To develop and evaluate the psychometric properties of a modified Naturalistic Action Test (m-NAT) for Korean patients with impaired cognition. The NAT was originally designed to assess everyday action impairment associated with higher cortical dysfunction.

We developed the m-NAT by adapting the NAT for the Korean cultural background. The m-NAT was modeled as closely as possible on the original version in terms of rules and scoring. Thirty patients receiving neurorehabilitation (twenty-three stroke patients, five traumatic brain injury patients, and two dementia patients) and twenty healthy matched controls were included. Inter-rater reliability was assessed between two raters. Validity was evaluated by comparing the m-NAT score with various measures of attention, executive functions, and daily life.

Performance on the m-NAT in terms of the total score was significantly different between patients and controls (p<0.01). Patients made significantly more total errors than controls (p<0.01). Omissions error was the most frequent type of error in patient group. Intraclass correlation coefficients for total m-NAT score was 0.95 (95% confidence interval [CI], 0.92 to 0.97; p<0.001); total error was 0.91 (95% CI, 0.89 to 0.92; p<0.001). Total m-NAT score showed moderate to strong correlations with Stroop test interference score & index, Trail Making Test parts A and B, Sustained Attention to Response Task commission error, Functional Independence Measure, Korean instrumental activities of daily living, Korean version of the Cognitive Failures Questionnaire, and Executive Behavior Scale (p<0.05).

The m-NAT showed very good inter-rater reliability and adequate validity. The m-NAT adjusted to Korean cultural background can be useful in performance-based assessment of naturalistic action for clinical and research purposes.