Reliability of video-based identification of footstrike pattern and video time frame at initial contact in recreational runners

Abstract Introduction: In clinical practice, joint kinematics during running are primarily evaluated by clinicians by two-dimensional (2D) video recordings and motion-analysis software. The applicability of these methods depends on the clinicians´ ability to evaluate kinematics in a similar manner. Clinicians´ ability to evaluate knee- and hip angles at initial foot strike is uninvestigated. Objective: To investigate the intra- and inter-rater reliability of clinicians’ ability evaluate the knee- and hip angles at foot strike during running. Methods: Eighteen recreational runners were recorded twice during treadmill running using a simple 2D video setup. Two blinded raters evaluated each video twice with freeware motion-analysis software. Results: In 50% of the analyses, the mean differences between the raters’ quantification of angles were significant different (p<0.05), although less than two degrees in all analyses, and less than one degree in 75% of all analyses. The range from upper to lower prediction limit for the knee angles varied three to eight degrees for the within day reliability and nine to 14 degrees for the between day reliability. Similarly, the ranges in the hip angles were three to seven degrees and nine to 11 degrees, respectively. Conclusion: The mean differences in evaluation of knee- and hip angles at foot strike was less than two degrees between raters. In a clinical setting this discrepancy between raters could be acceptable in evaluation of the individual runners foot strike pattern, but the method has limited applicability in research.