Stott N Susan, Atherton W Guy, Mackey Anna H, Galley Ian J, Nicol Richard O, Walsh Stewart J
Department of Surgery, University of Auckland, Auckland, New Zealand.
Arch Phys Med Rehabil. 2005 Dec;86(12):2337-41. doi: 10.1016/j.apmr.2005.06.021.
To assess the reliability and validity of a newly described classification of sagittal plane alignment in spastic diplegic gait.
Twenty split-screen videos of children with spastic diplegia, Gross Motor Function Classification System levels I to III, were viewed on 2 occasions, 6 weeks apart, by 5 raters. The sagittal plane alignments of the right and left lower limbs in gait were classified separately as true equinus, jump knee, apparent equinus, or crouch, based on the published classification. A fifth category, nonclassifiable, was used if classification was not possible. We then used sagittal plane kinematic data to confirm the classification for each subject and these were compared with rater classification scores, which used the video information only.
Tertiary-level children's hospital.
Three pediatric orthopedic surgeons and 2 pediatric orthopedic residents.
Not applicable.
Gait classification scores derived from visual observation were compared among and within raters. The gait classification scores derived from visual observation were compared with the scores derived from sagittal plane kinematic data to assess validity.
A moderate correlation was found among the 5 raters within each session, with an interrater weighted kappa score of .45 in session 1 and .49 in session 2. The intrarater, weighted kappa scores showed a moderate to substantial level of agreement between sessions, ranging from .50 to .68. The classification scores of individual raters had moderate validity when compared with classifications derived from the sagittal plane kinematic data. However, there was a substantial level of agreement between the consensus opinion and the classification obtained using the kinematic data as well as the video recordings (weighted kappa=0.8).
This classification has only moderate reliability and validity when a single experienced rater views the 2-dimensional gait videos. However, the consensus opinion derived from the scores of 5 raters considerably improves the validity of the assessment.
评估一种新描述的痉挛性双侧瘫步态矢状面排列分类方法的可靠性和有效性。
选取20名痉挛性双侧瘫儿童（粗大运动功能分类系统I至III级）的分屏视频，由5名评估者分两次观看，两次观看间隔6周。根据已发表的分类方法，将步态中右下肢和左下肢的矢状面排列分别分类为真性马蹄足、跳跃膝、假性马蹄足或蹲伏步态。若无法分类，则使用第五类“不可分类”。然后，我们使用矢状面运动学数据来确认每个受试者的分类，并将这些数据与仅使用视频信息的评估者分类分数进行比较。
三级儿童医院。
三名小儿骨科医生和两名小儿骨科住院医师。
不适用。
比较评估者之间以及评估者内部通过视觉观察得出的步态分类分数。将通过视觉观察得出的步态分类分数与矢状面运动学数据得出的分数进行比较，以评估有效性。
在每个阶段，5名评估者之间存在中等程度的相关性，阶段1的评估者间加权kappa分数为0.45，阶段2为0.49。评估者内部加权kappa分数显示，各阶段之间的一致性为中等至高度，范围从0.50至0.68。与矢状面运动学数据得出的分类相比，个别评估者的分类分数具有中等有效性。然而，共识意见与使用运动学数据以及视频记录得出的分类之间存在高度一致性（加权kappa = 0.8）。
当由一名经验丰富的评估者观看二维步态视频时，这种分类方法的可靠性和有效性仅为中等。然而，由5名评估者的分数得出的共识意见显著提高了评估的有效性。
