Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Apr 5.
Published in final edited form as: J Clin Child Adolesc Psychol. 2023 Oct 5;54(5):567–577. doi: 10.1080/15374416.2023.2261547

Therapeutic Alliance, Attendance, and Outcomes in Youths Receiving CBT or Client-Centered Therapy for Anxiety

Deepika Bose a, Jeremy W Pettit b, Jennifer S Silk c, Cecile D Ladouceur c, Thomas M Olino d, Erika E Forbes e, Greg J Siegle e, Ronald E Dahl f, Phillip C Kendall d, Neal D Ryan e, Dana L McMakin b
PMCID: PMC10995113  NIHMSID: NIHMS1933555  PMID: 37796228

Abstract

Objective:

Positive associations between therapeutic alliance and outcome (e.g., youth symptom severity) have been documented in the youth anxiety literature; however, little is known about the conditions under which early alliance contributes to positive outcomes in youth. The present study examined the relations between therapeutic alliance, session attendance, and outcomes in youths (N = 135; 55.6% female) who participated in a randomized clinical trial testing the efficacy of cognitive-behavioral therapy or client-centered therapy for anxiety (Silk et al., 2018).

Method:

We evaluated a conceptual model wherein: (1) early alliance indirectly contributes to positive outcomes by improving session attendance; (2) alliance-outcome associations differ by intervention type, with stronger associations in cognitive-behavioral therapy compared to client-centered therapy; and (3) alliance-outcome associations vary across outcome measurement timepoints, with the effect of early alliance on outcomes decaying over time.

Results:

Contrary to hypotheses, provider ratings of early alliance predicted greater youth-rated anxiety symptom severity post-treatment (i.e., worse treatment outcomes). Session attendance predicted positive youth-rated outcomes, though there was no indirect effect of early alliance on outcomes through session attendance.

Conclusions:

Results show that increasing session attendance is important for enhancing outcomes and do not support early alliance as a predictor of outcomes.

Keywords: therapeutic alliance, treatment attendance, child/adolescent, anxiety, cognitive-behavioral therapy, supportive therapy

Therapeutic alliance refers to the emotional bond and collaborative relationship between patients and providers (Bordin, 1979). Research on therapeutic alliance consistently documents small associations with outcomes in youth psychotherapy (mean effect sizes from r = 0.14 to r = 0.22; Shirk, Karver, & Brown, 2011; McLeod, 2011; Karver et al., 2018; Bose et al., 2022). Therapeutic alliance is theorized to enhance outcomes in part by increasing youth attendance and involvement in treatment (Chu & Kendall, 2004; Shirk et al., 2008; Karver et al., 2008; McLeod et al., 2014). Within the context of youth anxiety treatment specifically, a strong alliance is theorized to improve outcomes by increasing youth involvement in skill-building tasks (Chu et al., 2004) and facilitating the completion of therapeutic tasks that may be perceived as distressing (McLeod et al., 2014). For example, in exposure-based therapy for youth anxiety, youth may feel more comfortable confronting their fears (a challenging exposure task) in a therapeutic setting if they perceive a strong, positive alliance with their provider early in treatment (Crawford et al., 2018).

In spite of the strong theoretical basis for how alliance contributes to outcomes in youth anxiety interventions, empirical findings have been mixed (e.g., Chiu et al., 2009; Hudson et al., 2014; Marker et al., 2013), with an estimated mean effect size of r = 0.14 (Bose et al., 2022). Possible explanations for mixed findings include methodological differences between studies (e.g., measurement timing, informant source, participant characteristics, intervention types) and variations in how alliance is conceptualized and measured (Cummings et al., 2013; Fjermestad et al., 2016). The mixed findings highlight the need for continued examination of therapeutic alliance in relation to outcomes in youth anxiety treatment, including how alliance might contribute to outcomes and variables that may influence the strength of the alliance-outcome association in youth anxiety interventions (Bose et al., 2022). Below, we present relevant theory and research that provided the impetus for the conceptual model we evaluated.

Alliance, Attendance, and Outcomes

Models of the alliance-outcome association posit that a positive, early alliance predicts increased session attendance (therapy “dosage”), which in turn predicts positive outcomes (Shirk, 2001; Hawley & Weisz, 2005; Shirk et al., 2008). Consistent with these models, research has documented positive associations between alliance and youth session attendance (Kazdin, Holland & Crowley, 1997; Garcia & Weisz, 2002; Hawley & Weisz, 2005; Shirk et al., 2008; Ormhaug & Jensen, 2018), and between session attendance (higher “doses” of treatment) and youth intervention outcomes (Angold et al., 2000; Nock & Ferriter, 2005).

Although links between alliance, attendance, and outcomes have been established, we are aware of only one study from the youth literature that has tested the full model wherein the association between alliance and outcomes is accounted for by attendance (Shirk et al., 2008). In this study, Shirk and colleagues found that in a sample of 54 adolescents receiving cognitive-behavioral therapy (CBT) for depression, therapist-rated alliance was associated with session attendance but the indirect path from alliance to outcome through attendance was not supported. As noted by Shirk and colleagues (2008), additional testing of this model in larger samples with higher statistical power to detect indirect effects is needed before drawing conclusions. Further, this model has never been tested in a sample of youth receiving CBT for anxiety disorders. As noted, the alliance may be particularly important in the context of exposure-based therapy for youth anxiety (McLeod et al., 2014).

Factors Influencing the Strength of the Alliance-Outcome Association

In addition to the emerging literature on the mechanisms by which alliance contributes to outcomes, there is growing evidence that the strength of the alliance-outcome association depends on substantive factors (e.g., youth and intervention characteristics) and study methodology (McLeod, 2011; Karver et al., 2018). Below, we describe two factors that may influence the strength of the alliance-outcome association in youth psychotherapy: treatment type and outcome measurement timing.

