RESEARCH REPORT POSTER DISPLAY

VALIDATION OF A PREDICTIVE MODEL FOR LOCOMOTION IN STROKE PATIENTS: INVESTIGATION USING THE FUNCTIONAL BALANCE SCALE. Sugimoto S¹, Maruya K², Saitoh H³, Asano S⁴; ¹Health Science University, Minamitsurugunn Yamanashi, Japan. ²Musashidai Hospital, Hidaka Saitama, Japan. ³Tukuba Memorial Hospital, Tsukuba Ibaraki, Japan. ⁴Tukuba Central Hospital, Ushiku Ibaraki, Japan

PURPOSE: The Functional Balance Scale [FBS] is a balance test including not only sitting balance but also higher-level balance activities such as standing on one leg and stepping. The FBS is composed of 14 sub-tasks, which are allotted from 0 to 4 points (0 indicating an inability to complete any aspect of the task) and the total score is commonly used as the index of balance ability. Recently, many researchers have reported that the FBS is useful in predicting locomotive levels in stroke patients and risks of falling in the elderly. We also previously investigated the relationship between locomotive levels and the FBS scores in chronic stroke patients (Sugimoto et al, 2005). Our report emphasized that each sub-task of the FBS predicted locomotive levels because the point which was needed to get a good locomotive level was different from one task to another. However, we couldn’t confirm the validity of the predictive model in the previous study. Additionally, five-graded measurements were too many to judge whether the activity of each sub-task was good or not. The purpose of this study was to simplify the sub-tasks of the FBS and to investigate the validity of a predictive model using the simplified sub-tasks for locomotion in stroke patients. RELEVANCE: This study is useful in order to clarify methods to easily predict locomotive levels in stroke patients. PARTICIPANTS: Eighty two chronic stroke patients (34 women and 48 men, 69.5±11.3 years old), who could walk indoors by themselves consented to participate in this study. METHODS: The subjects were randomly divided into the predictive group and the verifiable group after consideration of two locomotive levels (indoors/ outdoors). After frequency distribution tables were made for locomotive levels and 14 sub-tasks of the FBS, a threshold level to distinguish the difference between the two locomotive levels was investigated in each sub-task in the predictive group. ANALYSIS: The predictive expression for locomotive levels was formulated by discriminant analysis method using the two-graded sub-tasks. The hitting ratio of the locomotive level in the verifiable group was calculated with this predictive expression. RESULTS: The results of the discriminant analysis using the stepwise selection method indicated that “Turn to look behind” and “Turn 360 degrees” in the sub-tasks were the most effective predictors of locomotive levels. In the above two tasks, if a patient could turn to look behind from both sides and he could turn 360 degrees slowly and safely, then the patient could be predicted as being in good condition. The hitting ratio of locomotive levels was 80.4% in the predictive group and 68.2% in the verifiable group. CONCLUSIONS: Our results suggest that this predictive model using the simplified sub-tasks of the FBS is useful in the prognosis of locomotive levels in chronic stroke patients. IMPLICATIONS: Physical therapy practice KEYWORDS: stroke patients, locomotion, Functional Balance Scale. FUNDING ACKNOWLEDGEMENTS: N.A. CONTACT: s.sugimoto@kenkoudai.ac.jp