Predictors of Response to Trauma-Focused Psychotherapy for PTSD: A Meta-Analysis

Keyan, D., Garland, N., Choi-Christou, J., Tran, J., O’Donnell, M., & Bryant, R. A. (2024). A systematic review and meta-analysis of predictors of response to trauma-focused psychotherapy for posttraumatic stress disorder. Psychological Bulletin, 150(7), 767–797. https://doi.org/10.1037/bul0000438

Rationale

This systematic review and meta-analysis was conducted to comprehensively examine baseline factors that predict response to trauma-focused psychotherapy (T-F psychotherapy) for posttraumatic stress disorder (PTSD). The rationale for this study stems from several key points in the existing literature:

What we know:

• T-F psychotherapy is considered the frontline treatment for PTSD, but up to 50% of patients do not respond optimally to this treatment (Bradley et al., 2005; Loerinc et al., 2015).
• Most attempts to augment T-F psychotherapy have focused on promoting extinction or inhibitory learning processes, based on fear conditioning models of PTSD (Ressler et al., 2022).
• These augmentation attempts have had limited success in improving overall treatment outcomes (Mataix-Cols et al., 2017).
• A broad array of patient and clinical factors beyond fear extinction processes have been associated with T-F psychotherapy outcomes in individual studies (Fonzo et al., 2020; Kline et al., 2023).

What’s the next step:

Given the limited success of current augmentation strategies and the diverse range of factors potentially influencing treatment response, a data-driven approach examining the relative importance of various predictors was needed. This meta-analysis aimed to:

1. Quantify the relationships between a comprehensive range of baseline factors and PTSD treatment outcome.
2. Elucidate potential candidates for augmenting T-F psychotherapy beyond fear extinction-based mechanisms.
3. Inform the development of more personalized, process-based approaches to PTSD treatment.

Method

The study followed PRISMA guidelines for systematic reviews and meta-analyses. The protocol was preregistered with PROSPERO (CRD42020162112).

Electronic database searches were conducted in three cycles:

1. PubMed, APA PsycInfo, PTSDpubs, and Cochrane Library
2. Additional search including “predictor” OR “moderator” terms
3. Hand searches of reference lists, relevant reviews, and major PTSD journals

Key search terms included variations of:

• PTSD and related terms
• Psychotherapy and treatment-related terms
• Study design terms (e.g., randomized, controlled)

Inclusion criteria:

• Full-text, peer-reviewed articles in English
• Adult samples (18+ years) exposed to DSM-defined traumatic events
• Participants meeting full or subthreshold PTSD criteria
• Any form of clinical trial for PTSD (RCT, uncontrolled, etc.)
• Use of validated PTSD outcome measures
• T-F psychotherapy as primary intervention

Exclusion criteria:

• Primary acute stress disorder or mixed PTSD/ASD samples
• Interventions targeting only select PTSD symptoms
• Non-English language articles without translations
• Grey literature (dissertations, conference proceedings, etc.)

Statistical measures:

• Random effects meta-analyses were conducted for 24 predictor categories
• Effect sizes were computed as bivariate correlations (r) between predictors and PTSD outcomes
• Heterogeneity was assessed using Q, I2, and T2 statistics
• Publication bias was evaluated using funnel plots, Egger’s test, and trim-and-fill analyses
• Moderator analyses examined the influence of study design and sample characteristics

Results

Fear Biology Factors:

• Greater activation in fear-related brain regions (r = -0.44) and higher psychophysiological reactivity (r = -0.46) predicted better treatment outcomes.
• Carrying risk alleles for genes modulating fear expression predicted poorer outcomes (r = 0.49).

Other Clinical Factors:

• Better executive functioning (r = -0.29), higher social support (r = -0.30), and better quality of life (r = -0.19) predicted better outcomes.
• Higher levels of trauma-related cognitions (r = 0.37), anger (r = 0.31), pain (r = 0.15), sleep issues (r = 0.45), and service-related disability (r = 0.12) predicted poorer outcomes.

Demographic Characteristics:

• Being male (r = 0.12), non-Caucasian (r = 0.16), and older (r = 0.28) predicted poorer outcomes.

Trauma-Related Characteristics:

• Childhood trauma (r = 0.28), higher trauma load (r = 0.37), longer time since trauma (r = 0.55), combat exposure (r = 0.33), and higher pretreatment PTSD severity (r = 0.29) predicted poorer outcomes.