Treatment Type

In a study examining the relation between alliance, treatment involvement, and outcomes in adolescents receiving CBT or nondirective supportive therapy (NST) for depression, Karver and colleagues (2008) found that alliance and involvement may be differentially associated with outcomes. Specifically, findings suggested stronger alliance-outcome associations (r = .35, .31) and a stronger involvement-outcome association (r = .56) in the CBT arm than the NST arm (r = −.18, −.11 for alliance-outcome and r = −.08 for involvement-outcome). These findings led the researchers to hypothesize that therapeutic alliance and youth involvement may partially explain treatment outcome in CBT but not supportive therapy (Karver et al., 2008). In addition, a meta-analysis by Karver and colleagues (2018) suggested that the alliance-outcome association in youth psychotherapy may be larger in behavioral treatments (r = .23) compared to nonbehavioral treatments like supportive therapy (r = .10).

Altogether, there is emerging evidence that the alliance-outcome association differs across interventions for youth, with some evidence for stronger effects in CBT compared to supportive therapy, at least in the treatment of youth depression. CBT is a structured, goal-oriented intervention that focuses on identifying and changing maladaptive thoughts and behaviors, and typically involves active collaboration between providers and patients to develop coping skills that reduce distress and impairment (Beck & Dozois, 2011). Supportive therapies (e.g., client-centered therapy) tend to provide a more nondirective approach to treatment, emphasizing the therapeutic relationship and providing emotional support (Bothart & Watson, 2020).

We are unaware of any studies that have compared alliance-outcome associations across multiple intervention types in youth anxiety. Given key differences between CBT for depression and anxiety (e.g., use of behavioral exposures in anxiety interventions, motivational factors that may factor uniquely in depression), it is useful to evaluate this model in a sample of youths receiving treatment for anxiety. From a theoretical standpoint, we propose a strong early alliance may be particularly important in promoting positive outcomes in CBT, given that youth may feel more comfortable confronting challenging situations (i.e., facing their fears) when a strong alliance with their provider has been established early in treatment (McLeod et al., 2014). Further, exposure tasks in CBT may provide additional opportunities to strengthen the provider-youth alliance (and thereby enhance outcomes) as transient ruptures are repaired within the session (Kendall et al., 2009). In contrast, the need for a strong early alliance to complete intervention tasks may not be equally critical in nondirective, supportive interventions. Examination of the moderating effect of treatment type on alliance-outcome associations in youth anxiety will advance theory and provide insights on whether therapeutic alliance and alliance-building should (or need not be) emphasized equally across interventions.

Outcome Measurement Timing

The strength of alliance-outcome association may also vary depending on the timepoint at which outcome is assessed (Chu, Suveg, Creed, & Kendall, 2010). In a study of 34 youths receiving CBT for anxiety, Chiu et al. (2009) found that early alliance predicted mid-treatment outcomes, but not post- or follow-up outcomes. Additionally, a systematic review and meta-analysis of alliance-outcome associations in interventions for youth internalizing disorders found that alliance-outcome associations tend to be stronger when outcome is measured mid-treatment compared to post-treatment or follow-up (Bose et al., 2022). Together, these findings suggest that the predictive effect of alliance on outcomes may decay as time passes, such that early alliance is a better predictor of outcomes at mid-treatment than post- or follow-up. Under that scenario, the predictive effect of early alliance on outcomes is limited to shorter-term (e.g., next session or mid-treatment) outcomes. We evaluated that scenario in the present study.

Present Study

As evident from the preceding review, empirical testing on the role of therapeutic alliance in interventions for youth internalizing disorders has lagged far behind theory development and has occurred in a small number of studies with small sample sizes. The present study provides a necessary expansion of this empirical literature by examining the relation between therapeutic alliance, session attendance, and anxiety severity outcomes in youths who participated in a randomized clinical trial testing the efficacy of CBT and client-centered therapy (CCT), a form of supportive therapy, for anxiety (Silk et al., 2018). We evaluated a conceptual model wherein early (session 2) alliance influences post-treatment anxiety severity indirectly through session attendance.

We also examined whether the strength of the alliance-outcome association differs by treatment type or outcome measurement timing. Specifically, we tested whether prior research findings of interventions for adolescent depression extend to youth anxiety with respect to stronger alliance-outcome associations in CBT compared to supportive therapy (CCT). We hypothesized that treatment type would moderate the alliance-outcome association, such that alliance is a stronger predictor of outcomes in CBT than CCT. We also hypothesized that the strength of the alliance-outcome association will decrease over time in both treatment arms, with stronger alliance-outcome associations when outcome is measured at mid-treatment than post-treatment. Findings from this study will provide novel insights on potential intervention targets for youth anxiety (e.g., increasing session attendance by improving the alliance) and will inform the field with regards to the circumstances under which early alliance contributes most strongly to outcomes.

Method

Participants

Participants included 135 children and adolescents (age range: 9–14; 55.6% female) who were randomized using a 2:1 ratio to receive CBT (N = 92) or CCT (N = 43). Eighty-eight percent of youths were White, 5.9% Multiracial, 3.7% Black, 3.0% Hispanic or Latino, and 1.5% Other. Family income was $10,000–30,000 (10.4%), $30,000–60,000 (25.2%), and $60,000–100,000+ (58.5%). Eight participants (5.9%) did not indicate their income.

Providers included seven master’s and doctoral-level clinicians. Providers were trained in both CBT and CCT by experts, received weekly supervision, and provided therapy to youths in both treatment arms.

Measures

Therapeutic Alliance

Therapeutic alliance was assessed using the revised Therapeutic Alliance Scale for Children – Child/Therapist Versions (TASC-C/T; Creed & Kendall, 2005). The revised TASC-C/T are 12-item questionnaires that measure youth- and provider-reports of the youth’s perspective on the alliance (e.g., “I like spending time with my therapist,”/ “The child likes spending time with you, the therapist”). Respondents rank items on a 4-point Likert scale ranging from “Not at all” to “Very much,” with 5 items that are reversed scored (e.g., “The child appears eager to have session end,” “The child is resistant to coming to therapy”) such that higher scores indicating a stronger alliance. Past research on the TASC-C/T supports internal consistency and convergent and predictive validity (Hawley & Weisz, 2005; Fjermestad et al., 2012; Accurso & Garland, 2015).

