Internal structure and factorial invariance of the Patient Health Questionnaire −9 (PHQ-9) in a large Argentinean sample

TL;DR

This study confirms that the PHQ-9 depression questionnaire works well in Argentina, with consistent results across different groups and formats.

Contribution

The study provides evidence for the PHQ-9's unidimensional structure and measurement invariance in a large Argentine sample.

Findings

The PHQ-9 unidimensional model showed good fit and high correlations between latent factors in two-factor models.

Measurement invariance was confirmed across gender, age groups, and administration formats.

ESEM analyses supported the unidimensional structure without estimation bias.

Abstract

Depression is one of the most prevalent psychological disorders and the leading cause of disability worldwide. The development of effective strategies for the prevention and treatment of depression requires valid and reliable instruments capable of accurately assessing this condition. The present study examined the internal structure and measurement invariance of the Patient Health Questionnaire–9 (PHQ-9) in an Argentine population.

Several measurement models proposed in the international literature were tested in a large Argentine sample (N = 5,950), comprising three independent subsamples: an online sample (n = 4,803), a clinical sample of individuals seeking psychological assistance (n = 98), and a random telephone sample (n = 1,049). PHQ-9 item responses were analyzed using Confirmatory Factor Analysis (CFA) and Exploratory Structural Equation Modeling (ESEM). In addition, measurement invariance was evaluated across gender, age groups, and administration formats.

The unidimensional model demonstrated good fit across most indices. Although two-factor models showed marginal improvements in fit, correlations between latent factors were consistently high (> 0.80). ESEM analyses indicated that these high correlations were not attributable to estimation bias arising from restrictive CFA specifications, further supporting a unidimensional structure. Moreover, the unidimensional model exhibited measurement invariance across gender, age groups, and administration formats.

The findings support the use of the PHQ-9 to derive a global measure (e.g., sum score) of depressive symptom severity in the Argentine population, without systematic bias associated with demographic characteristics or mode of data collection. Further studies are warranted to replicate these results and extend validity evidence.

The online version contains supplementary material available at 10.1186/s12888-026-07820-9.