Validity, reliability and diagnostic accuracy of the Thai version of the cognitive assessment scale of stroke patient (T-CASP) in non-aphasic stroke

A total of 66 stroke patients from a neurology clinic completed the Thai Mental State Examination (TMSE), Montreal Cognitive Assessment (MoCA), and T-CASP. All participants scored above the cutoff on the TMSE; therefore, the MoCA was used as the reference instrument. T-CASP scores showed a strong positive correlation with the MoCA (r = 0.71, p < 0.001). Internal consistency of the original T-CASP was acceptable (Cronbach's α = 0.67) and improved after removal of items C2 (Comprehension) and C4 (Drawing a picture series) (Cronbach's α = 0.72). Test-retest reliability was excellent for both the original version (ICC = 0.89) and the modified version excluding items C2 and C4 (ICC = 0.87). Both versions demonstrated high diagnostic performance, with sensitivity of 79.5% and 72.7%, specificity of 95.5% and 99.5%, and area under the ROC curve values of 0.87 and 0.84, respectively. Taken together, these findings suggest that the T-CASP shows promise as a screening tool for PSCI in Thai patients.