Psychiatric Comorbidities:

• Higher levels of depression (r = 0.42), problematic alcohol use (r = 0.19), and personality disorder symptoms (r = 0.44) predicted poorer outcomes.

Moderator Analyses:

• Age moderated the effect of baseline depression, with older individuals showing worse outcomes.
• Trauma type moderated pretreatment PTSD severity effects, with combat trauma associated with poorer outcomes.

Insight

This comprehensive meta-analysis provides several key insights into predicting response to T-F psychotherapy for PTSD:

1. Multi-faceted predictors: While fear-related biological factors are important predictors of treatment response, a broad array of other clinical, demographic, and trauma-related factors also significantly influence outcomes. This suggests that a more holistic approach to understanding and augmenting treatment response is needed.
2. Relative importance of predictors: The study quantifies the relative strength of various predictors, allowing for prioritization in clinical assessment and treatment planning. For example, psychophysiological reactivity and cognitive distortions showed stronger associations with outcomes compared to demographic factors.
3. Complex interplay of factors: The findings highlight the complex nature of PTSD and its treatment. Factors like comorbid depression, executive functioning, and social support likely interact with fear-based mechanisms in influencing treatment response.
4. Potential treatment targets: The identified predictors suggest potential avenues for augmenting T-F psychotherapy beyond fear extinction processes. For instance, targeting cognitive distortions, improving sleep, or addressing comorbid depression may enhance overall treatment outcomes.
5. Personalized treatment approaches: The diverse range of predictors supports the development of more personalized, process-based approaches to PTSD treatment. Tailoring interventions based on individual patient profiles may improve overall response rates.

These findings extend previous research by providing a comprehensive, quantitative synthesis of predictor effects across a large number of studies. They challenge the dominant focus on fear extinction-based augmentation strategies and suggest a need for more multifaceted approaches to improving PTSD treatment outcomes.

Future research directions include:

• Examining interactions between multiple predictors to develop more sophisticated prediction models
• Investigating the effectiveness of personalized treatment approaches based on baseline predictor profiles
• Exploring the underlying mechanisms linking various predictors to treatment outcomes
• Developing and testing novel augmentation strategies targeting non-fear-based predictors of poor response

Strengths

1. Comprehensive scope: The meta-analysis included a wide range of predictor categories, providing a holistic view of factors influencing T-F psychotherapy outcomes.
2. Large sample size: With 114 studies and over 60,000 participants, the analysis had substantial statistical power to detect effects.
3. Rigorous methodology: The study adhered to PRISMA guidelines, used preregistration, and employed robust statistical techniques including random effects models and publication bias analyses.
4. Inclusion of diverse study designs: By including both randomized controlled trials and uncontrolled studies, the analysis captured a broader range of clinical contexts.
5. Examination of moderators: The study investigated how study design and sample characteristics influenced predictor effects, providing nuanced insights.
6. Data-driven approach: The categorization of predictors was based on the available literature rather than predetermined theories, allowing for a more objective assessment.
7. Clinically relevant effect size metric: Using bivariate correlations as the effect size measure facilitates interpretation and application of findings in clinical contexts.

Limitations

1. Arbitrary grouping of predictors: The categorization of predictors into meta-analytic groups, while necessary, may have obscured some nuances in individual predictor effects.
2. Exclusion of mediation and dropout studies: The focus on baseline predictors meant that important process variables and factors influencing treatment engagement were not captured.
3. Mixed risk of bias: The inclusion of studies with varying methodological quality may have introduced some bias into the overall findings.
4. Lack of consensus on treatment response: The absence of a standardized definition of treatment response across studies may have introduced variability in outcome measurement.
5. English language bias: Limiting inclusion to English language articles may have missed relevant studies from non-English speaking countries.
6. Focus on published literature: The exclusion of unpublished studies may have resulted in some publication bias, although this was assessed and accounted for in the analysis.
7. Limited exploration of predictor interactions: The analysis focused on individual predictor effects rather than more complex interactions between multiple factors.

