Less Subjectiveness In Clinical Assessments For Cutting And Pivoting Athletes Using Markerless Motion Capture

This study explores an AI-driven, open-source motion capture system (OpenCap) to evaluate landing errors in the drop jump assessment. Assessing landing mechanics is vital for physical therapists to stratify injury risk, and the traditional Landing Error Scoring System (LESS) relies on subjective assessments to determine this. PURPOSE: This study aims to address inconsistencies in subjective assessments with a more objective, standardized approach, requiring limited software knowledge. METHODS: An Excel-based scoring template was created, using kinematic data from drop- jump recordings. The scoring template was used to classify landings as “pass” or “fail” across 15 items based on kinematic thresholds, with a weighted score for item 16. The single-session reliability of scores from 14 subjects over three landings was assessed using a two-way mixed- effects intraclass correlation coefficient (absolute agreement) (ICC). Additionally, the template- derived scores were evaluated against two novice raters using ICCs for absolute agreement and consistency. RESULTS: Across the three selected landings (N = 42), template-derived scores showed good reliability with a 95% confidence interval (CI95) from moderate to excellent (ICC [CI95]: .76 [.53, .91], p < .001). Compared to novice raters, the template-derived scores showed moderate agreement (ICC [CI95]: .62 [.37, .78], p < .001) and consistency ICC [CI95]: .66 [.43, .81], p < .001) across subjects. The mean template-derived score across trials (N = 57) was 4.7 ± 0.2 (CI95: 4.3, 5.1), with scores ranging between 2 and 8. The largest proportion of landings were rated as excellent (43.9%), followed by good (35.1%), poor (14.0%), and moderate (7.0%). For scoring items, the highest error rates were observed for items 1 (89.5%) and 2 (82.5%), whereas no errors were detected for items 6, 9, 10, and 11. For item 16, most landings scored as average (score of 1, 70.2%) and stiff (score of 2, 28.0%). CONCLUSION: OpenCap combined with the Excel-based template standardizes landing assessments, reducing subjectivity. Preliminary findings suggest consistent data across multiple trials; however, further exploration is needed to evaluate the influence of kinematic thresholds, the overall impression score, and different athletic cohorts.