Preliminary data on micro errors as a sensitive measure of everyday functioning in older adults

Abstract Background Mild difficulties with daily tasks are associated with cognitive abilities and future cognitive decline but are difficult to measure objectively. Performance‐based tests with high task demands (e.g., multitasking, prospective memory) have been fruitful in capturing mild functional difficulties, however measuring certain aspects like overt errors or completion time often shows floor effects and thereby fail to discriminate between individuals with subtle functional decline. We propose a different approach that involves coding micro‐error defined as inefficient but not overtly erroneous actions (e.g., mis‐reaching toward the wrong object) during relatively simple, performance‐based tests, such as the Naturalistic Action Test (NAT). We present preliminary findings on the validity of micro‐errors as sensitive measures of cognition in participants with healthy cognition, mild cognitive impairment (MCI), and mild dementia. Method 27 older adults (Healthy, MCI, Dementia) completed the NAT, which requires participants to make a breakfast and a lunch using objects presented on a table. Standard cognitive tests (memory, language, etc.) were administered. NAT scores reflecting micro‐errors, total steps accomplished, and completion time were examined across participant groups and included in correlation analyses. Result Worse performance on the NAT was associated with disease severity (micro errors Mean (SD) ‐ Healthy: M = 3.05 (.79); MCI: M = 8.80 (1.28); Dem: M = 9.33 (4.84); completion time(sec) ‐ Healthy: M = 211.84 (10.70); MCI: M = 352.00 (110.50); Dem: M = 363.00 (36.67); total steps accomplished‐ Healthy: M = 32.79 (0.12); MCI: M = 31.80 (0.73); Dem: M = 30.67 (2.33)). Among those with healthy cognition and MCI, greater micro errors were associated with lower accomplishment scores (r = ‐.422, p = .040) and longer completion times (r = .388, p = .061). Micro‐errors, but not accomplishment scores or completion time, were significantly associated with cognitive tests, who made more micro‐errors obtained a lower overall cognitive composite (r = ‐.434, p = .034) and a lower memory score (r = ‐.438, p = .032). Conclusion Subtle errors and inefficient actions (micro‐errors) during highly familiar everyday tasks are valid indicators of cognitive difficulties in older adults without dementia. Micro‐errors may be more sensitive than time to completion and task accomplishment and may be obtained from 5‐10 minutes of task performance. Micro‐errors have potential to identify risk of decline and as sensitive functional outcomes in clinical trials.