Clinical Implications

The results of this meta-analysis have significant implications for clinical psychology practice and PTSD treatment:

1. Comprehensive assessment: Clinicians should conduct thorough baseline assessments covering multiple domains (e.g., fear-related biology, cognitive factors, comorbidities) to better predict treatment response and tailor interventions.
2. Personalized treatment planning: The diverse range of predictors supports a more personalized approach to PTSD treatment, potentially combining elements of T-F psychotherapy with targeted interventions for specific risk factors.
3. Expanded augmentation strategies: Treatment developers should consider a broader range of augmentation approaches beyond fear extinction, such as targeting cognitive distortions, improving sleep, or addressing comorbid conditions.
4. Process-based therapy: The findings support the development of more flexible, process-based treatment approaches that can address the heterogeneous nature of PTSD presentations.
5. Prioritizing risk factors: Clinicians and researchers can use the relative strength of predictor effects to prioritize assessment and intervention targets.
6. Trauma-informed care: The influence of trauma-related factors (e.g., childhood trauma, combat exposure) highlights the importance of trauma-informed approaches across various clinical settings.
7. Holistic treatment models: The relevance of factors like social support and quality of life suggests the value of more holistic treatment models that address broader aspects of functioning beyond PTSD symptoms.
8. Treatment expectations: Clinicians can use predictor information to set realistic treatment expectations and identify patients who may require more intensive or specialized interventions.
9. Research priorities: The findings inform research priorities for developing and testing novel PTSD treatment approaches and augmentation strategies.

Variables influencing results include study design characteristics, sample demographics, and trauma types. The consistent effects across diverse samples and methodologies, however, suggest broad applicability of the findings.

References

Primary reference

Keyan, D., Garland, N., Choi-Christou, J., Tran, J., O’Donnell, M., & Bryant, R. A. (2024). A systematic review and meta-analysis of predictors of response to trauma-focused psychotherapy for posttraumatic stress disorder. Psychological Bulletin, 150(7), 767–797. https://doi.org/10.1037/bul0000438

Other references

Bradley, R., Greene, J., Russ, E., Dutra, L., & Westen, D. (2005). A multidimensional meta-analysis of psychotherapy for PTSD. American Journal of Psychiatry, 162(2), 214-227.

Fonzo, G. A., Federchenco, V., & Lara, A. (2020). Predicting and managing treatment non-response in posttraumatic stress disorder. Current Treatment Options in Psychiatry, 7(2), 70-87.

Kline, A. C., Panza, K. E., Lyons, R., Kehle-Forbes, S. M., Hien, D. A., & Norman, S. B. (2023). Trauma-focused treatment for comorbid post-traumatic stress and substance use disorder. Nature Reviews Psychology, 2(1), 24-39.

Loerinc, A. G., Meuret, A. E., Twohig, M. P., Rosenfield, D., Bluett, E. J., & Craske, M. G. (2015). Response rates for CBT for anxiety disorders: Need for standardized criteria. Clinical Psychology Review, 42, 72-82.

Mataix-Cols, D., Fernández de la Cruz, L., Monzani, B., Rosenfield, D., Andersson, E., Pérez-Vigil, A., … & DCS Anxiety Consortium. (2017). D-cycloserine augmentation of exposure-based cognitive behavior therapy for anxiety, obsessive-compulsive, and posttraumatic stress disorders: a systematic review and meta-analysis of individual participant data. JAMA Psychiatry, 74(5), 501-510.

Ressler, K. J., Berretta, S., Bolshakov, V. Y., Rosso, I. M., Meloni, E. G., Rauch, S. L., & Carlezon Jr, W. A. (2022). Post-traumatic stress disorder: clinical and translational neuroscience from cells to circuits. Nature Reviews Neurology, 18(5), 273-288.

Keep Learning

Socratic questions for a college class to discuss this paper:

1. How might the diverse range of predictors identified in this study challenge our current understanding of PTSD and its treatment?
2. What are the potential benefits and drawbacks of adopting a more personalized, process-based approach to PTSD treatment based on these findings?
3. How could the strong predictive value of non-fear-based factors (e.g., cognitive distortions, social support) inform the development of novel treatment augmentation strategies?
4. In what ways might the interaction between different predictor categories (e.g., fear biology and comorbid depression) complicate our approach to PTSD treatment?
5. How could the findings regarding demographic and trauma-related predictors inform broader public health approaches to PTSD prevention and early intervention?
6. What ethical considerations might arise in using predictor information to guide treatment decisions or resource allocation in PTSD care?
7. How might the limitations of this study, such as the focus on published literature and English language articles, impact the generalizability of its findings to diverse global contexts?
8. In what ways could future research build upon these findings to develop more sophisticated, multifaceted models of PTSD treatment response?