Related Experiment Video
Updated: Feb 18, 2026
14:09
Tilt Testing with Combined Lower Body Negative Pressure: a "Gold Standard" for Measuring Orthostatic Tolerance
Published on: March 21, 2013
22.0K
AI-assisted video analysis of the Trendelenburg test: a feasibility study
Kieran O'Sullivan1,2, Tom Doyle3, Eoghain Quinn3
1Galway University Hospitals, Galway, Ireland. kieranosullivan14@gmail.com.
Scientific Reports
|February 16, 2026
Summary
Artificial intelligence (AI) analysis of the Trendelenburg test offers objective hip abductor function assessment. This feasible method quantifies pelvic and trunk motion, revealing compensatory trunk lean in post-total hip arthroplasty patients.
Area of Science:
- Orthopaedics
- Biomechanics
- Artificial Intelligence in Medicine
Background:
- The Trendelenburg test is standard for assessing hip abductor function but suffers from subjective interpretation and moderate reliability.
- Compensatory trunk lean can mask subtle pelvic drop, reducing diagnostic accuracy.
- Objective quantification of pelvic and trunk motion is needed for improved Trendelenburg test assessment.
Purpose of the Study:
- To evaluate the feasibility and clinical informativeness of AI-based markerless motion analysis for the Trendelenburg test.
- To objectively quantify pelvic, trunk, and knee angles during the Trendelenburg test using standard video recordings.
- To compare compensatory movement patterns between post-total hip arthroplasty (THA) patients and those with native hip pathology.
Main Methods:
- A single-centre cross-sectional feasibility study involving 12 adults (7 post-THA, 5 native hip pathology).
- Standardised single-leg Trendelenburg tests performed bilaterally, recorded with a single smartphone camera.
- Offline video analysis using AI-based markerless motion software (OnForm) to derive pelvic obliquity, trunk lean, and knee angle changes.
Main Results:
- AI-assisted analysis was feasible, rapid (median total workflow time 215.5s), and provided quantifiable data.
- Post-THA patients demonstrated significantly greater trunk lean (median 9.0°) compared to native hip patients (median 3.0°).
- Knee angle deviations ≥3° were observed in 67% of patients, indicating widespread compensatory strategies.
Conclusions:
- AI-assisted single-camera analysis of the Trendelenburg test is a viable, efficient, and informative method.
- The AI approach objectively quantifies key movement parameters and highlights trunk lean as a common compensation in post-THA patients.
- This technology holds potential for enhancing objective documentation and monitoring rehabilitation post-THA, warranting further validation in larger studies.
