Psychometric properties of the GAD-7 and PROMIS-Anxiety-4a among youth with depression and suicidality: Results from the Texas youth depression and suicide research network

Abstract

There is a tremendous need for brief, valid, and free assessments of anxiety in child mental healthcare. The goal of this study was to determine the psychometric properties of two such measures, the GAD-7 and PROMIS-Anxiety-4a, in 1000 children, adolescents, and young adults (8-20 years-old) with depression and/or suicidality. The GAD-7, the PROMIS-Anxiety-4a, and other validated assessments of anxiety, physical functioning, and psychiatric diagnoses were completed. Confirmatory factor analyses showed an acceptable fit for a single factor in both measures via all indices but the RMSEA. They demonstrated measurement invariance across pre-adolescents (8-12 years-old) and adolescents and emerging adults (13-20 years-old), though scalar invariance was not observed for the GAD-7. Both measures showed strong convergent validity, GAD-7: r = 0.68; PROMIS-Anxiety-4a: r = 0.75, divergent validity with a measure of physical function, GAD-7: r = -0.24; PROMIS-Anxiety-4a: r = -0.28, good internal consistency, ω = 0.89 for both, and high test-retest reliability, GAD-7: r = 0.69; PROMIS-Anxiety-4a: r = 0.71. Both measures also showed acceptable sensitivity and specificity in detecting the presence of any anxiety disorder, GAD-7 cut-off score of 10: AUC = 0.75; PROMIS-Anxiety-4a cutoff score of 12: AUC = 0.79. The GAD-7 correlated similarly with the Screen for Child Anxiety Related Disorders total score and generalized anxiety subscale, and also showed similar diagnostic sensitivity and specificity when used to detect the presence of any anxiety disorder vs. generalized anxiety disorder specifically. Results suggest that both of these brief, publicly available instruments are valid and reliable assessments of anxiety among youth in treatment for depression and/or suicidality.