Objective
To develop a Korean version of Modified Barthel Index (K-MBI) for subjects with stroke and to test the reliability and validity of K-MBI. Method: Six senior physiatrists translated the 5th version of MBI into K-MBI and we administered K-MBI to 30 subjects with stroke. Fifteen different examiners working at the 5 different university hospitals evaluated video-recorded examination cases independently to test the reliability and validity of K-MBI. We analyzed intra- and inter-rater reliabilities of the K-MBI by the Kendall's coefficient of concordance and Spearman's correlation coefficients, respectively. Cronbach's alpha coefficient was used for assessing internal consistency of the K-MBI and Spearman's correlation between the K-MBI and Brunnström stage was employed to evaluate the validity of the K-MBI. Results: The intra-rater reliabilities of physiatrists, resident physicians of rehabilitation medicine and occupational therapists were 0.93∼1.00, 0.87∼0.99, and 0.97∼1.00 (p＜ 0.01), respectively. The inter-rater reliabilities were 0.93∼0.98 (p＜0.01) and Cronbach's alpha was 0.84 (p＜0.01) as the internal consistency reliability of K-MBI. For construct validation study, each item of K-MBI had significant correlation with total score of K-MBI (r=0.54∼0.78, p＜ 0.01). Conclusion: The K-MBI is a reliable and valid instrument for measuring functional status of subjects with stroke. (J Korean Acad Rehab Med 2007; 31: 283-297)