Open Source, Open Science: Development Of OpenLESS As The Automated Landing Error Scoring System

INTRODUCTION: The Open Landing Error Scoring System (OpenLESS) is a novel development aimed at automating clinical measures for lower extremity movement quality, specifically the landing error scoring system (LESS). With increasing utilization of clinical measures to monitor outcomes and limited time during medical visits for a lengthy analysis of functional movement, there is a pressing need to extend automation efforts. Addressing these issues, OpenLESS utilizes freely available markerless motion capture software to automate the LESS using 3D kinematics. PURPOSE: Describe the development and validity of OpenLESS in a preliminary group of healthy, college-aged adults. METHODS: This package was developed in Python (v.3.7.1) and includes several functions including reading in single or batched files, filtering, detecting key events for the stance phase of the jump-landing task, scoring based on the operational definitions of the LESS from the stance phase, and returning the item-specific and total scores in a results table. The event detection function derived from kinematics was validated against force plate-derived timing of events (12 subjects, 98 trials). OpenLESS scores were validated in a preliminary sample against expert rater LESS scores (12 trials). Validation was assessed via Intraclass Correlation Coefficients and Bland Altman Plots. RESULTS: The event detection pipeline was found to be a valid method for identifying initial ground contact and toe-off compared to force-plates (ICC2,1 >.99, p<.001). OpenLESS returned comparable results to that of expert rater consensus scores of the LESS (ICC2,1 >.84, p<.001). The automation significantly reduced the clinician observed LESS scoring duration, <1 minute for batched processing compared to 7 minutes per trial, respectively. CONCLUSIONS: OpenLESS was found to be valid compared to expert rater and it reduced the time to analyze jump-landing movement quality. Table 1. - Key Event Detection with Markerless Motion Capture Intraclass Correlation Coefficient (2,1) Jump Landing Event Time Method Mean (SD) Value P value Initial Ground Contact (seconds) Force Plate 6.745 (4.475) 0.999 <.001 OpenLESS-based 6.775 (4.511) Toe-Off (seconds) Force Plate 7.365 (4.524) 0.999 <.001 OpenLESS-based 7.277 (4.552)