Background The Rapid Assessment of Disability (RAD) questionnaire measures the magnitude and impact of disability and aims to inform the design of disability inclusive development programs. This paper reports the psychometric evaluation of the RAD. Methods The initial version of the RAD comprised five sections: 1) demographics, 2) functioning, 3) rights awareness, 4) well-being, and 5) access to the community. Item functioning and construct validity were assessed in a population-based study in Bangladesh. Data were analysed using descriptive statistics (sections 2 and 5) and Rasch modelling (sections 3 and 4). A subsequent case–control study in Fiji tested the refined questionnaire in a cross-cultural setting and assessed the sensitivity and specificity of the RAD section 2 to identify people with disability. Results 2,057 adults took part in the study (1,855 in Bangladesh and 202 in Fiji). The prevalence of disability estimated using RAD section 2 in Bangladesh was 10.5% (95% CI 8.8-12.2), with satisfactory sensitivity and specificity (62.4% and 81.2%, respectively). Section 3 exhibited multidimensionality and poor differentiation between levels of rights awareness in both Bangladesh (person separation index [PSI]?=?0.71) and Fiji (PSI?=?0.0), and was unable to distinguish between people with and without disability (Bangladesh p?=?0.786, Fiji p?=?0.403). This section was subsequently removed from the questionnaire pending re-development. Section 4 had good ability to differentiate between levels of well-being (PSI?=?0.82). In both countries, people with disability had significantly worse well-being scores than people without disability (p? Conclusions Filed-testing in Bangladesh and Fiji confirmed the psychometric robustness of functioning, well-being, and community access sections of the RAD. Information from the questionnaire can be used to inform and evaluate disability inclusive development programs.
展开▼
