The reliability, validity and factorial structure of the Swahili version of the 7-item generalized anxiety disorder scale (GAD-7) among adults living with HIV from Kilifi, Kenya

Background Generalized Anxiety Disorder (GAD) is under-investigated in people living with HIV/AIDS from sub-Saharan Africa. In part, this is due to paucity of culturally appropriate measures for GAD which are psychometrically robust. This study aimed to evaluate the reliability, factorial structure, and validity of Swahili version of the 7-item GAD questionnaire (GAD-7) among adults living with HIV. Study design Descriptive cross-sectional study. Methods 450 adults receiving comprehensive care from an HIV specialized clinic in Kilifi County, coastal Kenya, were consecutively recruited. Swahili versions of GAD-7, Patient Health Questionnaire (PHQ-9) and a 12-item HIV stigma scale were administered alongside measures of psychosocial and health-related characteristics. Internal consistency, test-retest reliability, factorial structure, convergent validity, and discriminant validity of Swahili GAD-7 were examined using Cronbach's alpha (alpha), intra-class correlation coefficient (ICC), Confirmatory Factor Analysis (CFA), Pearson's correlation, and analysis of covariance (ANCOVA), respectively. Results Internal consistency of Swahili GAD-7 was good, alpha = 0.82 (95% CI 0.78, 0.85). Its test-retest reliability (2 weeks apart) was acceptable, ICC = 0.70 (95% CI 0.55, 0.81). A confirmatory analysis of a one-factor solution indicated an excellent fit to the hypothesized structure (RMSEA = 0.00 [95% confidence interval 0.00, 0.05], CFI = 1.00, TLI = 1.00). Multi-group CFA substantiated factorial invariance for sex and age for the one-factor structure of Swahili GAD-7. Scores of GAD-7, Swahili version, significantly correlated with those of PHQ-9 (r = 0.73; p < 0.001) and the HIV stigma scale (r = 0.36; p < 0.001) suggesting good convergent validity. Statistically significant differences were observed between participants on first-line antiretroviral therapy compared to those on second-line treatment (F [1, 441] = 5.55, p = 0.02) indicative of good discriminant validity of Swahili GAD-7. Conclusion GAD-7 Swahili version retained its original unidimensional latent structure with good psychometric properties among adults living with HIV from Kilifi, Kenya. It can be used to identify symptoms of GAD in similar research settings. However, to confidently identify those in need of mental health treatment or referral services in HIV primary care clinics, more research on the validity of Swahili GAD-7 is needed especially its discriminant validity and diagnostic accuracy at different cut-off scores.