Convergent and concurrent validity of a report- versus performance-based evaluation of everyday functioning in the diagnosis of cognitive disorders in a geriatric population

Background:: Several methods have been developed to evaluate activities of daily living (ADLs) in mild cognitive impairment (MCI) and mild dementia. This study evaluated the convergent and concurrent validity between (1) two report-based methods (the advanced (a)- and instrumental (i)-ADL tools) and (2) a performance-based method (the Naturalistic Action Test (NAT)) to check if their ability to differentiate between cognitively healthy comparisons (HCs), persons with MCI, and persons with mild Alzheimer's disease (AD) are comparable to each other. Method:: This was a cross-sectional study, undertaken in a geriatric day hospital. The participants comprised community-dwelling HCs (n = 21, median age 78.0 years, 61.9% female), MCI (n = 20, median age 79.5 years, 55.0% female), and AD (n = 20, median age 80.0 years, 85.0% female) adults. A diagnostic procedure for neurocognitive disorders was employed. In addition, the a- and i-ADL tools and the NAT were administered separately by blinded raters. Results:: The NAT and both the a- and i-ADL tools showed significant differences between HCs, MCI, and AD participants. Convergent validity showed moderate to strong significant correlations between the NAT, and a- and i-ADL tools (range −0.583 to −0.663; p < 0.01). Concurrent validity showed that the NAT (AUC 0.809–1.000) and the a- and i-ADL tools (AUC 0.739–0.964) presented comparable discriminatory accuracy (p = 0.0588). Conclusions:: In contrast to prior studies comparing report-based and performance-based methods of assessing ADL, this study indicates that the NAT and the a- and i-ADL tools have strong convergent and concurrent validity, and appear to have similar discriminatory power in differentiating between HCs, MCI, and AD.