, Schueler CM, Garber J, & Hankin BL (2021). Personalized Depression Prevention: A Randomized Controlled Trial to Optimize Effects Through Risk-Informed Personalization. Journal of the American Academy of Child and Adolescent Psychiatry, 60(9), 1116–1126.e1. 10.1016/J.JAAC.2020.11.004 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1986842-supplement-1.docx (204.6KB, docx)

RESOURCES