Kazuhiro Miyata, Satoshi Hasegawa, Hiroki Iwamoto, Tomohiro Otani, Yoichi Kaizu, Tomoyuki Shinohara, Shigeru Usuda
Department of Physical Therapy, Ibaraki Prefectural University of Health Science, 300-0394 Inashiki-gun, Japan. E-mail: miyatak@ipu.ac.jp
DOI: 10.2340/16501977-2870

Objective: To determine which sections of the Balance Evaluation Systems Test (BESTest) distinguish levels of post-stroke functional walking status and to establish their cut-off scores.
Design: A retrospective cross-sectional study.
Subjects and methods: The BESTest was administered to 87 stroke patients who were able to walk without physical assistance upon discharge from the hospital. Subjects were divided into 3 functional walking status groups: namely, household ambulators, limited community ambulators, and unlimited community ambulators. The receiver operating characteristic curve was determined and the cut-off score and area under the receiver operating characteristic curve (AUROC) of each section calculated.
Results: In the comparison of household and limited community ambulators, the accuracies of all BESTest sections were moderate (AUROC>0. 7), and the cut-off scores were 36. 1–78. 6%. In the comparison of limited and unlimited community ambulators, one section (stability in gait) had high accuracy (AUROC=0. 908, cut-off scores=73. 8%) and 3 sections (biomechanical constraints, anticipatory postural adjustments, and postural response) had moderate accuracy (AUROC=0. 8120–0. 834, cut-off scores=75. 0–83. 4%).
Conclusion: This study demonstrated that different sections of the BESTest had different abilities to discriminate levels of post-stroke functional walking status, and identified cut-off values for targeted improvement.

The Balance Evaluation Systems Test (BESTest), a clinical postural control measure, categorizes postural control systems in 6 different sections. This study investigated which sections of the BESTest distinguish levels of post-stroke functional walking status, which, in turn, is based on walking speed. Among the slower walkers, all sections of the BESTest showed moderate relationships to categories of walking status. Among the faster walkers, 4 sections showed moderate to strong relationships and 2 sections showed weak relationships. This study may have clinical implications for rehabilitation aimed at improving functional walking status in individuals with stroke. These findings will help rehabilitation professionals assess postural control in relation to stroke patients' ability to walk in different settings (e.g. their household or the community) and determine which postural control systems should be prioritized in therapeutic interventions.