The early recognition of dementia in the elderly: A preliminary communication about a longitudinal study using the gms–agecat package (community version)

Abstract A random community sample of 1070 subjects aged over 65 was interviewed by trained non‐medical interviewers using the Geriatric Mental State, community version (GMSA). A sub‐sample of 126 subjects was selected so as to contain possible early cases of dementia, pseudo‐dementia, and normal subjects; and re‐interviewed, a mean 1 year and 23 weeks later, by a group of psychiatrists in training. The computer diagnosis AGECAT, based on GMSA applied by non‐medical raters, had predicted at initial interview, nine out of twelve cases of dementia at follow‐up and five out of nine borderline cases. An Organic/Depression Index may prove useful in predicting which of those cases with early organic levels will eventually develop dementia, depression or recover.