ABSTRACT
BACKGROUND: Many screening instruments for the diagnosis of dementia are not education and culture fair. In addition, despite good sensitivity and specificity, they result in unacceptable levels of false positives when used in the community. This study aimed to develop appropriate instruments for populations with low literacy and to consider strategies to reduce the false positive rates in low prevalence settings. METHODS: Activities of daily living, which are not influenced by education and culture, were used to develop a patient screen and an informant version to identify people with dementia. The instruments were validated in the hospital and in the community setting against the standards of DSM-IV and the education-adjusted 10/66 Dementia Research Group diagnosis of dementia. RESULTS: The instruments were administered to 90 patients attending a hospital and 101 subjects living in the community in Vellore, South India. The psychometric properties and inter-rater reliability of these screening instruments were good. While the sensitivity and specificity of the patient screen and the informant version were good their false positive rates were high in the community setting. However, the false positive rates reduced when these instruments were used in combination. CONCLUSION: The patient screen and the informant version are short culture- and education-fair instruments. They reduce false positive rates, when used in combination in the community.