Outcome: Youth Symptom Severity

The Screen for Child Anxiety and Related Emotional Disorders – Child/Parent Versions (SCARED-C/P ; Birmaher et al., 1997) are 41-item youth- and caregiver-reports of anxiety symptom severity across multiple anxiety domains (e.g., separation, social, school, panic, general anxiety). Respondents rank items on a 3-point Likert scale (“Not true or hardly ever true” to “Very true or often true”), with higher scores reflecting higher anxiety severity. Scores on the SCARED-C/P range from 0 to 82, with a recommended clinical cut off of 25 for SCARED total scores (Birmaher et al., 1997). The SCARED-C/P has satisfactory to excellent test-retest reliability, and good convergent and divergent validity (Birmaher et al., 1999; Monga et al., 2000).

Procedures

Study procedures from the original trial were approved by the principal investigator’s Institutional Review Board. All participants provided written informed consent and assent to participate in the study. Youth demographic information was collected at intake and session attendance (i.e., total number of sessions attended, including rescheduled sessions) was tracked by providers. Therapeutic alliance was measured at Session 2 and youth symptom severity was measured every session. Youths and therapists completed the TASC independently. Youths and parents completed the SCARED-C/P before the session, while waiting for the provider in the waiting room. Youths and parents were instructed to place the completed forms in a blank envelope which they then placed in a dropbox located in the waiting room. Youths and therapists were masked to all alliance ratings. In the present study, we examined SCARED-C/P scores at three timepoints: “early” (Session 2), “mid-” (Session 7), and “post” (post-treatment). “Initial symptom severity” refers to SCARED-C/P scores assessed at pre-treatment. “Outcome” refers to SCARED-C/P scores assessed at post-treatment.

Youths in both intervention arms received 14 sessions of individual therapy and 2 parent-only sessions (16 sessions total). Providers delivered CBT from the Coping Cat manual (Kendall & Hedtke, 2006) and CCT from a manual developed by Cohen and colleagues (2004). In Coping Cat, youth are taught skills for anxiety management (e.g., identifying somatic cues and thought patterns, progressive muscle relaxation, developing coping thoughts) and complete graded exposures to approach anxiety-provoking situations. CCT is a supportive, nondirective intervention in which providers use active listening, empathy, and encouragement to address the youth’s problems (Cohen et al., 2004). Treatment integrity and fidelity were measured in both treatment arms by reviewing 16% of recorded video tapes. Ratings indicated 98% fidelity for CBT and 99% fidelity for CCT. Youths in both treatment arms responded positively to treatment, with no between-arm differences on response rates. However, youths who received CBT were more likely to achieve diagnostic recovery, and a higher percentage of youth in CBT compared to CCT were in recovery at 1-year follow up. Additional information about the study design, procedures, and findings are provided in Silk and colleagues (2018).

Data Analytic Plan

Missing data occurred for the TASC-T (3.7%), TASC-C (3.0%), SCARED-P pre (3.0%), SCARED-P early (10.4%), SCARED-P mid (11.9%), SCARED-P post (14.8%), SCARED-C pre (3.7%), SCARED-C early (8.1%), SCARED-C mid (11%), and SCARED-C post (16.3%). Analysis of missing data patterns yielded significant associations between missingness on the SCARED-C post and being Hispanic (r = 0.28, p < .001) and missingness on the SCARED-C and SCARED-P mid and lower income (r = −0.30 and −0.29, ps <.001) after Holm-Bonferonni corrections (Holm, 1979). To address missing data bias, TASC and SCARED data were estimated using full information maximum likelihood (FIML) procedures, which estimate missing data based on model parameters (Enders, 2001).

To test the indirect effect of alliance on post-treatment anxiety severity through session attendance, we conducted an indirect effects test with bootstrapping (Preacher & Hayes, 2008). An indirect effects test is conceptually similar to a mediation test, but does not require temporal precedence for all paths in the model. The bootstrapping procedure produces a point estimate for the indirect path after repeatedly sampling the data. The indirect path is the product of path a (the effect of the independent variable on the proposed mediator) and path b (the effect of the proposed mediator on the dependent variable). An indirect effect is considered present if the indirect indirect path is significant, regardless of the significance of the direct path or paths a and b (Hayes & Rockwood, 2017).

To test the moderating effect of treatment condition on the alliance-outcome association, we tested moderated regressions using procedures outlined by Aiken and West (1991). Specifically, we employed ordinary least squares regression to examine the effect of the independent variable, X (early alliance) on the dependent variable, Y (youth outcome); the effect of each moderator (M) on Y; and the interaction between X and M on Y (i.e., the effect of X*M on Y). Significant interactions were probed by computing the conditional effect of X on Y at different values of M (i.e., 1 SD below mean, mean, 1 SD above mean).

To explore the effect of outcome measurement timing on the alliance-outcome association, we ran individual linear regression analyses for each outcome timepoint (early, mid, post-treatment).

We conducted analyses on 4 models, with 2 informants for alliance (youth and provider) and 2 informants for outcome (youth and parent). To control for the possibility of family-wise error rates, we applied Holm-Bonferroni adjustments (Holm, 1979).

Youth sex designated at birth was associated with TASC-T and SCARED-C scores (see Results), and so was added as a covariate in models including the TASC-T and SCARED-C. To control for initial symptom severity, we included pre-treatment SCARED scores as a covariate in each of the tested models.

Regression, indirect effect, and moderation analyses were conducted in the R statistical program (R Core Team, 2017), lavaan package (Roseel, 2012).

Results

Means, standard deviations, and bivariate correlations between variables are presented in Table 1. Female sex designated at birth was associated with lower income, higher provider-rated alliance, and higher youth-rated symptom severity at pre-, early-, mid-, and post-treatment. Youth age was not associated with any other measured variable. Pre-treatment youth-rated symptom severity was lower in the CBT arm than the CCT arm. Higher income was associated with higher session attendance. Provider-rated alliance was also associated with youth-rated alliance and higher youth-rated symptom severity at all timepoints. Youth-rated alliance was associated with lower youth-rated symptom severity at pre-treatment. Provider- and youth-rated alliance did not differ by treatment arm.

Table 1.

