Mechanisms of CBT and IPT in Youth Depression: A 30-Year MASEM Review

Ng, M. Y., DiVasto, K. A., Gonzalez, N.-a-r., Cootner, S., Lipsey, M. W., & Weisz, J. R. (2023). 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, 149(9-10), 507–548. https://doi.org/10.1037/bul0000395

Rationale

Youth depression treatments yield only modest effects, and the mechanisms driving therapeutic change remain unclear.

Identifying how interventions work can inform refinement and enhance efficacy (Kazdin & Nock, 2003).

Prior studies suggest cognitive shifts and behavioral activation contribute to symptom reduction, but findings are inconsistent and underpowered when tested individually (Chu & Harrison, 2007; Ng et al., 2020).

No quantitative synthesis has yet pooled mediation effects across randomized trials to determine robust and treatment‐specific mechanisms.

This systematic review and meta‐analytic structural equation modeling addresses this gap by integrating evidence on candidate mediators in CBT and IPT for youth depression, clarifying common versus unique pathways of change.

Method

The review followed PRISMA guidelines, with a detailed flowchart documenting identification, screening, eligibility, and inclusion of studies.

Although no protocol was preregistered, all data, extraction code, and syntaxes are openly available on OSF.

Search Strategy

• Databases: APA PsycInfo and PubMed
• Time Frame: January 1963 through September 2020
• Search Terms: A combination of 21 psychotherapy-related root terms (e.g., “psychother-,” “counsel-”) and MeSH headings (e.g., “clinical assessment,” “child/adolescent,” English language filter) .
• Supplementary Sources: Backward and forward citation tracking, author–year searches, NIMH repositories, and ProQuest Dissertations.

Eligibility Criteria

• Inclusion:
  • Randomized controlled trials comparing CBT or IPT to control conditions (waitlist, placebo, usual care)
  • Participants aged 4–18 with elevated depressive symptoms or formal diagnosis
  • Continuous measures of at least one candidate mediator (CM) and depressive symptoms assessed at posttreatment
• Exclusion:
  • Non‐RCT designs or non‐youth samples
  • Interventions not primarily targeting depression
  • Insufficient or incompatible outcome/mediator data

Study Selection & Yield

From 15,361 records initially identified, 9,544 abstracts were screened; 611 depression‐relevant RCTs were assessed in full, yielding 51 trials targeting youth depression.

Of these, 28 CBT trials (N = 3,521) and 7 IPT trials (N = 370) reported suitable CM and outcome data and were included .

Data Extraction

For each arm, we extracted means, standard deviations, and sample sizes to compute Hedges’s g for treatment effects on mediators and outcomes (Path a and c), and zero‐order correlations (r) between mediator change and symptom change (Path b).

Effect‐size conversions aligned all paths so that positive values indicated beneficial change.

Extraction was conducted in duplicate by independent coders, with discrepancies resolved by consensus; risk of bias was assessed using standard domains (randomization, blinding, attrition).

Results

Study Selection and Characteristics

• Trials Included: 28 CBT trials (N = 3,521) and 7 IPT trials (N = 370) met inclusion criteria out of 51 depression‐focused RCTs identified .
• Participants: Youth aged 4–18 with elevated depressive symptoms or diagnosis.
• Candidate Mediators (CMs): Seven CMs were assessed across studies: negative cognition, pleasant activities, social engagement, family functioning, problem solving, cognitive reframing, and avoidance.

---

CBT Mediator Findings

1. Negative Cognition
   • Path a (Treatment → Mediator): Hedges’s g = 0.25 (95% CI [0.18, 0.32], p < .001)
   • Path b (Mediator → Outcome): r = 0.43 (95% CI [0.36, 0.50], p < .001)
   • Indirect Effect (ab): β = .030 (95% CI [.019, .043], p < .01)
   • Interpretation: Change in negative cognition reliably mediates CBT’s impact on depressive symptoms .
2. Pleasant Activities
   • Full Sample (k = 5 RCTs): Path a g = 0.12 (ns); ab β = .011 (ns)
   • Independent Comparisons (k = 3): ab β = .023 (p < .01)
   • Interpretation: Limited evidence suggests behavioral activation may mediate symptom change in a subset of trials .
3. Other Candidate Mediators
   • Social Engagement, Family Functioning, Problem Solving, Reframing, Avoidance:
     • Path a effect sizes uniformly below g = 0.10; deemed too small for formal mediation testing in MASEM.

