Reliability and validity tests of an evaluation tool based on the modified Barthel Index

Abstract

Aims: To establish the reliability and validity of a performance evaluation tool based on the modified Barthel Index (PET-MBI) for elderly residents of long-term care facilities. Methods: To develop the PET-MBI, the 10-task modified Barthel Index was translated into Japanese by three therapists, taking Japanese culture and lifestyles into account. Its reliability was independently assessed by two therapists who evaluated 16 participants (mean age 85.6 years; standard deviation (SD)=7.3) from one facility. The internal consistency and validity of the PET-MBI was then assessed by two therapists at three facilities, who evaluated a total of 110 participants (mean age 85.0 years; SD=7.1). Findings: We excluded the item ‘stair climbing’ because participants did not have to do this. Kappa coefficients of intra-tester reliability based on the test-retest method were 0.70–1.00 for individual tasks of the PET-MBI; kappa coefficients for inter-tester reliability were 0.72–1.00. Cronbach's coefficient alpha of internal consistency was 0.93. Factor analysis according to the PET-MBI revealed that a single factor explained self-care in elderly residents. Conclusions: The PET-MBI demonstrated high reliability, validity and internal consistency for the assessment of elderly residents in facilities.