Bivariate Correlations, Means, and Standard Deviations

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)
(1) Age --
(2) Sexa −.08 --
(3) Ethnicity −.06 .07 --
(4) Income .05 −.24** −.13 --
(5) Treatment Cond. −.02 −.004 .03 .01 --
(6) Attendance −.04 .02 −.04 .24** .03 --
(7) TASC-T .12 .23** −.09 −.11 .02 −.10 --
(8) TASC-C −.16 .16 .05 −.02 −.02 −.12 .27** --
(9) SCARED-P pre −.09 .14 .10 −.03 .01 −.10 −.07 −.18* --
(10) SCARED-P early −.16 .02 .13 −.15 .04 −.10 .04 −.01 .52** --
(11) SCARED-P mid −.10 .03 .15 −.12 .10 .03 −.02 .02 .47** .71** --
(12) SCARED-P post −.18 .07 .05 −.10 −.06 −.05 −.04 −.01 .55* .60** .69** --
(13) SCARED-C pre −.03 .19* .06 .01 −.21* .03 .11 .03 .18 .25** .15 .16 --
(14) SCARED-C early −.10 .25** .08 −.16 .02 −.07 .25* .14 .27** .40** .37** .24** .64** --
(15) SCARED-C mid −.11 .18** −.02 −.08 −.02 −.11 .31* .17 .28** .36** .43** .40** .51** .75** --
(16) SCARED-C post −.08 .27** .11 −.06 −.08 −.08 .34* .09 .30** .24* .23* .46** .41** .52** .78* --
Mean 10.94 -- -- -- -- 14.36 36.21 41.23 36.38 30.07 27.15 22.21 38.50 28.54 25.33 17.80
SD 1.46 -- -- -- -- 3.73 5.85 5.03 11.81 12.45 13.66 12.97 12.70 14.92 17.80 15.10
Open in a new tab

Note. TASC-T = Therapeutic Alliance Scale for Children – Therapist version. Therapeutic Alliance Scale for Children – Child version. SCARED-P = Screen for Child Anxiety and Related Emotional Disorder – Parent version. SCARED-C = Screen for Child Anxiety and Related Emotional Disorders – Child version.

a

Binary sex was coded as 0 for male and 1 for female.

Indirect Effect Model: Session Attendance

The indirect path between early alliance and outcomes through session attendance was not significant for either alliance or outcome informant. The path between session attendance and youth-rated outcomes was significant using youth- and provider-ratings of alliance, with higher attendance associated with lower youth-rated symptom severity at post-treatment. The direct path between provider-rated alliance and youth-rated outcome was also significant, with higher ratings of alliance associated with higher post-treatment youth-rated symptom severity (i.e., worse outcomes). The indirect effect models using parent-rated outcomes were not significant. See Table 2 for path estimates of the 4 indirect models that evaluated youth- and provider-rated alliance with youth- and parent-rated outcome.

Table 2.

Indirect Path Between Alliance and Outcome through Session Attendance

Youth-rated Alliance Provider-rated Alliance b
Parent-rated outcome a Est. SE p Est. SE p
 Alliance and attendance (path a) −0.09 0.06 .08 −0.07 0.05 .15
 Attendance and outcome (path b) −0.87 0.75 .24 −0.98 0.75 .19
 Direct path between alliance and outcome (path c’) 0.26 0.21 .22 0.04 0.19 .84
 Indirect path between alliance and outcome (path a*path b) 0.08 0.09 .33 0.07 0.08 .33
Youth-rated Outcome a,b Est. SE p Est. SE p
 Alliance and attendance (path a) −0.10 0.06 .07 −0.07 0.05 .15
 Attendance and outcome (path b) −2.29* 0.94 .02 −2.34** 0.88 <.01
 Direct path between alliance and outcome
(path c’)		 0.26 0.27 0.34 0.89** 0.23 <.01
 Indirect path between alliance and outcome (path a*path b) 0.24 0.16 0.14 0.18 0.14 .20
Open in a new tab

Note. Est. = Estimate. SE = standard error.

a

Initial symptom severity was added as a covariate in each of the models.

b

Sex was added as a covariate for models including youth-rated outcome and provider-rated alliance.

*

p < .05.

**

p < .01.

Interaction Model: Treatment Arm

Moderation analyses revealed that treatment arm did not significantly moderate the effect of early alliance on outcome according to either alliance or outcome informant (Table 3).

Table 3.

Interaction Models Examining Moderating Effect of Treatment Type on the Alliance-Outcome Association

Youth-rated Alliance Provider-rated Alliance
Est. SE p Est. SE p
Parent-rated Outcome
 Initial severity 0.62** 0.09 <.0001 0.61** 0.09 <.0001
 Sex -- -- -- 0.004 2.26 .99
 Alliance 0.19 0.90 .84 −0.29 0.67 .67
 Tx.Arm −3.94 20.46 .84 −8.76 14.00 .53
 Alliance*Tx. Arm 0.05 0.49 .91 0.19 0.38 .61
Youth-rated Outcome a
 Initial severity 0.45** 0.11 <.001 0.46** 0.11 <.001
 Sex 6.08* 2.80 .03 3.84 2.67 .15
 Alliance −0.67 1.11 .55 0.53 0.82 .52
 Tx. Arm. −16.74 25.20 .51 −26.65 17.16 .12
 Alliance*Tx. Arm. 0.41 0.61 .50 0.74 0.47 .11
Open in a new tab
*

p < .05

**

p < .01. Tx. = Treatment.

a

Sex was added as a covariate for models including youth-rated outcome and provider-rated alliance.

Alliance-Outcome Association Across Measurement Timepoints

See Table 4 for the predictive effect of alliance on outcomes across measurement timepoints and informants. Controlling for initial symptom severity and sex designated at birth, higher provider-rated early alliance significantly predicted higher youth-rated anxiety symptoms at all three measurement timepoints, after applying Holm-Bonferroni adjustments (Holm, 1979). The strength of the effect appeared to increase from early treatment to mid-treatment and from mid-treatment to post-treatment.

Table 4.

Predictive Effect of Alliance on Outcomes Across Measurement Timepoints

Youth-rated Alliance Provider-rated Alliance b
Est. SE p Est. SE p
Parent-rated outcome a
 Early (Session 2) 0.25 0.19 .21 0.22 0.17 .19
 Mid (Session 7) 0.39 0.22 .08 0.10 0.20 .62
 Post 0.33 0.21 .11 0.04 0.19 .84
Youth-rated outcome a,b
 Early (Session 2) 0.31 0.21 .14 0.52** 0.18 .004
 Mid (Session 7) 0.39 0.26 .13 0.74** 0.23 .001
 Post 0.02 0.27 .95 0.72** 0.23 .002
Open in a new tab

Note.

a

Initial symptom severity was added as a covariate in each of the models.

b

Sex was added as a covariate for models including youth-rated outcome and provider-rated alliance.

