Older Canadians who go without: predictors of unmet need for activities of daily living

Paul Spin, Dalhousie University
Janice Keefe, Mount Saint Vincent University
Samuel Vézina, Mount Saint Vincent University

Prevalence of unmet and undermet needs for personal assistance with activities of daily living (ADLs) and instrumental ADLs among older Canadians with disabilities is examined. Characteristics that may differ between persons with met and unmet needs are also investigated. Using the 2006 Participation and Activity Limitation Survey (PALS) data, eight ADLs and IADLs are examined. The overall prevalence of the population with unmet or undermet needs is calculated as well as the prevalence of met, undermet and unmet need for each activity. Binary and multinomial logistic regressions are performed, respectively, for the two models. The first one estimates the probability of having at least one unmet or undermet need compared to having all needs met. In the second model, we distinguish individuals with (1) at least one undermet need and no unmet needs; (2) at least one undermet need and no unmet needs; (3) at least one unmet need and one undermet need. The independent variables in the model include: age, sex, country of birth, living arrangements, number of surviving children, schooling level, region of residence, number of activities for which help is needed, and disability level. Bootstrap weights are used to weight the results since the PALS uses a complex survey design that increases the variance of the parameter estimates. Results show that 503,000 older Canadians with a disability, with needs, and living in private households had at least one unmet or undermet need in 2006. Among them, 47.3% don’t receive all the help for 2 or more needs (activity). Highest prevalences of unmet/undermet needs are observed for heavy household chores and moving about. The number of needs, the disability severity, and to a lesser extent education, significantly increases the probability of having unmet/undermet need; age significantly decreases this probability.

  See paper

Presented in Poster Session 1