A randomized trial of interpersonal therapy versus supportive therapy for social anxiety disorder

Joshua D. Lipsitz, Merav Gur, Donna Vermes, Eva Petkova, Jianfeng Cheng, Nina Miller, Joseph Laino, Michael R. Liebowitz, Abby J. Fyer

Research output: Contribution to journalArticlepeer-review

53 Scopus citations


Seventy patients seeking treatment for social anxiety disorder (SAD) were randomly assigned to 14 weekly individual sessions of interpersonal therapy (IPT) or supportive therapy (ST). We hypothesized that IPT, a psychotherapy with established efficacy for depression and other psychiatric disorders, would lead to greater improvement than ST. Patients in both groups experienced significant improvement from pretreatment to posttreatment. However, improvement with IPT was not superior to improvement with ST. Mean scores on the Liebowitz Social Anxiety Scale decreased from 67.7 to 46.9 in the IPT group and 64.5 to 49.8 in the ST group. There were also no differences in proportion of responders between IPT and ST. Only for a scale measuring concern about negative evaluation (Brief Fear of Negative Evaluation Scale) was IPT superior. Limitations of this initial controlled trial of IPT include a nonsequential recruitment strategy and overlap in the administration of the two therapies. It is recommended that future studies of IPT for SAD include a more carefully defined control therapy condition, different therapists administering each therapy, a larger sample, and a more rigorous strategy for long-term follow-up assessments.

Original languageEnglish
Pages (from-to)542-553
Number of pages12
JournalDepression and Anxiety
Issue number6
StatePublished - 4 Jul 2008
Externally publishedYes


  • Anxiety
  • Clinical trial
  • Interpersonal
  • Psychotherapy
  • Social phobia

ASJC Scopus subject areas

  • Clinical Psychology
  • Psychiatry and Mental health


Dive into the research topics of 'A randomized trial of interpersonal therapy versus supportive therapy for social anxiety disorder'. Together they form a unique fingerprint.

Cite this