**

p < .01.

Post-hoc Exploration of Alliance-Outcome Association in CBT

Given that unequal sample sizes in treatment groups (N = 43 in CCT; N = 92 in CBT) can lead to the underestimation of moderation effects, even when significant group differences are present (Memon et al., 2019), we ran post-hoc exploratory analyses to examine whether the positive association between provider-rated early alliance and youth-rated anxiety symptoms (i.e., higher early alliance associated with worse outcomes) differed by intervention groups in the absence of a significant interaction. Findings revealed that, while controlling for initial symptom severity and youth sex, provider-rated early alliance predicted greater youth anxiety severity (i.e., worse outcomes) in the CBT arm (β = .34, p < .01), but not the CCT arm (β = .31, p > .05). To gain insight into the positive association between early alliance and youth anxiety severity in the CBT arm, we ran additional post-hoc exploratory analyses to examine whether higher alliance may have predicted a lower number of exposures administered and/or lower intensity of exposures, thereby leading to worse youth outcomes. Intensity of exposures was measured by averaging the youth’s subjective units of distress (SUDS) across in-session exposures. Specifically, we explored: (1) whether higher provider-rated early alliance predicted a lower number and/or lower intensity of exposures, and (2) the indirect effect of number of exposures and exposure intensity on the alliance-outcome association. Findings revealed that higher provider-rated alliance did not predict number of exposures administered or exposure intensity, although there was a nonsignificant trend toward predicting higher exposure intensity (β = .34, p = .06). The indirect paths of alliance on outcomes via number of exposures and exposure intensity were nonsignificant (effect sizes = −.01 and .10, 95%Cis = −.13-.08 and −.06-.36, respectively).

Discussion

Findings from the current study provide insights on the associations between early alliance, attendance, and outcomes in youths receiving CBT or CCT for anxiety. In the original trial (Silk et al., 2018), CBT outperformed CCT in terms of diagnostic recovery at post-treatment and 1-year follow-up. Our study expanded on this research by examining how therapeutic alliance may have differentially affected outcomes across the two treatment arms. Findings indicate that higher provider-ratings of early alliance predicted greater youth-rated anxiety symptom severity (i.e., worse outcomes), but did not support a conceptual model wherein alliance predicted outcomes indirectly through session attendance. In addition, alliance-outcome associations were not moderated by treatment arm and did not vary across intervention timepoints.

The nonsignificant indirect path of alliance on outcome through session attendance may be partially attributed to a weak association between early alliance and attendance. Consistent with prior research (e.g., Angold et al., 2000), session attendance predicted positive outcomes in the indirect effect model using youth- and provider-ratings of alliance and youth-ratings of outcome. This finding suggests that attendance plays an important role in predicting youth outcomes. Notably, session attendance was also associated with higher income, which may be indicative of the structural issues facing lower income individuals.

The predictive effect of session attendance on outcomes implies that session attendance should be targeted in treatment, especially in lower income families. Session attendance may be targeted by assessing and resolving barriers to treatment prior to treatment initiation, measuring patient strengths and needs, soliciting youth and parent commitment to the intervention early-on, and sending appointment reminders on a consistent basis (Lefforge, Donohue, & Strada, 2007; Becker et al., 2015). With lower income families, who may face challenges such as transportation, cost of services, and child-care for siblings (Gonzalez, 2005; Bringewatt & Gershoff, 2009; Kazdin & Blase, 2011), accessibility of services (and thereby, treatment attendance) may be improved by offering alternative intervention formats, including teletherapy (Crum & Comer, 2016; Nelson & Patton, 2016).

Treatment type did not moderate the alliance-outcome association. However, closer exploratory examination of the alliance-outcome association revealed a positive association between provider-rated early alliance and youth-rated symptom severity (i.e., worse outcomes) in the CBT arm but not the CCT arm. This finding was unexpected and inconsistent with past research, and suggests that providers would benefit from reflecting on alliance early on in CBT with the understanding that it may be a poor prognostic sign of recovery; in our study, it predicted worse outcome.

To gain insight into variables that might explain the unexpected alliance-outcome finding, in post-hoc analyses we explored whether a higher early alliance was associated with a lower number and/or lower intensity of exposures in CBT. Our rationale was that providers who perceived a strong initial alliance with youths might have shied away from challenging youths with intense exposures, out of fear of rupturing the alliance (Whiteside et al., 2016; Deacon et al., 2013). Given that exposures are considered the “active ingredient” in CBT for youth anxiety (Whiteside et al., 2020), shying away from challenging exposures may have inadvertently led to worse youth outcomes. Alternatively, given that providers may fear that strong exposures would rupture the alliance (although research suggests the contrary; Kendall et al., 2009), providers who perceived a weaker early alliance may have felt as though there was less to “lose,” and subsequently felt more comfortable challenging youths during exposures, facilitating more positive outcomes. Our exploratory analyses did not support these possibilities, as alliance was not associated with number or intensity of exposures implemented in CBT. An alternative possibility we were unable to explore is that providers experienced higher levels of empathy for more severe or less responsive-to-treatment cases and higher empathy drove higher provider ratings of early alliance. Additionally, providers may have perceived a stronger early alliance with participants who were agreeable and perhaps eager to please, but were not motivated or willing to engage in treatment in an effective manner. Future research is encouraged to evaluate these and other potential explanations of the associations between early alliance and outcome in CBT for youth anxiety disorders. We cannot definitively rule out the possibility that alliance predicting worse outcomes in CBT is a spurious finding, and should therefore be interpreted cautiously.

Therapeutic alliance did not predict outcomes in the CCT arm or when youths rated alliance in the CBT arm, and did not vary across outcome measurement timepoints. These findings add to the mixed literature on alliance-outcome associations in youth anxiety interventions (Bose et al., 2022), and are consistent with past findings on nonsignificant alliance-outcome associations in supportive therapy (Karver et al., 2008) and when youths rate the alliance (Creed, 2007; Anderson et al., 2012; Marker et al., 2013; Stjerneklar et al., 2019; Fjermestad et al., 2016). These findings also suggest that the effect of early alliance on outcomes (by informant) does not vary across outcome timepoints, thereby adding to the mixed literature on timing effects (Bose et al., 2022).

