短版巴氏量表應用於中風患者之最小可偵測變化值及反應性

Minimal Detectable Change and Responsiveness of the 5-item Barthel Index in Patients with Stroke

呂文賢(Wen-Shian Lu) ； 李雅珍(Ya-Chen Lee) ； 凃富籌(Fuh-Chour Twu) ； 謝清麟(Ching-Lin Hsieh)

《台灣復健醫學雜誌》 41卷2期 (2013/06) Pp. 113-119

https://doi.org/10.6315/2013.41(2)04

摘要

Background and Purpose: A concise measurement tool can be easily accepted by clinicians. The 5-item Barthel Index (BI-5) has only 5 items and thus the potential to be used routinely in clinical settings. However, the minimal detectable change (MDC) and responsiveness of the BI-5 have not been verified sufficiently, limiting the interpretability of the change scores of the BI-5. The purposes of this study were to investigate the MDC and responsiveness of the BI-5 in patients with stroke and to compare the MDC and responsiveness of the BI-5 to those of the original BI.Methods: This study used secondary data analysis. A prior data set (with 56 stroke patients administered the BI twice, 14 days apart) was used to calculate the MDCs of the BI and the BI-5. Another data set (with 226 stroke patients administered the BI on the 14th day and the 90th day post onset) was used to calculate the responsiveness of the BI and that of the BI-5. Responsiveness indices included group-level responsiveness (i.e., effect size d, standardized response mean, SRM and paired t test) and individual-level responsiveness (i.e., the difference in number of patients whose change scores on the BI and BI-5 exceeded the respective MDC, and the difference in value of MDC ratio).Results: The MDCs (MDC%) of the BI and BI-5 were 4.3 (28.8%) and 3.1 (41.1%), respectively. The group-level responsiveness of the BI and that of the BI-5 were both large (both d and SRM were more than 0.8). The mean differences of the BI and the BI-5 between two measurements were significant. In individual-level responsiveness, the number of patients administered the BI having change scores exceeding the MDC was greater than that of those administered the BI-5 (155 vs. 141). However, the difference was not statistically significant. The value of the MDC Ratio of the BI was also significantly higher than that of the BI-5.Conclusion: The group-level responsiveness of the BI-5 was good. However, the random measurement error of the BI-5 was larger than that of the original BI, and the individual-level responsiveness of the BI-5 was slightly less than that of the BI. Thus, we recommend that given sufficient time for assessment, clinicians and researchers should use the original BI rather than the BI-5 to sufficiently show the change of ADL function in stroke patients. The results of this study can help clinicians and researchers interpret patients' change scores and select an appropriate ADL measure.

關鍵字

短版巴氏量表 ； 中風 ； 最小可偵測變化值 ； 反應性

並列摘要

Background and Purpose: A concise measurement tool can be easily accepted by clinicians. The 5-item Barthel Index (BI-5) has only 5 items and thus the potential to be used routinely in clinical settings. However, the minimal detectable change (MDC) and responsiveness of the BI-5 have not been verified sufficiently, limiting the interpretability of the change scores of the BI-5. The purposes of this study were to investigate the MDC and responsiveness of the BI-5 in patients with stroke and to compare the MDC and responsiveness of the BI-5 to those of the original BI.Methods: This study used secondary data analysis. A prior data set (with 56 stroke patients administered the BI twice, 14 days apart) was used to calculate the MDCs of the BI and the BI-5. Another data set (with 226 stroke patients administered the BI on the 14th day and the 90th day post onset) was used to calculate the responsiveness of the BI and that of the BI-5. Responsiveness indices included group-level responsiveness (i.e., effect size d, standardized response mean, SRM and paired t test) and individual-level responsiveness (i.e., the difference in number of patients whose change scores on the BI and BI-5 exceeded the respective MDC, and the difference in value of MDC ratio).Results: The MDCs (MDC%) of the BI and BI-5 were 4.3 (28.8%) and 3.1 (41.1%), respectively. The group-level responsiveness of the BI and that of the BI-5 were both large (both d and SRM were more than 0.8). The mean differences of the BI and the BI-5 between two measurements were significant. In individual-level responsiveness, the number of patients administered the BI having change scores exceeding the MDC was greater than that of those administered the BI-5 (155 vs. 141). However, the difference was not statistically significant. The value of the MDC Ratio of the BI was also significantly higher than that of the BI-5.Conclusion: The group-level responsiveness of the BI-5 was good. However, the random measurement error of the BI-5 was larger than that of the original BI, and the individual-level responsiveness of the BI-5 was slightly less than that of the BI. Thus, we recommend that given sufficient time for assessment, clinicians and researchers should use the original BI rather than the BI-5 to sufficiently show the change of ADL function in stroke patients. The results of this study can help clinicians and researchers interpret patients' change scores and select an appropriate ADL measure.

並列關鍵字

5-item Barthel index ； stroke ； minimal detectable change ； responsiveness