Calibrating ADL-IADL scales to improve measurement accuracy and to extend the disability construct into the preclinical range: a systematic review

Abstract

BackgroundInterest in measuring functional status among nondisabled older adults has increased in recent years. This is, in part, due to the notion that adults identified as 'high risk' for functional decline portray a state that is potentially easier to reverse than overt disability. Assessing relatively healthy older adults with traditional self-report measures (activities of daily living) has proven difficult because these instruments were initially developed for institutionalised older adults. Perhaps less evident, are problems associated with change scores and the potential for 'construct under-representation', which reflects the exclusion of important features of the construct (e.g., disability). Furthermore, establishing a formal hierarchy of functional status tells more than the typical simple summation of functional loss, and may have predictive value to the clinician monitoring older adults: if the sequence task difficulty is accelerated or out of order it may indicate the need for interventions.MethodsThis review identified studies that employed item response theory (IRT) to examine or revise functional status scales. IRT can be used to transform the ordinal nature of functional status scales to interval level data, which serves to increase diagnostic precision and sensitivity to clinical change. Furthermore, IRT can be used to rank items unequivocally along a hierarchy based on difficulty. It should be noted that this review is not concerned with contrasting IRT with more traditional classical test theory methodology.ResultsA systematic search of four databases (PubMed, Embase, CINAHL, and PsychInfo) resulted in the review of 2,192 manuscripts. Of these manuscripts, twelve met our inclusion/exclusion requirements and thus were targeted for further inspection.ConclusionsManuscripts presented in this review appear to summarise gerontology's best efforts to improve construct validity and content validity (i.e., ceiling effects) for scales measuring the early stages of activity restriction in community-dwelling older adults. Several scales in this review were exceptional at reducing ceiling effects, reducing gaps in coverage along the construct, as well as establishing a formal hierarchy of functional decline. These instrument modifications make it plausible to detect minor changes in difficulty for IADL items positioned at the edge of the disability continuum, which can be used to signal the onset of progressive type disability in older adults.

Studying Disability Trends in Aging Populations

“…We are aware of some recent calls for improvement for construct validities of the conventional measurements of the ADL and IADL disability scales due to their ceiling effects (i.e., many people reported no disability in these two scales) (Fieo et al 2011). We also notice that the United Nations (Washington City Group 2009) has recently developed a new measure called the Washington City Group general disability measure.…”

Development of the Central Africa Daily Functioning Interference Scale for Dementia Diagnosis in Older Adults: The EPIDEMCA Study

“…It has been shown that combining IADL and ADL into a single scale enlarges the coverage of daily functioning measurement [27, 29], yet in the present study, only IADL limitations and participation restrictions provided good fit to unidimensionality. Moreover, as the scale was developed for dementia diagnosis, only dressing and bathing limitations have been evaluated, since they constitute the entry into severe dependency [30].…”

Physical disability and cognitive impairment among recipients of long-term care

“…When instruments show gaps between the least difficult variables, it becomes more difficult to separate between the least disabled individuals; conversely, gaps between the most difficult variables make it more difficult to distinguish between the most disabled persons. [14,30,54] This scenario may be related to both the number and the types of variables included. [14] For both groups of elderly, we found relatively large gaps at both ends of the physical disability measures.…”

“…There are no clear guidelines for the recommended size of the gaps between items, although some reports have suggested values between 0.15 and 0.30. [30,47,48] The analyses were performed with SPSS version 21. IBM Corp. Armonk, NY.…”

“…[19] Measures of ADL, mobility, IADL and cognitive impairment have been shown to vary in their ability to detect disability. [13,16,[28][29][30] Thus, important variables may be missing in the instruments that are used ("gaps"), and the aggregated factors may consequently be less precise.…”