Female sex was associated with lower income, higher youth-rated symptom severity, and higher provider-rated alliance. This finding is consistent with past research indicating higher anxiety rates and higher initial alliance among females compared to males (Bekker & van Mens-Verhulst, 2007; Langer, McLeod, & Weisz, 2011). To adjust for the significant associations, we added sex as a covariate in models using provider-rated alliance and youth-rated symptom severity. This finding highlights the need for further research on associations between youth sex and provider-rated alliance.

Results from the current study should be interpreted in light of its strengths and limitations. Strengths include a rigorous, RCT treatment design; examination of the prospective association between early alliance and outcomes; analysis of alliance-outcome associations in multiple treatment arms (CBT, CCT); multiple outcome assessment timepoints (pre, early, mid, post); and multi-informant (youth, parent) ratings of outcome. With respect to limitations, the unequal sample sizes in the treatment arms (N = 43 in CCT; N = 92 in CBT) may have led to an underestimation of moderation effects (Memon et al., 2019). Second, we relied on single timepoint assessments of alliance, which may not fully capture the complexity of the relationship between alliance and outcomes. Other alliance-related factors, including alliance shifts and provider-youth agreement on alliance change, may be stronger indicators of outcome in both CBT (Chiu et al., 2009; Marker et al., 2013; Fjermestad et al., 2016) and supportive therapy interventions. Third, we did not have access to parent- or observer ratings of alliance or mid-treatment assessments of provider-rated outcome, and relied on ratings of anxiety severity as the primary outcome measure for this study. Observer ratings of alliance provide complementary information and address limitations in youth-, parent-, and provider-report (e.g., demand characteristics, social desirability, etc.) (McLeod & Weisz, 2005). Parent alliance contributes to youth outcomes differently from child alliance (Hawley & Weisz, 2005), so should also be assessed when possible. Further, although assessment of symptom severity maximizes construct specificity when evaluating intervention efficacy, it may over- or underestimate the effect of a predictor (Becker, Chorpita, & Daleiden, 2011). Thus, measuring the effect of alliance on other measures of outcome (e.g., adaptive functioning, life satisfaction, hope) may also provide important insights. Fourth, all therapists in the study were women, which limits the generalizability of findings. Fifth, our binary assessment of sex and gender may have led to gender misclassifications. Future research is encouraged to adopt inclusive measures of sex and gender in order to promote gender inclusivity, prevent misclassications, and increase the generalizability of findings (Cameron & Stinson, 2019). Sixth, the lack of ethnic diversity and tightly controlled nature of the RCT from which this sample was derived from also limits the generalizability of findings.

Altogether, results suggest that increasing session attendance is important for enhancing youth outcomes and that early provider perceptions of alliance predict worse youth-rated outcomes in CBT for anxiety but not CCT. Future research aimed at understanding the extent to which exposure intensity, client distress, and specific aspects of alliance interact in CBT for youth anxiety could be useful in creating formal recommendations for protocol modification.

Funding details:

This work was supported by the National Institute of Mental Health under Grant 5P50MH080215.

Footnotes

Disclosure statement: The authors report there are no competing interests to declare.

