Subjective cognitive complaints of older adults at the population level: an item response theory analysis

Subjective cognitive complaints (SCCs) are increasingly a focus in studies on prodromal Alzheimer disease (AD) and risk for dementia. Little is known about the optimal approach to measure SCCs. We used item response theory (IRT) to examine the characteristics of 24 SCC items in a sample of 3495 older adults pooled from 4 community-based studies. We investigated the potential advantages of IRT scoring over conventional scoring on the basis of participants' item response patterns. Items most likely endorsed by individuals low in SCC severity relate to word retrieval and general subjective memory decline. Items likely endorsed only by individuals high in SCC severity relate to nonepisodic memory changes, such as decline in comprehension, judgment and executive functions, praxis and procedural memory, and social behavior changes. Above and beyond conventional total score, IRT scoring of SCCs was associated with performance on objective cognitive tests, and was associated with cognitive test performance among participants endorsing only 1 SCC item. Thus, IRT scoring captures additional information beyond a simple sum of SCC symptoms. Modern psychometric approaches including IRT may be useful in developing: (1) brief community screening questionnaires; and (2) more sensitive measures of very subtle subjective decline for use in prodromal Alzheimer disease research.