Screening for depression in Andean Latin America: Factor structure and reliability of the CES-D short form and the PHQ-8 among Bolivian public hospital patients

Abstract

Objective We assessed the validity and reliability of the Center for Epidemiologic Studies-Depression scale (CES-D) short form and the Patient Health Questionnaire-8 (PHQ-8) (two measures of depressive symptoms) among urban, low-income patients from a heavily indigenous area of Andean Latin America. This is a patient population that is in many ways culturally distinct from the populations that have been included in previous Spanish language validation studies of these instruments. Methods We administered the CES-D Yale version short form and the PHQ-8 to 107 outpatients at public hospitals in La Paz and El Alto, Bolivia. We conducted exploratory factor analysis, checked internal reliability, and examined concordance between the two measures in identifying patients with clinically significant depressive symptoms. Results Internal reliability was high for both scales: PHQ-8 (Cronbach's α = 0.808) and CES-D (α = 0.825). A one-factor solution for the PHQ-8 was maintained after the removal of one poorly loading item. The CES-D showed a two-factor solution, with items for somatic symptoms and interpersonal problems loading on the same factor as negative affect. Using both the cutoff for depression recommended in prior studies and a sample specific cutoff of ≥1 standard deviation above the mean, the two scales demonstrated only moderate agreement (κ = 0.481-0.486). Conclusions The PHQ may be more appropriate for clinical use in the Andean region, given its greater specificity; however, lower than expected agreement between the scales suggests that they also measure different elements of depression. Irregularities with the factor structure of both scales suggest that further research with this population is needed to better understand depressive symptomology and improve screening in the region.