Understanding age differences in the scalability of ADL items: the role of diagnosis and difficulty

Saturday, November 20 1:30 PM - 3:00 PM
Session Number/Code: 2563
Room: Marriott Salon 1

Mitchell P. LaPlante
Associate Professor, University of California San Francisco

Although the scaling properties of ADL items have been well established for older adults, they have not been established for younger adults, who make up half of the population who need help from another person in one or more ADLs. This study investigates the scalability of five basic ADL items (bathing dressing, transferring, toileting, and eating) among adults ages 18 and above, comparing working age and older adults. Data are from the National Health Interview Survey on Disability, a national survey of people living in households conducted from 1994-1997 (N=126,704). We examine scaling properties of ADL items for which an individual indicates help from another person is needed. Among persons ages 65 and above, the ADL items form a Guttman scale, with high reliability (Cronbach's Alpha =0.88), high reproducibility (coefficient of reproducibility=0.97, high scalability (coefficient of scalability=0.83), and high correlations among items (correlations measured by phi/phi max all >0.7). Among working age adults, all of the scaling properties are weaker, with a coefficient of scalability=0.63 which barely passes the minimum threshold of 0.60. Results are replicated in a follow-up sample. There is a higher than expected prevalence of needing help transferring in the working age population. Multivariate analyses reveal that the reduced scalability of ADL items in the working ages owes mainly to differences in the types of diagnoses (e.g., more back impairments) and a lesser degree of difficulty experienced when perfoming ADLs solo. We conclude that ADL scalability is highest for advanced chronic diseases associated with aging.