Psychometric properties of the 7-item Generalized Anxiety Disorder scale in antepartum women from Ghana and Côte d'Ivoire

BackgroundTo explore the psychometric properties of the 7-item Generalized Anxiety Disorder scale (GAD-7) in West African pregnant women.MethodsIn a cross-sectional study, the GAD-7, the PHQ-9 (depression module of the Patient Health Questionnaire) and the WHO-DAS II (12-item World Health Organization Disability Scale) were administered to n=647 Ivorian and n=395 Ghanaian antepartum women with low-risk pregnancies (third trimester; gestational age range: 28-40 weeks) in 2010-2011. Statistical analyses combined classical test theory (CTT) and item response theory (IRT).ResultsCronbach׳s α was close to acceptable (α=.69 in Côte d'Ivoire; α=.67 in Ghana). Construct validity was demonstrated by correlations between the GAD-7 and both the PHQ-9 and the WHO-DAS II. Results regarding factorial validity were ambiguous: confirmatory and exploratory factor analyses indicated that the data fit the unidimensionality assumption; however, the explained variance was low. IRT methods suggested excellent item infit and outfit measures. Yet, residual principal component analyses showed a lack of unidimensionality. Rasch analyses revealed that the response options were not used as intended. Item-person-fit was not optimal since the items did not discriminate well on the range of the latent construct where most of the women drawn from the general population were located.LimitationsBoth samples were somewhat homogenous regarding gestational age and urban residence; test-retest reliability was not investigated.ConclusionThe GAD-7 met the requirements of CTT. However, applying the stricter criteria of IRT suggested areas for potential improvements. Cross-cultural adaption and a dichotomized response format might ameliorate the GAD-7 for the use in sub-Saharan countries.