References

  1. Accurso EC, & Garland AF (2015). Child, caregiver, and therapist perspectives on therapeutic alliance in usual care child psychotherapy. Psychological Assessment, 27(1), 347–352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Aiken LS, & West SG (1991). Multiple Regression: Testing and Interpreting Interactions. Thousand Oaks, CA, US: Sage Publications, Inc. [Google Scholar]
  3. Anderson RE, Spence SH, Donovan CL, March S, Prosser S, & Kenardy J. (2012). Working alliance in online cognitive behavior therapy for anxiety disorders in youth: comparison with clinic delivery and its role in predicting outcome. Journal of Medical Internet Research, 14(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Angold A, Costello EJ, Burns BJ, Erkanli A, & Farmer EM (2000). Effectiveness of nonresidential specialty mental health services for children and adolescents in the “real world”. Journal of the American Academy of Child & Adolescent Psychiatry, 39(2),154–160. [DOI] [PubMed] [Google Scholar]
  5. Beck AT, & Dozois DJ (2011). Cognitive therapy: Current status and future directions. Annual Review of Medicine, 62, 397–409. [DOI] [PubMed] [Google Scholar]
  6. Becker KD, Chorpita BF, & Daleiden EL (2011). Improvement in symptoms versus functioning: How do our best treatments measure up? Administration and Policy in Mental Health and Mental Health Services Research, 38(6), 440–458. [DOI] [PubMed] [Google Scholar]
  7. Becker KD, Lee BR, Daleiden EL, Lindsey M, Brandt NE, & Chorpita BF (2015). The common elements of engagement in children’s mental health services: Which elements for which outcomes?. Journal of Clinical Child & Adolescent Psychology, 44(1), 30–43. [DOI] [PubMed] [Google Scholar]
  8. Birmaher B, Brent DA, Chiappetta L, Bridge J, Monga S, & Baugher M. (1999). Psychometric properties of the screen for child anxiety related emotional disorders (SCARED): A replication study. Journal of the American Academy of Child and Adolescent Psychiatry, 38, 1230–1236. [DOI] [PubMed] [Google Scholar]
  9. Birmaher B, Khetarpal S, Brent D, Cully M, Balach L, Kaufman J, et al. (1997). The Screen for Child Anxiety Related Emotional Disorders (SCARED): Scale construction and psychometric characteristics. Journal of the American Academy of Child and Adolescent Psychiatry, 36(4), 545–553. [DOI] [PubMed] [Google Scholar]
  10. Bordin ES (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research & Practice, 16(3), 252–260. [Google Scholar]
  11. Bose D, Arencibia D, Costales G, Viswesvaran C, Bickman L. & Pettit JW (2022). Therapeutic alliance in psychosocial interventions for youth internalizing disorders: A systematic review and preliminary meta-analysis. Clinical Psychology: Science and Practice, 29(2), 124–136. [Google Scholar]
  12. Bothart AC, & Watson JC (2020). Person-centered and emotion-focused psychotherapies. In Messer SB & Kaslow NJ (Eds.) Essential Psychotherapies, Theory and Practice (4th ed.). New York, NY: Guilford Press. [Google Scholar]
  13. Bringewatt EH, & Gershoff ET (2010). Falling through the cracks: Gaps and barriers in the mental health system for America’s disadvantaged children. Children and Youth Services Review, 32(10), 1291–1299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Cameron JJ, & Stinson DA (2019). Gender (mis) measurement: Guidelines for respecting gender diversity in psychological research. Social and Personality Psychology Compass, 13(11), e12506. [Google Scholar]
  15. Chiu AW, McLeod BD, Har K, & Wood JJ (2009). Child–therapist alliance and clinical outcomes in cognitive behavioral therapy for child anxiety disorders. Journal of Child Psychology and Psychiatry, 50(6), 751–758. [DOI] [PubMed] [Google Scholar]
  16. Chu BC, & Kendall PC (2004). Positive association of child involvement and treatment outcome within a manual-based cognitive-behavioral treatment for children with anxiety. Journal of Consulting and Clinical Psychology, 72(5), 821–829. [DOI] [PubMed] [Google Scholar]
  17. Chu BC, Choudhury MS, Shortt AL, Pincus DB, Creed TA, & Kendall PC (2004). Alliance, technology, and outcome in the treatment of anxious youth. Cognitive and Behavioral Practice, 11(1), 44–55. [Google Scholar]
  18. Chu BC, Suveg C, Creed TA, & Kendall PC (2010). Involvement shifts, alliance ruptures, and managing engagement over therapy. In Castro-Blanco D. & Karver MS (Eds.), Elusive alliance: Treatment engagement strategies with high-risk adolescents (p. 95–121). American Psychological Association. [Google Scholar]
  19. Cohen JA, Deblinger E, Mannarino AP, & Steer RA (2004). A multisite, randomized controlled trial for children with sexual abuse- related PTSD symptoms. Journal of the American Academy of Child & Adolescent Psychiatry, 43(4), 393–402. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Crawford EA, Frank HE, Palitz SA, Davis JP, & Kendall PC (2018). Process factors associated with improved outcomes in CBT for anxious youth: Therapeutic content, alliance, and therapist actions. Cognitive Therapy and Research, 42(2), 172–183. [Google Scholar]
  21. Creed TA (2007). A mediation model of early predictors of treatment outcome within cognitive-behavioral therapy for children with anxiety disorders: Child involvement, therapist behavior, and alliance. Dissertation Abstracts International: Section B: The Sciences and Engineering, 68(3-B), 1919. [Google Scholar]
  22. Creed TA, & Kendall PC (2005). Therapist alliance-building behavior within a cognitive-behavioral treatment for anxiety in youth. Journal of Consulting and Clinical Psychology, 73, 498–505. [DOI] [PubMed] [Google Scholar]
  23. Crum KI, & Comer JS (2016). Using synchronous videoconferencing to deliver family-based mental healthcare. Journal of Child and Adolescent Psychopharmacology, 26(3), 229–234. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Cummings CM, Caporino NE, Settipani CA, Read KL, Compton SN, March J,…& Kendall PC (2013). The therapeutic relationship in cognitive-behavioral therapy and pharmacotherapy for anxious youth. Journal of Consulting and Clinical Psychology, 81(5), 859–864. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Deacon BJ, Farrell NR, Kemp JJ, Dixon LJ, Sy JT, Zhang AR, & McGrath PB (2013). Assessing therapist reservations about exposure therapy for anxiety disorders: The Therapist Beliefs about Exposure Scale. Journal of Anxiety Disorders, 27(8), 772–780. [DOI] [PubMed] [Google Scholar]
  26. Enders CK (2001). The performance of the full information maximum likelihood estimator in multiple regression models with missing data. Educational and Psychological Measurement, 61(5), 713–740. [Google Scholar]
  27. Fjermestad KW, Lerner MD, McLeod BD, Wergeland GJH, Heiervang ER, Silverman WK, Ost L, De Los Reyes A, Havik OE, & Haugland BS (2016). Therapist‐youth agreement on alliance change predicts long‐term outcome in CBT for anxiety disorders. Journal of Child Psychology and Psychiatry, 57(5), 625–632. [DOI] [PubMed] [Google Scholar]
  28. Fjermestad KW, McLeod BD, Heiervang ER, Havik OE, Öst LG, & Haugland BS (2012). Factor structure and validity of the therapy process observational coding system for Child Psychotherapy–Alliance Scale. Journal of Clinical Child & Adolescent Psychology, 41(2), 246–254. [DOI] [PubMed] [Google Scholar]
