Measurement Invariance by Gender, Ethnicity, and International Student Status with the PHQ-9 in a University Setting

Background: Measurement invariance on mental health outcome measures is vital to establish validity for using measures in clinical practice across varying demographics. Objective: This study analyzes the measurement invariance of the Patient Health Questionnaire-9 (PHQ-9) to support its use in practice across various groups. Method: In this study we analyzed the validity of the PHQ-9 in a large college counseling sample across gender (cis-gender women and men), ethnicity (White and Hispanic / Latina/o/x), and international student status (international and domestic). The sample consisted of 4,363 participants; the gender spread of the sample was 60.9% women, 35.9% men, 11.7% other gendered or multi-gendered, 0.6% questioning or unsure, 0.5% transgendered, and 0.3% who preferred not to answer. The racial/ethnic spread of the sample was 71.2% White, 13.7% Hispanic and/or Latina/o/x, 3.6% Asian or Asian American, 2.4% African American / Black, 1.1% Pacific Islander, 0.7% Native American, 3.6% multi-ethnic, and 2.7% other. Results: Analyses demonstrated configural, metric, scalar, and residual measurement invariance as indicated by model fit indices across all groups tested using both a unidimensional and second-order model, with relevant factors in each model also demonstrating composite reliability. Limitations: The conclusions focus on the sample of a university counseling setting, as specified. Conclusions: These results support the use of the PHQ-9 to assess levels of depression across the different populations studied and an overall score.