Representative samples of people aged 65 or over on October 31, 1990 were drawn in 39 urban centres and nearby rural areas in the ten Canadian provinces. Geographic sampling areas were defined using postal codes; we estimate that these areas contain roughly 66% of the elderly Canadian population. The final sample included 9,008 participants from the community and 1,255 from long-term care institutions.

Community Sample:

Except in Ontario, the community sampling frame used the computerized records of the provincial health insurance plans (which provide universal access to free medical care in Canada). A random sample, stratified by age, was drawn in each sampling area. Optimal allocation was used to guide the relative sample sizes in each age stratum; the sampling ratio for those aged 75-84 was twice that of those aged 65-74; those aged 85+ were sampled at a ratio 2.5 times that of the 65-74 year-olds.

The institutional sampling frame was developed from a comprehensive list of nursing homes, homes for the aged and other group dwellings such as convents. From this list we first drew a random sample of institutions, stratified by size. Within the larger institutions selected, we then drew simple random samples of people 65 years of age or over; in the small institutions everyone was included.

Sample weights were calculated for each person in the study to correct for the over-sampling of the very old, and for the fact that equal samples were drawn in each study region despite their different population sizes. Sample weights were also calculated at CSHA-2 and CSHA-3 to adjust for losses to follow-up.

The caregiver studies selected sub-groups of CSHA participants according to their cognitive status, and then interviewed their caregivers. The CSHA-1 caregiver sample included 1,048 study participants diagnosed with dementia; a comparison group included 638 caregivers of people who were cognitively normal (this group included some who were physically frail). In each group we interviewed the principal caregiver, whether a family member or paid caregiver.

At CSHA-2 we re-interviewed 1,057 of the informal caregivers who had been included in the CSHA-1 caregiver study. The CSHA-3 caregiver study compared 220 caregivers of people with incident dementia to 330 caregivers of people with CIND, and to a comparison group of 300 caregivers of people without cognitive problems, but who had functional limitations.

The table shows participation rates in the three waves of the study. Those not participating at CSHA-2 and 3 included people who declined to participate, or who were lost to follow-up (e.g., moved out of the study area), or who were too sick to participate.

* The “not contactable” group refers to people on the sampling lists whom, after multiple attempts, we could not locate. It seems likely that many of these people had died, or had moved away from the study areas, or would have been ineligible for other reasons.

Further information on study sampling:

McDowell I, et al. Study sampling in the Canadian Study of Health and Aging. International Psychogeriatrics 2001;13(Suppl 1):19-28