How does the therapeutic alliance develop throughout cognitive behavioral therapy for panic disorder? Sawtooth patterns, sudden gains, and stabilization

Abstract

Objective: There has been little research on the development of the therapeutic alliance in cognitive behavioral therapy (CBT). This study aims to examine the development of therapeutic alliance in CBT for panic disorder. Method: Nineteen patients were treated with CBT for panic disorder. Pre- and post-session data of the therapeutic alliance and panic symptoms and cognitions were collected. Results: Several patterns were observed, including a sawtooth pattern (within-session improvements followed by decline between sessions; 63% of the patients), sudden gains in the alliance (58%), and late stabilization (89%). The sawtooth pattern was related to less symptom reduction between sessions (explained variance = 20–48%). Though not statistically significant, there were moderate effect sizes for the relationships between outcomes and early alliance and sudden gains (explained variance = 13–17%). Conclusions: Overall, results suggest that intensive data collection is likely to yield understanding of the relationship between therapy processes and outcomes.

Obiettivo: sono state effettuate poche ricerche sullo sviluppo dell'alleanza terapeutica nella terapia cognitivo comportamentale (CBT). Questo studio si propone di esaminare lo sviluppo dell'alleanza terapeutica nella CBT del disturbo di panico. Metodo: 19 pazienti con disturbo di panico sono stati trattati con la CBT. Sono stati raccolti dati circa l'alleanza terapeutica, i sintomi legati al panico e le cognizioni sia in fase pre- che in fase post-trattamento. Risultati: sono stati osservati diversi modelli, incluso un pattern a denti di sega (miglioramenti nel corso della seduta seguiti da peggioramenti fra le sedute; 63% dei pazienti), improvvisi miglioramenti dell'alleanza (58%), e una stabilizzazione tardiva (89%). Il pattern a denti di sega era legato ad una minore riduzione dei sintomi tra le sedute (varianza spiegata = 20–48%). Sebbene non statisticamente significativo, sono stati osservati moderati effect-size nelle relazioni tra i risultati finali, l'alleanza precoce e miglioramenti improvvisi (varianza spiegata = 13–17%). Conclusioni: nel complesso, i risultati suggeriscono che una ricca raccolta di dati consente di comprendere la relazione tra i processi terapeutici e gli esiti.

Objetivos: A investigação acerca do desenvolvimento da aliança terapêutica na terapia cognitivo-comportamental (TCC) é escassa. Este estudo pretende analisar o desenvolvimento da aliança terapêutica na TCC para a perturbação de pânico. Método: Dezanove clientes receberam TCC para a perturbação de pânico. Dados pré e pós-sessão acerca da aliança terapêutica e dos sintomas e cognições de pânico foram recolhidos. Resultados: Vários padrões foram observados, incluindo um padrão dente de serra (melhoria intrasessão seguida de agravamentos entre sessões; 63% dos clientes), ganhos súbitos na aliança terapêutica (58%) e estabilização tardia (89%). O padrão dente de serra associou-se a menor redução sintomática entre sessões (variância explicada = 20–48%). Embora estatisticamente não significativos, existiram effect sizes moderados para as relações entre resultado, aliança inicial e ganhos súbitos (variância explicada = 13–17%). Conclusão: Em geral, os resultados que sugerem que a recolha intensiva de dados poderá promover a compreensão da relação entre os processos e os resultados da terapia.

Ziel: Es gibt wenig Forschung über die Entwicklung der therapeutischen Allianz innerhalb der kognitiven Verhaltenstherapie (CBT). Die Studie zielt darauf ab, die Entwicklung von therapeutischer Allianz in CBT für Panikstörung zu untersuchen. Methode: Neunzehn Patienten mit Panikstörung wurden mit CBT behandelt. Prä- und Post-Sitzungsmessungen der therapeutischen Allianz, Paniksymptomen und Kognitionen wurden erhoben. Ergebnisse: Unterschiedliche Verläufe wurden beobachtet, inklusive eines Sägezahnverlaufs (Verbesserungen innerhalb der Sitzungen gefolgt von Verschlechterungen zwischen den Sitzungen; 63% der Patienten), plötzlicher Veränderungssprünge in der Allianz (58%) und späte Stabilisierung (89%). Der Sägezahnverlauf war mit geringerem Symptomrückgang zwischen den Sitzungen assoziiert (aufgeklärte Varianz = 20–48%). Obwohl statistisch nicht signifikant, gab es moderate Effektstärken für die Beziehungen zwischen Therapieergebnissen und früher Allianz und plötzlichen Veränderungen (aufgeklärte Varianz = 13–17%). Schlussfolgerung: Insgesamt legen die Ergebnisse nahe, dass intensive Datenerhebung wahrscheinlich dazu beiträgt, den Zusammenhang zwischen Therapieprozessen und -ergebnissen zu verstehen.

目的: 鮮少研究探討認知行為治療(CBT)中治療同盟關係之發展情況。本研究旨在檢視接受CBT恐慌症治療時，同盟關係的發展情況。方法：針對19位因恐慌症接受CBT治療的病人，搜集治療前／後的治療同盟關係、恐慌症狀及認知等資料。結果：觀察到許多的模式，包含一個鋸齒模式（63%的病人，在當次晤談時有改善，但是隨之在下次晤談間又退步），同盟關係的立即進展（58%），及延宕穩定性(89%)。鋸齒模式與晤談間較少的症狀減緩有關(可解釋變異量20–48%)。雖未達到顯著效果，但治療效果與早期的同盟關係、立即效果間有中等的效應值(可解釋變異量13–17%)。結論：整體而言，研究結果指出密集的資料搜集可能有助於了解治療過程與結果間的關係。

Keywords:

• therapeutic alliance
• CBT
• panic
• sudden gains

Acknowledgment

The authors would like to thank Robert Gallop for his invaluable assistance with statistical advice for this manuscript

Notes

¹ We considered the gains that occurred within session, since we had measured alliance both before and after each session, and theoretically it makes sense that alliance changes would occur mostly within session. We also modified the criteria suggested by Tang and DeRubeis (1999) in order to fit the alliance scale. These modifications are reported in the method section.

² However, our approach is different from that of Baldwin et al. (2007) in that Baldwin et al.'s dataset included therapist and patient levels. Therefore, they calculated therapist-level WAI scores by averaging patient-level WAI scores for each therapist separately. Our dataset, on the other hand, includes patient and time levels. Therefore, we calculated patient-level WAI scores by averaging time-level WAI scores for each patient separately.

³ Note that although the effect sizes are reported as percentages they do not represent the percentage of patients responding to the therapy but rather the percent of variance explained by a given predictor.

⁴ To assess the degree of mutli-collinearity between the two slopes, we fit multilevel models predicting change in ASI/WAI between sessions from the respective change in the session that preceded it. The results showed that change within a given session explained only a small proportion of the variance in change between sessions (8.77% for WAI; 6.45% for ASI), corresponding to a modest correlation only (−.30; −.25).

⁵ Note that due to some missing post-session data, the sum of the within- and between-session changes (1.75) resembles but is not entirely equal to the weekly changes in WAI (1.82).