Validation of the Patient Health Questionnaire (PHQ–9) in an African context

Abstract

This research was aimed at validating the PHQ–9 in an African context. This study forms part of the project of Psychosocial Health and Biomarkers in an African context (FORT3, Wissing, 2008). The Patient Health Questionnaire (PHQ–9) is a nine–item depression scale that has the potential of being a dual–purpose instrument to establish the diagnosis of a depressive disorder, as well as the grade of symptom severity (Kroenke, Spitzer & Williams, 2001). The PHQ–9 was administered with criterion related measures to a multicultural convenience sample of 2214 participants from the North West Province of South Africa, including two groups of adolescents (n1 = 1480 and n2 = 559) and an availability sample of adults (n3 = 185). Instruments to determine criterion validity were the General Health Questionnaire (GHQ), designed to detect symptoms of mental disorders; the Mental Health Continuum - Short Form for Adults (MHC–SF) which measures the degree of emotional, social and psychological well–being; and the New General Self–Efficacy Scale (NGSE) designed to measure an individual’s general self–efficacy. Descriptive statistics for the PHQ–9 including its reliability in the various groups is reported. The PHQ–9 manifested a Cronbach Alph are liability index of 0.86. Criterion–related validity was supported by significant correlations between the PHQ–9 and criterion measures. Confirmatory factor analysis for the PHQ–9 yielded a one–factor solution in all groups. The percentage variance explained ranged between 34.71% and 46.62%. Exploratory factor analyses yielded two factors in all groups with the second factor comprised of no more than 2 items and thus interpreted as a minor factor. The construct validity obtained in this research indicates that the PHQ–9 may be a valid measure to identify depression in a South African context. Based on the psychometric properties found in this study, it can be concluded that the PHQ–9 is a valid measure of depression in two of the samples selected for this study. Future studies may further validate this instrument in specific language and cultural groups, and explore the cross–cultural measurement equivalence.