Background: Cultural differences in how the symptoms, causes, consequences, and treatments of dementia are understood and interpreted by South Asian people are a commonly expressed reason for late- or nonuse of mental health and care services. However, systematic collection of information on South Asian perceptions of dementia is hindered by a lack of appropriate instrumentation.
Objectives: To produce a shortened version of the Barts Explanatory Model Inventory for Dementia (BEMI-D) schedule.
Methods: A two stage item reduction approach was employed first using multidimensional scaling categorizing items as core, intermediate, or outlier. Then, item review was undertaken using three criteria: literature importance, clinical face validity, and sub-group prevalence. The analysis followed a nonmetric multidimensional scaling method based on a two-way proximity matrix.
Results: The original BEMI-D had 197 items allocated to four checklists: symptoms, causes, consequences, and treatments. The two stage item reduction approach resulted in the removal of 75 items. These reductions were achieved across all four checklists in relatively equal proportions. There was no evidence of substantive content loss in the revised schedule. The reduced version of the schedule comprises 122 items.
Conclusions: A condensed version of the BEMI-D is more efficient as an assessment schedule that captures the culturally diverse perceptions of memory problems for South Asians offering a balanced trade-off between feasibility of use and content validity.
Keywords: South Asian; dementia; ethnic minorities; explanatory models; multidimensional scaling.
© 2020 John Wiley & Sons Ltd.