Abstract
Objectives: This study investigated the validity and reliability of subtle errors and slowing in simple everyday tasks (Naturalistic Action Test [NAT]) to assess mild functional difficulties in older adults with mild cognitive impairmen (MCI). Method: Older adults (N = 111, MAge= 73.45; SD= 6.53) classified as having healthy cognition (HC) or MCI completed neuropsychological testing and two NAT tasks (breakfast and lunch) twice, separated by 4-6 wk. NATs were scored for subtle, inefficient actions (i.e., micro-errors) and the average time (in sec) to complete a task step. Results: Participants with MCI made significantly more micro-errors [F (1, 109) = 8.78, p = .004, partial η(2) = 0.07] and had a significantly longer average time per step [F (1, 109) = 13.98, p < .001, partial η(2) = 0.11] than participants with HC. Micro-errors correlated with tests of episodic memory (r = -0.237, p = .012) and executive functioning (r = -0.201, p = .035), whereas average time per step correlated only with episodic memory (r = -0.300, p = .0001). Test-retest reliability was good for average time per step (ICC = .872, p < .001) and moderate for micro-errors (ICC = .675, p < .001). Conclusions: Measures of inefficient actions and performance time in familiar everyday tasks demonstrated strong to adequate construct and concurrent validity, as well as test-retest- and inter-rater reliability, supporting their use for the quantifying mild functional difficulties. Future studies should explore this scoring approach to develop early markers of functional disability or dementia risk.