[Agreement on the Barthel Index. A rapid analysis of other and self-assessment in elderly stroke patients]

The relationship between staff and patient ratings of activity restrictions (Barthel Index, BI) was investigated in 120 elderly stroke patients (on average 78 years old) using the Rasch model and the rating scale analysis. In addition, the relationship between the rated activity restrictions and measures from the Geriatric Assessment was analyzed. We found good patient-staff agreement (r = 0.90) with the poorest agreement in the item "bathing" of the BI. There was also a highly significant correlation between staff and patient ratings and the Tinetti Gait and Balance Scales (r = 0.72 and r = 0.76, respectively). Correlations between other measures of the Geriatric Assessment and the rated activity restrictions were low explaining less than 8% of variance. Our findings merit the use of patient ratings of activity restrictions in stroke outcome research. However, self-ratings of activity restrictions were measured by an unstructured interview and it cannot be ruled out that this method had an influence on the correlation between self-rating and staff rating.