Projection of young-old and old-old with functional disability: does accounting for the changing educational composition of the elderly population make a difference?

Abstract

This study compares projections, up to year 2040, of young-old (aged 60-79) and old-old (aged 80+) with functional disability in Singapore with and without accounting for the changing educational composition of the Singaporean elderly. Two multi-state population models, with and without accounting for educational composition respectively, were developed, parameterized with age-gender-(education)-specific transition probabilities (between active, functional disability and death states) estimated from two waves (2009 and 2011) of a nationally representative survey of community-dwelling Singaporeans aged ≥ 60 years (N=4,990). Probabilistic sensitivity analysis with the bootstrap method was used to obtain the 95% confidence interval of the transition probabilities. Not accounting for educational composition overestimated the young-old with functional disability by 65 percent and underestimated the old-old by 20 percent in 2040. Accounting for educational composition, the proportion of old-old with functional disability increased from 40.8 percent in 2000 to 64.4 percent by 2040; not accounting for educational composition, the proportion in 2040 was 49.4 percent. Since the health profiles, and hence care needs, of the old-old differ from those of the young-old, health care service utilization and expenditure and the demand for formal and informal caregiving will be affected, impacting health and long-term care policy.

Fig 1. Dynamic Multi-state Population Model Not Accounting for Education.

Population model with fertility rates, mortality rates and disability transition probabilities that are not education-specific.

Fig 2. Dynamic Multi-state Population Model Accounting for Education.

Population model with education-specific fertility rates, mortality rates and disability transition probabilities.

Fig 3. Transition Probabilities by Age, Gender and Education.

(A) Transition probabilities from active to functional disability. (B) Transition probabilities from functional disability to active. (C) Transition probabilities from active to death. (D) Transition probabilities from functional disability to death.

Fig 4. Young-old (aged 60–79).

(A) Projected number of young-old with functional disability. (B) Distribution of young-old with functional disability at 2040. (C) Difference between not accounting and accounting for educational composition at 2040.

Fig 5. Old-old (aged 80+).

(A) Projected number of old-old with functional disability. (B) Distribution of old-old with functional disability at 2040. (C) Difference between not accounting and accounting for educational composition at 2040.