Five-year incidence and prediction of dementia and cognitive decline in a population sample of women aged 70-79 at baseline.

Objective. To determine the incidence of dementia and cognitive decline and their predictors in a population sample. Settings. A rural geographically defined population served by a single health centre. Sample. 75% of the 70–74-year-old, 100% of the 75–79-year-old women registered with the practice. Method. Interview the total sample using an augmented version of CAMDEX, including informant interview, and reaction times at baseline and 5 years later. IQ was estimated at baseline only. Results: 365 women were interviewed at baseline, 237 at follow-up. Incidence of dementia of all severities was 3.1% per annum (95% CI, 2.2–4.3). CAMCOG declined by 12.6 points in those who became demented, MMSE by 3 points. Observed, self-reported and informant-reported cognitive difficulty at baseline all predicted later incidence of dementia. The additional use of estimated IQ and reaction times did not increase the efficiency of either cross-sectional detection of dementia or prediction of dementia at 5 years. Conclusions. The incidence of dementia was in the range reported by other studies; slight cognitive decline was noted in the total population, increasing with age and marked in incident dementia. Of clinical importance was the finding that both the individuals and the informants of those who subsequently demented were aware of decline 5 years earlier. Simple and relatively easily administered cognitive tests were found to be the most efficient method of detection of dementia. © 1997 John Wiley & Sons, Ltd.