This paper examines the effect of cohort heterogeneity upon the potential needs for assistance in ADL and IADL of an aging population. Variables that are important predictors of the differentiating pattern of perceived needs in the older population are subject to on-going changes as a result of cohort differentiation and cohort succession in the older ages. For those reasons, a forecasting methodology involving three phases has been developed: an identification phase of the key explanatory variables, a demographic-projection phase of socio-demographic variables identified as key variables and a phase in which risk-factor probabilities are applied to the demographic-projection output to determine the overall impact of the population change. The use of logistic regression is shown as the best strategy to determine risk-factor probabilities. This method is applied to measure the assistance needed in ADL and IADL by disabled elderly persons in Québec. The results point to the need to consider changes in intra-cohort heterogeneity between new cohorts entering the older population and old ones leaving it.
