Psychometric characteristics of the Barthel activities of daily living index in stroke patients.

BACKGROUND AND PURPOSE

Although the Barthel activities of daily living (ADL) index (BI) is widely used to measure disability in Taiwan, the usage of the BI has not been scientifically justified. This study examined the reliability, validity, and responsiveness of the BI in a cohort of stroke patients who were followed for half a year after stroke.

METHODS

One hundred and twenty-one patients were assessed using the BI at 14, 30, 90, and 180 days after stroke (DAS). Reliability of the BI was determined by estimation of inter-rater agreement and internal consistency. Validity of the BI was determined by examining convergent construct validity and predictive validity. Responsiveness was determined using effect size and t-statistics calculations.

RESULTS

There was moderate to excellent agreement between raters for individual items (kappa value range, 0.53-0.94) and total score (ICC = 0.94). An excellent internal consistency was found within the BI at four DAS points (alpha value range, 0.89-0.92). The BI scores closely correlated with scores of the Fugl-Meyer motor assessment (measuring impairment) and the Berg balance scale (measuring balance) (Pearson's r > or = 0.78, p < 0.0001). The Frenchay activities index (measuring instrumental ADL) at 180 DAS showed moderate correlation with the BI scores obtained at 14, 30, and 90 DAS (Pearson's r > or = 0.59, p < 0.0001). The effect size d, standardized response mean, t-statistics, and p-values showed that the BI was moderately to highly responsive at each stage of stroke recovery except the late stage (90-180 DAS).

CONCLUSION

This study has demonstrated that the BI is a useful instrument with high inter-rater reliability, internal consistency, convergent and predictive validity, and adequate responsiveness in assessing ADL functions in stroke patients in Taiwan.