Differential Item Functioning on the Patient Health Questionnaire 8 by Disease Subtype, Language, Sex, and Age Among People With Systemic Sclerosis: A Scleroderma Patient-Centered Intervention Network Cohort Study.

Abstract

Somatic items used in depression assessments can potentially overlap with symptoms related to physical illness, including systemic sclerosis (SSc). No studies have looked at whether somatic depression items may be influenced by diffuse versus limited SSc disease subtypes, which are associated with varying degrees of symptom presentation. The objective of this study was to evaluate differential item functioning (DIF) in items of the 8-item Patient Health Questionnaire (PHQ-8) across SSc subtypes. We also assessed the PHQ-8 for DIF across language (English and French), sex, and age. Participants enrolled in the Scleroderma Patient-Centered Intervention Network Cohort who completed the PHQ-8 at enrollment between April 2014 and October 2020 were included. Confirmatory factor analysis (CFA) was used to evaluate the unidimensional structure of the PHQ-8, and DIF analyses based on SSc subtype, language, sex, and age were conducted using Multiple Indicators Multiple Causes models. In total, 2,191 participants were included. CFA with several covarying error terms supported a one-factor structure for the PHQ-8 (Tucker-Lewis Index = 0.99, Comparative Fit Index = 0.98, Root Mean Square Error of Approximation = 0.08). We did not identify statistically significant DIF based on SSc subtype. Statistically significant DIF was found in one item for language, one item for sex, and two items for age. However, the effect of DIF on overall PHQ-8 scores was negligeable in all cases. We did not find evidence that the PHQ-8 performs differently across SSc subtypes, language of administration, sex, and age groups.