Validation of the World Health Organization Disability Assessment Schedule in people with severe mental disorders in rural Ethiopia

Kassahun Habtamu, Atalay Alem, Girmay Medhin, Abebaw Fekadu, Michael Dewey, Martin Prince, Charlotte Hanlon

The World Health Organization Disability Assessment Schedule (WHODAS-2.0) has been adapted and validated in several cultures, but data on performance in the African context are lacking. The aim of the study was to evaluate the validity and psychometric properties of the WHODAS-2.0 among people with severe mental disorders (SMD) and their caregivers in a rural African setting. The content validity of the 36 item WHODAS was assessed using free listing and pile sorting in 36 community members. Cognitive interviewing was conducted with 20 people with SMD and 20 caregivers to assess comprehensibility. Convergent validity and sensitivity to change were evaluated in a facility-based cohort study of new or acutely relapsed cases of people with SMD (n = 150) and their caregivers (n = 150) consecutively recruited from a psychiatric clinic. A repeat assessment was conducted in a sub-sample (n = 84) after 6 weeks. Confirmatory factor analysis was used to evaluate construct validity in people with SMD (n = 250) and their caregivers (n = 250). Internal consistency of the items of the overall scale and each domain ranged from very good (alpha = 0.82) to excellent (alpha = 0.98). Scores on the WHODAS-2.0 correlated highly with a locally developed measure of functioning (r = 0.88) and moderately with clinical symptom severity (r = 0.52). The WHODAS- 2.0 was sensitive to treatment changes (effect size = 0.50). As hypothesized, the six sub-scales loaded highly onto the general disability factor and each item loaded significantly onto their respective domains. The factor loadings of each item in the one factor model of the brief version of WHODAS (12 item) were also high. For both 12- and 36-item scales the goodness of fit indices, were close to, but outside of, recommended ranges. The caregiver data of both the 36 and 12 item versions had similar psychometric properties, but higher mean values and better responsiveness to change. Our study showed that both the 12 and 36 item versions of the WHODAS 2.0 have acceptable validity and psychometric properties and can be used as a cross-cultural measure; however, careful and rigorous adaptation is required for rural African settings.