---

Treatment Specificity: CBT versus IPT

1. Negative Cognition
   • CBT (k = 28): Path a g = 0.25 → ab β = .030 (p < .01)
   • IPT (k = 3): Path a g = 0.62 (p < .01) → ab β = .046 (ns)
   • Subgroup Test: No significant difference in indirect effects (p = .053), indicating negative cognition is a common mediator across both therapies.
2. Social Engagement
   • CBT (k = 14): Path a g = 0.11 → ab β = .007 (ns)
   • IPT (k = 5): Path a g = 0.46 (p < .01) → ab β = .052 (p < .01)
   • Subgroup Test: Significant difference (p = .009), supporting social engagement as an IPT‐specific mechanism .
3. Family Functioning
   • CBT (k = 13): Path a g = 0.01 → ab β = .001 (ns)
   • IPT (k = 6): Path a g = 0.28 (p < .05) → ab β = .029 (p < .05)
   • Subgroup Test: Difference reaches significance (p = .049), indicating family functioning mediates IPT but not CBT .

---

Summary of Main Findings

• CBT: Negative cognition is the only robust mediator; pleasant activities show tentative mediation in a small subset of trials.
• IPT: Social engagement and family functioning emerge as treatment‐specific mediators, while negative cognition operates as a common factor across both therapies.

Insight

• Negative Cognition: A consistent mediator across therapies but likely a common factor or correlate of symptom change rather than a CBT-specific mechanism.
• Behavioral Activation: Pleasant activities show promise; BA-focused protocols (with functional analysis, avoidance reduction) may yield stronger mediation.
• IPT-Specific Pathways: Enhancing social and family functioning aligns with IPT theory and predicts symptom change, underscoring the importance of interpersonal skills modules.
• CBT Refinements: Lack of mediation by problem solving and reframing suggests need to bolster these modules or improve their measurement.

Clinical Implications

Implementing these findings can sharpen youth depression care by focusing on validated change pathways.

Across therapies, routinely assessing and targeting negative thought patterns through age-appropriate cognitive restructuring can enhance outcomes.

In CBT, structured behavioral activation—using activity scheduling and functional analysis—should replace vague “pleasant activities” to boost engagement and symptom relief.

For IPT, prioritizing interpersonal strategies that improve social connectedness and family dynamics will capitalize on its unique mechanisms.

Embedding brief, session-by-session mediator assessments enables clinicians to monitor progress and adapt focus in real time.

Finally, training and supervision must emphasize mechanism mastery, and funding should support trials that isolate and strengthen these high-yield components.

References

Ng, M. Y., DiVasto, K. A., Gonzalez, N.-a-r., Cootner, S., Lipsey, M. W., & Weisz, J. R. (2023). 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, 149(9-10), 507–548. https://doi.org/10.1037/bul0000395

Kazdin, A. E., & Nock, M. K. (2003). Delineating mechanisms of change in child and adolescent therapy: Methodological issues and research recommendations. Journal of Child Psychology and Psychiatry, 44(8), 1116–1129. https://doi.org/10.1111/1469-7610.00195 PubMedSciepub

Chu, B. C., & Harrison, T. L. (2007). Disorder-specific effects of CBT for anxious and depressed youth: A meta-analysis of candidate mediators of change. Clinical Child and Family Psychology Review, 10(4), 352–372. https://doi.org/10.1007/s10567-007-0028-2 Rutgers University

Ng, M. Y., DiVasto, K. A., Cootner, S., González, N.-A.-R., & Weisz, J. R. (2023). 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, 30(4), 396–419. https://doi.org/10.1111/cpsp.12367