RESUMEN
INTRODUCTION: Barthel index consisting 10 items is used to estimate physical dependence of elderly and physically disabled. A shorter version with 5 items has been developed and it is more suitable as a rapid screening tool of physical dependence. OBJECTIVES: To assess the measurement agreement between 10-item Barthel index and 5-item shorter version. METHODS: The 10-item Barthel index was translated to Sinhala, adhering to the standard protocols and validated among 286 patients with varying degree of physical disability selected by stratified quota system. From the same data sheets, scores given for bathing, transfer, toileting, walking and climbing steps were used to make the shorter version. RESULTS: The 5-item Barthel index showed a high internal consistency (global Cronbach's alpha = 0.93). The 10-item version also showed a similar internal consistency (global Cronbach's alpha = 0.92) while its item-total correlations varied between 0.64- 0.90 for all items except for the urinary and bladder functions. In the factor analysis, urinary and bowel functions factored together and independent to other items and these two factors accounted for 73% variation of the score. The total scores of the 10-item and 5-item versions showed a high correlation (r = 0.9, p<0.001). In the Bland-Altman plot, more than 95% of data points were within the +/- 1.96 SD tolerance limits. CONCLUSIONS: This analysis illustrates the reliability and validity of the Sinhala version of 10-item Barthel index in estimating physical activities of daily living and the high measurement concordance between the standard 10-item and 5-item shorter versions.
