Discriminant analysis for predictor of falls in stroke patients by using the Berg Balance Scale.

Noriaki Maeda; Yukio Urabe; Masahito Murakami; Keisuke Itotani; Junichi Kato

Return

Discriminant analysis for predictor of falls in stroke patients by using the Berg Balance Scale.

Author: Noriaki MAEDA ¹ ; Yukio URABE ¹ ; Masahito MURAKAMI ² ; Keisuke ITOTANI ² ; Junichi KATO ³
Author Information

1. Department of Sport Rehabilitation, Graduate of Biomedical & Health Sciences, Hiroshima University, Hiroshima, Japan.
2. Department of Physical Therapy, Faculty of Rehabilitation, Kobe International University, Kobe, Hyogo, Japan.
3. Department of Internal Medicine, Hyogo Prefectural Rehabilitation Center at Nishi-harima, Tatsuno, Japan.
Publication Type:Journal Article
Keywords: Berg Balance Scale; Functional Independence Measure; discriminant; falls; stroke
MeSH: Accidental Falls; prevention & control; Aged; Discriminant Analysis; Female; Hemiplegia; complications; Humans; Male; Middle Aged; Postural Balance; Risk Assessment; Risk Factors; Severity of Illness Index; Stroke; complications; Stroke Rehabilitation
From:Singapore medical journal 2015;56(5):280-283
CountrySingapore
Language:English
Abstract:
INTRODUCTIONAn observational study was carried out to estimate the strength of the relationships among balance, mobility and falls in hemiplegic stroke inpatients. The objective was to examine factors that may aid in the prediction of the likelihood of falls in stroke patients.
METHODSA total of 53 stroke patients (30 male, 23 female) aged 67.0 ± 11.1 years were interviewed regarding their fall history. Physical performance was assessed using the Berg Balance Scale (BBS) and the Functional Independence Measure (FIM) scale. Variables that differed between fallers and non-fallers were identified, and a discriminant function analysis was carried out to determine the combination of variables that effectively predicted fall status.
RESULTSOf the 53 stroke patients, 19 were fallers. Compared with the non-fallers, the fallers scored low on the FIM, and differed with respect to age, time from stroke onset, length of hospital stay, Brunnstrom recovery stage and admission BBS score. Discriminant analysis for predicting falls in stroke patients showed that admission BBS score was significantly related to the likelihood of falls. Moreover, discriminant analysis showed that the use of a significant BBS score to classify fallers and non-fallers had an accuracy of 81.1%. The discriminating criterion between the two groups was a score of 31 points on the BBS.
CONCLUSIONThe results of this study suggest that BBS score is a strong predictor of falls in stroke patients. As balance is closely related to the risk of falls in hospitalised stroke patients, BBS might be useful in the prediction of falls.