BACKGROUND AND OBJECTIVES: There has been a need for functional status measurement tool with better validity than the existing tools such as New York Heart Association Functional Class. Duke Activity Status Index (DASI) is a representative example of a tool that was developed to enhance the validity of measurement by asking the patients about the ability to perform specific activities and scoring the response. Because such a tool must be culture-sensitive, it is desirable to use 'Koreanized' version of the tool. No Koreanized version of the functional status measurement tool has been developed yet. The objective of this study is to develop a Korean version of DASI. MATERIALS AND METHOD: In the developmental phase, a pilot questionnaire asking the ability to perform specific activity was made with reference to existing tools, such as Specific Activity Scale and DASI. Substitution, correction and addition of items were done through the pilot study. Ninety-nine patients was asked to fill developmental version of questionnaire, then underwent treadmill exercise test. Weight for each items were assigned to optimize the correlation between the calculated index (KASI) and total treadmill exercise time. Criteria for categorical functional classification were determined to maximize the agreement between KASI-estimated functional class (KASIFC) and functional class estimated by exercise time. In the validation phase, final version of questionnaire was tested in independent group of 159 patients. The questionnaire was self-administered. Canadian Cardiovascular Society Functional Class (CCSFC) was estimated by the physician who is in charge of treadmill exercise test. RESULTS: In the validation phase, Spearman correlation coefficient between KASI and treadmill exercise time was 0.62(p=.0001) and between CCSFC and exercise time -0.48(p=.0001). KASIFC agreed with functional class estimated by exercise time in 77% of cases, disagreed by 1 class in 20% and by 2 classes in 1%. KASIFC agreed with functional class estimated by exercise time in 77% of cases, disagreed by 1 class in 20% and by 2 classes in 1%. These two methods did not differ significantly in categorical classification. CONCLUSION: KASI is more accurate or at least as accurate as the existing tool in estimation of functional status. The characteristics such as self-administration, availability of outcome as a continuous variable are expected to make it a convenient, efficacious and useful tool in various clinical researches.