Reproducibility of the Increased Guttman Scalability with Age of ADL items

Items indicating the need for help in the activities of daily living (bathing, dressing, transferring, toileting and eating) are increasingly Guttman scalable with age, meaning they become more perfectly hierarchical with age. Here, results for several national surveys including the core National Health Interview Survey, the Survey of Income and Program Participation, and the National Long Term Care Survey are presented. Generally, these surveys show a highly reproducible pattern of greater Guttman scalability with age. In two surveys, at ages 18-34, the Guttman coefficient of scalability is approximately .75, but increases to almost .95 at ages 85 and older (perfect scalability is CS=1.0). The lower scalability in younger age groups means that disability measurement using ADL alone is substantially biased by age, which complicates the evaluation of need for long term services and supports in younger and older populations. However, a measure using both IADL and ADL items is not biased by age. Measurement of disability for purposes of providing long term services and supports to adults of all ages living in the community should be based on a combination of IADL and ADL items, but policy and practice still focuses on the number of ADL items a person needs help with (such as the CLASS provision under PPACA). Additional research is needed on the predictive validity of measures combining IADL and ADL items. Nevertheless, such measures have greater content validity for persons living in the community, for reasons Lawton and Brody outlined more than three decades ago.