Factor structure, measurement invariance, and concurrent validity of the Patient Health Questionnaire-9 and the Generalized Anxiety Disorder scale-7 in a Norwegian psychiatric outpatient sample

Abstract

Objective The aim of this study was to test factor structure, measurement invariance, and concurrent validity of the nine item Patient Health Questionnaire-9 (PHQ-9) and the seven item Generalized Anxiety Disorder scale-7 (GAD-7) in a heterogeneous outpatient sample.

Method Outpatients completed the PHQ-9, GAD-7, and the Working Social Adjustment Scale (WSAS) before starting treatment. Study design was cross-sectional, with convenience sampling. The total sample consisted of 831 participants (61% women).

Results Both PHQ-9 and GAD-7 demonstrated better fit statistics with two-factor and bifactor solutions consisting of a cognitive and somatic factor. Omega hierarchical was .78 for PHQ-9 and .81 for GAD-7. Both instruments achieved scalar invariance across gender, diagnosis, and comorbidity. However, the somatic factors demonstrated poor discriminant validity. These factors are not well separatable and risks being too similar if used together. The general factors of both instruments were most associated with functional impairment, although PHQ-9 demonstrated a stronger association with WSAS (γ = .74, r2 = .62) than GAD-7 (γ = .54, r2 = .32). Using latent mean difference, women and patients with comorbidity had significantly higher scores of both depression and anxiety.

Conclusion This study shows that the PHQ-9 and GAD-7 may be used as one-dimensional instruments in clinical settings. Tests for measurement invariance supported that both measures are understood and interpreted comparably across gender and diagnostic subgroups.