Objective The IDEA study cognitive screening tool was developed for identification of dementia in low-resource settings and populations with low levels of formal education. The aim of this study was to formally validate the tool as a screen for dementia and also delirium in hospitalised older adults and outpatient attendees alongside a lower prevalence community setting to assess usefulness in primary care.
Method The screening instrument was administered to 100 consecutive older adults admitted to medical wards of a Government hospital in Tanzania alongside 108 consecutive geriatric medicine attendees. A rural community sample of 530 older adults were also screened. All inpatients were subsequently fully assessed by a research doctor who was blind to the screening tool results. In the outpatient and community samples all those of high and intermediate probability and a random sample of low probability individuals were assessed by the same method. Assessment included psychiatric and cognitive examination, informant history and neurological examination. Dementia and delirium were diagnosed according to DSM-IV and ICD-10 criteria.
Results In hospital inpatients the area under the ROC curve (AUROC) for the IDEA cognitive screen was 0.903 (95% CI 0.84–0.965) for dementia and delirium combined. In the outpatient and rural community samples the AUROC values for dementia were 0.931 (95% CI 0.865–0.996) and 0.854 (95% CI 0.793–0.915) respectively.
Conclusion This six-item brief cognitive screening instrument performed well in this low-literacy population and should prove useful in screening for dementia in varied healthcare settings in Tanzania. Further testing in other low-resource settings is required.