Estimating total Barthel scores from just three items: the European Stroke Database 'minimum dataset' for assessing functional status at discharge from hospital

Background: the European Stroke Database (ESDB) Project aims to develop a 'common clinical language' for stroke care by agreeing on terminology, definitions and clinical assessments. Each area of stroke assessment has a 'minimum dataset', which can be collected routinely and more detailed information can be added for particular studies, Measurement of patients' functional status at discharge is essential for assessing the impact of hospital care, but even simple activities of daily living scales like the Barthel index may not be easy to use routinely on busy acute units. We thus aimed to further simplify the 20-point Barthel index by reducing it to a few key items. Methods: we initially analysed data on 169 consecutive stroke patients discharged from one British hospital and found that a simple formula involving the combined subscores for urinary continence (Blad), bed-chair transfers (Trans) and indoor mobility (Mob)-(Blad + Trans + Mob) x 2.39 + 0.14-predicted the total BI score to within 1 point in 79% and to within 2 points in 95% of cases. We then tested this three-item Barthel index (BI3) in four different stroke datasets (total n = 824). Results: the predictions were accurate to +/-1 point in 72-81% and to +/-2 points in 88-97% of cases, and BI3 accounted for 95% of the variance in total Barthel score. It was more accurate in patients without obvious mental impairment. Conclusions: for studies involving large groups of stroke patients, it is sufficient to know about each patient's continence, transfers and indoor mobility at discharge, in order to estimate the total Barthel score. These measures have therefore been incorporated, together with a simple observational measure of cognitive status, into the database minimum dataset for short-term functional outcome and are now being validated in international studies.