Psychometric Properties of the Generalized Anxiety Disorder Scale (GAD‐7) in Older Adults With Advanced Cancer

ABSTRACT Objectives Over half of new cancer diagnoses occur in patients aged 65 or older, with up to 40% experiencing anxiety. The American Society of Clinical Oncology recommends using the Generalized Anxiety Disorder Scale (GAD‐7) for anxiety screening, but the GAD‐7 psychometric properties in this population are unknown. This study examined the GAD‐7's reliability, validity, and item parameters, comparing its utility with the GAD‐2 in older adults with cancer. Methods This cross‐sectional secondary analysis of a nationwide multi‐site two‐arm cluster randomized trial in older adults (≥ 70) with advanced cancer. The GAD‐7 was administered at baseline. Properties were evaluated with Cronbach's α , Pearson correlation coefficients, and a 2‐parameter logistic model. Logistic regression models compared the GAD‐2 and GAD‐7. Results The sample included 718 participants (Mean age = 77, SD = 5) with mild anxiety ( M = 3.74, SD = 4.74). Internal consistency was strong (Cronbach's alpha = 0.89) and item‐total correlations ranged 0.53 to 0.78. Item 2 (Not being able to stop or control worrying) was the most discriminating and item 5 (Being so restless that it is hard to sit still) was least discriminating. Area Under the Curve (AUC) analyses demonstrated the GAD‐2 had a 0.93–0.96 AUC. Conclusions Establishing the psychometric properties of anxiety screening measures is crucial in the older adults with cancer to maximize referral efficiency and accuracy. This study indicates that the GAD‐7 is reliable and valid for older adults with cancer. Analyses suggest the GAD‐2 may be as sufficient as the GAD‐7 in identifying anxiety in older adults with cancer, thereby reducing assessment burden.