Aitkens S, Lord J, Bernauer E, Fowler W M, Lieberman J S, Berck P
Department of Physical Medicine and Rehabilitation, University of California, Davis.
Muscle Nerve. 1989 Mar;12(3):173-7. doi: 10.1002/mus.880120302.
This study investigates the relationship between manual muscle test scores (MMT) and quantitative isometric strength measurements (QIS). It also evaluates the implications of that relationship for design of therapeutic trials. Extension and flexion strength at the elbows, hips, and knees of 21 neuromuscular disease patients were tested a total of 26 times utilizing both MMT and QIS testing. Paired data were evaluated with Spearman ranked correlation coefficients, and then QIS was predicted from MMT using Lowess, a consistent form of nonparametric regression. Finally, the implications of the Lowess analysis for designing a therapeutic trial were evaluated. MMT and QIS measurements were significantly correlated in all movements tested. Lowess analysis yielded prediction errors ranging from 16 to 24% of QIS range. Analysis of the sample size needed for a therapeutic trial suggested that a protocol measuring MMT would require more subjects for the same level of statistical significance as a protocol measuring QIS. Since it was not possible to reliably predict QIS values from MMT scores, such conversions are not appropriate for clinical use. This inconsistent relationship between MMT and QIS carries major implications for the design of therapeutic trials. Since therapeutic trials are time consuming, expensive, and most centers do not have large numbers of individuals available, using QIS as an outcome measure is a preferable research design.
本研究调查了徒手肌力测试分数（MMT）与定量等长肌力测量（QIS）之间的关系。它还评估了这种关系对治疗试验设计的影响。使用MMT和QIS测试对21名神经肌肉疾病患者的肘部、髋部和膝部的伸展和屈曲力量总共进行了26次测试。配对数据采用Spearman等级相关系数进行评估，然后使用Lowess（一种非参数回归的一致形式）从MMT预测QIS。最后，评估了Lowess分析对设计治疗试验的影响。在所有测试的动作中，MMT和QIS测量结果显著相关。Lowess分析得出的预测误差范围为QIS范围的16%至24%。对治疗试验所需样本量的分析表明，对于相同的统计学显著性水平，与测量QIS的方案相比，测量MMT的方案需要更多的受试者。由于无法从MMT分数可靠地预测QIS值，因此这种转换不适合临床使用。MMT和QIS之间这种不一致的关系对治疗试验的设计具有重要意义。由于治疗试验耗时、昂贵，而且大多数中心没有大量可用的个体，因此使用QIS作为结果指标是一种更可取的研究设计。
