Factor structure and longitudinal measurement invariance of PHQ-9 for specialist mental health care patients with persistent major depressive disorder: Exploratory Structural Equation Modelling

Abstract

Background The Patient Health Questionnaire-9 (PHQ-9) is a widely used instrument for measuring levels of depression in patients in clinical practice and academic research; its factor structure has been investigated in various samples, with limited evidence of measurement equivalence/invariance (ME/I) but not in patients with more severe depression of long duration. This study aims to explore the factor structure of the PHQ-9 and the ME/I between treatment groups over time for these patients. Methods 187 secondary care patients with persistent major depressive disorder (PMDD) were recruited to a randomised controlled trial (RCT) with allocation to either a specialist depression team arm or a general mental health arm; their PHQ-9 score was measured at baseline, 3, 6, 9 and 12 months. Exploratory Structural Equational Modelling (ESEM) was performed to examine the factor structure for this specific patient group. ME/I between treatment arm at and across follow-up time were further explored by means of multiple-group ESEM approach using the best-fitted factor structure. Results A two-factor structure was evidenced (somatic and affective factor). This two-factor structure had strong factorial invariance between the treatment groups at and across follow up times. Limitations Participants were largely white British in a RCT with 40% attrition potentially limiting the study's generalisability. Not all two-factor modelling criteria were met at every time-point. Conclusion PHQ-9 has a two-factor structure for PMDD patients, with strong measurement invariance between treatment groups at and across follow-up time, demonstrating its validity for RCTs and prospective longitudinal studies in chronic moderate to severe depression.