The Canadian Study of Health and Aging (CSHA) is a national longitudinal study of the epidemiology of dementia in Canada. The CSHA also described patterns of disability, frailty and healthy aging. The study involved 10,263 people aged 65 or over, sampled representatively from the community and from long-term care institutions, and participants were interviewed and assessed clinically in 1991, 1996 and 2001. Over 200 papers have been published from the study; a brief summary of results is given. The current analyses provide descriptive statistics on patterns and correlates of early cognitive decline in the study cohort. Several salient findings emerged: almost half of this representative sample showed no cognitive decline over a 10-year period. Education showed a strong protective effect against decline, but not after age 80. There was little difference in crude rates of cognitive decline between men and women, but after adjustment for disability and cardiovascular risk factors, men showed a greater risk of decline, up to age 80. Among those who experienced cognitive decline, the extent of decline was predicted mainly by age, rather than by baseline cognitive scores. Among those who developed dementia, baseline tests were better able to predict Alzheimer's disease than vascular dementia. Further analyses designed to elucidate these complex associations are described.