Reliability and construct validity of three health-related self-report scales in HIV-positive adults in rural Rwanda

Depression, low health-related quality of life, and low perceived social support have been shown to predict poor health outcomes, including HIV-related outcomes. Mental health morbidity and HIV are important public health concerns in Rwanda, where approximately half of the current population is estimated to have survived the genocide and 3% is living with HIV. We examined the reliability and construct validity of the Hopkins Symptom Checklist-15 (HSCL-15), the Medical Outcomes Study HIV Health Survey (MOS-HIV), and the Duke/UNC Functional Social Support Questionnaire (DUFSSQ), which were used to assess depression, health-related quality of life, and perceived social support, respectively, among HIV-infected adults in rural Rwanda. We also studied whether scale reliability differed by gender, literacy status, or antiretroviral therapy (ART) delivery strategy. The Kinyarwanda versions of the HSCL-15, MOS-HIV, and DUFSSQ performed well in the study population. Reliability was favorable (Cronbach's alpha coefficients >= 0.75 or above) for the scales overall and across subgroups of gender, literacy, and mode of ART delivery. The scales also demonstrated good convergent, discriminant, and known-group validity.