Assessing knee hyperextension in patients following stroke: A comparison of Siliconcoach movement analysis software and experienced clinical observation

Background: Observational gait analysis used by clinicians has been shown to have moderate reliability in relation to joint angles. Siliconcoach movement analysis software uses a digital image to determine joint angles. The aim of this study was to compare Siliconcoach software against experienced clinicians in assessing knee hyperextension in patients following stroke. Method: This was an observational comparison study with a convenient sample of 20 ambulant participants with a unilateral stroke videoed during three walks. To determine knee hyperextension and the reliability of Siliconcoach the recordings were assessed by three raters using Siliconcoach and by three experienced clinicians. Findings: The Kappa statistic for intra-observer agreement for knee hyperextension ranged from 0.8 to 1.0 (substantial to almost perfect) and from 0.6 to 0.9 for inter-observer agreement (moderate to almost perfect). Intra-rater and inter-rater ICC scores for Siliconcoach were all >0.75 suggesting very high reliability. Clinicians compared with Siliconcoach ranged from 0.4 to 0.7(fair to substantial). Conclusion: Observational gait analysis appears to be an adequate measurement tool in clinical practice particularly when undertaken by the same experienced therapist. However, Siliconcoach may be a more reliable measure and provides an objective measure of quantifying knee joint angles for both research and clinical practice.