Putative temperament of patients with generalized anxiety disorder: two-years' interval test-retest reliability of a Japanese version of the Generalized Anxious Temperament.

This study examined long-interval test-retest reliability of the Generalized Anxious Temperament (GAT) Japanese version, which is a self-rated questionnaire to measure the temperament of patients with generalized anxiety disorder (GAD). First, we investigated the prevalence of GAT in 214 healthy Japanese subjects (107 females and 107 males). Second, we administered GAT to the 117 (77 females and 40 males) patients with GAD, together with GAF (Global Assessment of Functioning) and HAM-A (Hamilton Anxiety Scale), and readministered after 2 years to examine its reliability. GAT showed good test-retest reliability (Pearson's correlation coefficient = 0.8). Cronbach's alpha of the 26 GAT items was 0.93 that indicated a high level of internal consistency. GAT scores were not influenced by psychological, social or occupational functioning, nor by anxiety symptoms and remained stable over 2 years. GAT scores did not correlate with the age or sex of the control subjects. GAT scores were higher in patients with GAD than they were in control subjects, particularly in females. Twenty-three of 26 GAT items were significantly higher in patients with GAD than in healthy subjects. Stepwise multiple regression analysis revealed that GAT scores in patients with GAD were higher in females than in males (P < 0.01) and in those with early onset of the disorder (P < 0.05). These results suggest that the anxious temperament appears to increase the risk of early onset of GAD, particularly in females. This long interval and large sample research of test-retest reliability on GAT indicated its usefulness for understanding the core features of GAD.