Attachment Anxiety, Therapeutic Alliance Predict Treatment Response in Social Anxiety Disorder

upset woman
upset woman, depression, anxiety,
The researchers identified greater baseline attachment anxiety and poorer therapeutic alliance as predictors of low treatment response in patients with social anxiety disorder.

Alongside initial symptom severity, attachment anxiety and therapeutic alliance may be key predictors of treatment response in patients with social anxiety disorder (SAD), according to study data published in the Journal of Anxiety Disorders.

Researchers extracted data from a randomized clinical trial conducted as part of the Social Phobia Psychotherapy Network (SOPHO-NET) trial, which sought to investigate the effects of short-term cognitive behavior therapy (CBT) and short-term psychodynamic psychotherapy (PDT) on SAD. Outpatients from multiple clinics in Germany were randomly assigned to receive either CBT or PDT for up to 30 individual sessions. Therapy sessions were administered by study clinicians, each of whom had an advanced degree in clinical psychology.

The primary outcome measure was the change in Liebowitz Social Anxiety Scale (LSAS) score over time. The LSAS was administered at baseline/session 1, session 8, session 15, and the final session. The study used latent state variables and latent class analysis to identify change patterns in patients receiving CBT and PDT. Logistic regression was performed to assess predictors of symptom trajectories.  

The present analysis included 357 patients with SAD (mean age, 34.6±7.9 years; 54.6% women). In the full sample, overall symptom reduction was large (P <.05), and 38.4% of patients achieved remission. Analyses identified 3 latent classes in the patient sample: (1) patients with high initial impairment and significant improvements (n=57; 16%); (2) patients with moderate initial impairment and significant improvements (n=225; 63%); and (3) patients with high initial impairment and only moderate symptom reduction (n=75; 21%). Remission rates in these 3 classes were 68.4%, 43.6%, and 0%, respectively.

Related Articles

In logistic regression models, patients in Class 3 were less likely to have a long-term partner relationship, had greater attachment-related anxiety at baseline, more often received PDT vs CBT, and had lower-rated therapist alliance at session 8. In addition, patients in Class 3 were more likely to be treated by therapists in training than by therapists with completed training. Compared with Class 3, patients in Class 2 displayed fewer initial interpersonal and depressive symptoms, were more often treated with CBT vs PDT, and displayed greater alliance with therapists at session 8. In post-hoc analyses, both the CBT and PDT groups displayed similar improvement trajectories. However, patients in class 3 who received PDT had greater baseline symptom severity than patients who received CBT.

“Overall, our study suggests that patients with SAD can be assigned to different symptom change patterns and that the change patterns are specifically affected by the patients´ attachment style and the therapeutic alliance in an early stage of the therapy,” the researchers concluded. As a study limitation, they noted that patients with incomplete treatment whose data were not included in the study may represent a more severe symptomatology subset.

The investigators concluded that “psychotherapy of SAD should anticipate therapeutic strategies related to a patient’s attachment and specifically focus on the establishment of a solid therapeutic alliance in an early therapy stage.”


Altmann U, Gawlytta R, Hoyer J, et al. Typical symptom change patterns and their predictors in patients with social anxiety disorder: a latent class analysis. J Anxiety Disord. 2020;71:102200.