  29. Garcia JA, & Weisz JR (2002). When youth mental health care stops: therapeutic relationship problems and other reasons for ending youth outpatient treatment. Journal of Consulting and Clinical Psychology, 70(2), 439–443. [PubMed] [Google Scholar]
  30. Gonzalez MJ (2005). Access to mental health services: The struggle of poverty affected urban children of color. Child and Adolescent Social Work Journal, 22, 245–256. [Google Scholar]
  31. Hawley KM, & Weisz JR (2005). Youth versus parent working alliance in usual clinical care: Distinctive associations with retention, satisfaction, and treatment outcome. Journal of Clinical Child and Adolescent Psychology, 34(1), 117–128. [DOI] [PubMed] [Google Scholar]
  32. Hayes AF, & Rockwood NJ (2017). Regression-based statistical mediation and moderation analysis in clinical research: Observations, recommendations, and implementation. Behaviour Research and Therapy, 98, 39–57. [DOI] [PubMed] [Google Scholar]
  33. Holm S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6(2), 65–70. [Google Scholar]
  34. Hudson JL, Kendall PC, Chu BC, Gosch E, Martin E, Taylor A, & Knight A. (2014). Child involvement, alliance, and therapist flex-ibility: Process variables in cognitive-behavioural therapy for anxietydisorders in childhood. Behaviour Research and Therapy, 52, 1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Karver MS, DeNadai AS, Monahan M, & Shirk SR (2018). Meta-analysis of the prospective relation between alliance and outcome in child and adolescent psychotherapy. Psychotherapy, 55(4), 341–355. [DOI] [PubMed] [Google Scholar]
  36. Karver M, Shirk S, Handelsman JB, Fields S, Crisp H, Gudmundsen G, & McMakin D. (2008). Relationship processes in youth psychotherapy: Measuring alliance, alliance-building behaviors, and client involvement. Journal of Emotional and Behavioral Disorders, 16(1), 15–28. [Google Scholar]
  37. Kazdin AE, & Blase SL (2011). Rebooting psychotherapy research and practice to reduce the burden of mental illness. Perspectives on Psychological Science, 6(1), 21–37. [DOI] [PubMed] [Google Scholar]
  38. Kazdin AE, Holland L, & Crowley M. (1997). Family experience of barriers to treatment and premature termination from child therapy. Journal of Consulting and Clinical Psychology, 65(3), 453–463. [DOI] [PubMed] [Google Scholar]
  39. Kendall PC, & Hedtke K. (2006). The Coping Cat Workbook. Ardmore, PA: Workbook Publishing. [Google Scholar]
  40. Kendall PC, Comer JS, Marker CD, Creed TA, Puliafico AC, Hughes AA, ... & Hudson J. (2009). In-session exposure tasks and therapeutic alliance across the treatment of childhood anxiety disorders. Journal of Consulting and Clinical Psychology, 77(3), 517–525. [DOI] [PubMed] [Google Scholar]
  41. Lefforge NL, Donohue B, & Strada MJ (2007). Improving session attendance in mental health and substance abuse settings: a review of controlled studies. Behavior Therapy, 38(1), 1–22. [DOI] [PubMed] [Google Scholar]
  42. Marker CD, Comer JS, Abramova V, & Kendall PC (2013). The reciprocal relationship between alliance and symptom improvement across the treatment of childhood anxiety. Journal of Clinical Child & Adolescent Psychology, 42(1), 22–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. McLeod BD (2011). Relation of the alliance with outcomes in youth psychotherapy: A meta-analysis. Clinical Psychology Review, 31(4), 603–616. [DOI] [PubMed] [Google Scholar]
  44. McLeod BD, Islam NY, Chiu AW, Smith MM, Chu BC, & Wood JJ (2014). The relationship between alliance and client involvement in CBT for child anxiety disorders. Journal of Clinical Child & Adolescent Psychology, 43(5), 735–741. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. McLeod BD, & Weisz JR (2005). The therapy process observational coding system-alliance scale: measure characteristics and prediction of outcome in usual clinical practice. Journal of Consulting and Clinical Psychology, 73(2), 323–333. [DOI] [PubMed] [Google Scholar]
  46. Memon MA, Cheah JH, Ramayah T, Ting H, Chuah F, & Cham TH (2019). Moderation analysis: issues and guidelines. Journal of Applied Structural Equation Modeling, 3(1), 1–11. [Google Scholar]
  47. Monga S, Birmaher B, Chiappetta L, Brent D, Kaufman J, Bridge J, et al. (2000). Screen for child anxiety-related emotional disorders (SCARED): Convergent and divergent validity. Depression and Anxiety, 12, 85–91. [DOI] [PubMed] [Google Scholar]
  48. Nelson EL, & Patton S. (2016). Using videoconferencing to deliver individual therapy and pediatric psychology interventions with children and adolescents. Journal of Child and Adolescent Psychopharmacology, 26(3), 212–220. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Nock MK, & Ferriter C. (2005). Parent management of attendance and adherence in child and adolescent therapy: A conceptual and empirical review. Clinical Child and Family Psychology Review, 8(2), 149–166. [DOI] [PubMed] [Google Scholar]
  50. Ormhaug SM, & Jensen TK (2018). Investigating treatment characteristics and first-session relationship variables as predictors of dropout in the treatment of traumatized youth. Psychotherapy Research, 28(2), 235–249. [DOI] [PubMed] [Google Scholar]
  51. Preacher KJ, & Hayes AF (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879–891. [DOI] [PubMed] [Google Scholar]
  52. R Core Team (2017). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/. [Google Scholar]
  53. Rosseel Y. (2012). lavaan: An R Package for Structural Equation Modeling. Journal of Statistical Software, 48(2), 1–36. URL http://www.jstatsoft.org/v48/i02/. [Google Scholar]
  54. Rubin DB (1987). Multiple imputation for nonresponse in surveys. New York, NY: Wiley. [Google Scholar]
  55. Shirk S. (2001). The road to effective child psychological services: Treatment process and outcome research. In Hughes J, LaGreca A, & Conoley J. (Eds.), Handbook o Psychological Services for Children and Adolescents (pp. 43–59). New York: Oxford University Press. [Google Scholar]
  56. Shirk SR, Gudmundsen G, Kaplinski HC, & McMakin DL (2008). Alliance and outcome in cognitive-behavioral therapy for adolescent depression. Journal of Clinical Child & Adolescent Psychology, 37(3), 631–639. [DOI] [PubMed] [Google Scholar]
  57. Shirk SR, Karver MS, & Brown R. (2011). The alliance in child and adolescent psychotherapy. Psychotherapy, 48(1), 17–24. [DOI] [PubMed] [Google Scholar]
  58. Silk JS, Tan PZ, Ladouceur CD, Meller S, Siegle GJ, McMakin DL, ... & Ryan ND (2018). A randomized clinical trial comparing individual cognitive behavioral therapy and child-centered therapy for child anxiety disorders. Journal of Clinical Child & Adolescent Psychology, 47(4), 542–554. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Stjerneklar S, Hougaard E, & Thastum M. (2019). Guided internet-based cognitive behavioral therapy for adolescent anxiety: Predictors of treatment response. Internet Interventions, 15, 116–125. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Whiteside SP, Deacon BJ, Benito K, & Stewart E. (2016). Factors associated with practitioners’ use of exposure therapy for childhood anxiety disorders. Journal of Anxiety Disorders, 40, 29–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Whiteside SP, Sim LA, Morrow AS, Farah WH, Hilliker DR, Murad MH, & Wang Z. (2020). A meta-analysis to guide the enhancement of CBT for childhood anxiety: exposure over anxiety management. Clinical Child and Family Psychology Review, 23(1), 102–121. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES