The Psychometric properties of the generalized anxiety disorder-7 scale in individuals with anxiety and mood disorder

Abstract

Background: The Generalized Anxiety Disorder-7 (GAD-7) scale was originally developed and is commonly used as a screening instrument for generalized anxiety disorder in primary care [1], as well as in students [2] and general population [3]. Original authors of the scale recommend using an optimal cut-off point of ≥10 for identifying cases of generalized anxiety disorder [1], but novel research shows the GAD-7 has acceptable properties for identifying generalized anxiety disorder at cut-off scores 7-10 in different populations [4]. Despite the wide usage of the GAD-7, little is known about its psychometric properties in psychiatric populations, especially in individuals with anxiety and mood disorders. The knowledge about the multidimensional structure of the GAD-7 is also inconsistent and debatable [5]. The current study aimed to investigate the psychometric properties of the Lithuanian version of the GAD-7 in individuals with anxiety and mood disorders. Methods: Two hundred and seventy-six individuals (22.5% men and 77.5% women, mean age 39.7±12.1 years) attending the Stress-related Disorders Department of Lithuanian University of Health Sciences Neuroscience Institute Palanga Clinic, Lithuania, completed GAD-7, The Mini-International Neuropsychiatric Interview, as well as an objective measure of generalized anxiety disorder administered by trained clinical psychologists (Hamilton Anxiety Rating Scale [HAM-A]). Sociodemographic and clinical data were also collected. Descriptive statistics, internal consistency (Cronbach’s alpha), and confirmatory factor analysis were used to analyze reliability and factor structure of the questionnaire. Screening parameters including sensitivity, specificity, and predictive values were calculated for the GAD-7 scores to define the best cut-off point when screening for possible generalized anxiety disorder. Results: In individuals with anxiety and mood disorders, the GAD-7 (α¼0.800) as well as emerged factors (α¼0.752 and α¼0.582) demonstrated adequate internal consistency. The GAD-7 showed a moderate correlation with clinician rated scale of generalized anxiety disorder, HAM-A (r¼0.331, p<0.01), demonstrating adequate convergent validity. ROC analysis indicated index of discriminating ability of test AUC¼0.633 (p<0.001) and suggested a cut-off value of 7. Using a cut-off of 7, the GAD-7 showed good psychometric characteristics identifying current generalized anxiety disorder (sensitivity ¼ 72.4%, specificity ¼ 47.8%, positive predictive value ¼ 77.6%). The confirmatory factor analysis suggested a two-factor structure (one factor related to cognitive-emotional items and a second factor related to somatic symptoms) and indicated adequate fit to the model (comparative fit index ¼ 0.992, normed fit index ¼ 0.973, root mean square error of approximation ¼ 0.038). Conclusions: The findings support the good psychometric properties of GAD-7 as a screening tool for a current generalized anxiety disorder in individuals with anxiety and mood disorders in Lithuania using a lower cut-off than previously recommended for primary care samples. The findings also suggest two factor structure for GAD-7, measuring factors related to cognitive-emotional symptoms and somatic symptoms. Nonetheless, further psychometric studies regarding GAD-7 are recommended in more diverse psychiatric samples to address the inconsistencies in terms of its multidimensional structure and cut-off values in order to ensure more accurate diagnostic procedures.