Statistical Validation for Clinical Measures: Repeatability and Agreement of Kinect™-Based Software.

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Clinical Measurement

Background:

• Rehabilitation relies on subjective patient progress evaluation by clinicians.
• Objective, quantitative methods are needed to enhance rehabilitation assessment.
• Current methods lack precision and repeatability in tracking patient recovery.

Purpose of the Study:

• To develop and validate a Kinect-based software platform for objective upper limb rehabilitation tracking.
• To quantify patient movement, joint angles, and trajectories during rehabilitation.
• To statistically validate the software's performance, repeatability, and clinical utility.

Main Methods:

• Utilized Kinect sensor(s) to capture upper limb movements.
• Developed software to determine joint angles and trajectories.
• Statistically validated software measures against goniometer using Bland-Altman analysis.

Main Results:

• Software measurements showed excellent agreement and interchangeability with goniometer.
• Bland-Altman analysis confirmed high repeatability (95% confidence).
• Physiotherapists reported encouraging qualitative feedback on clinical usability.

Conclusions:

• The developed software provides objective patient data for rehabilitation success quantification.
• The system offers a simple, low-cost, and visually informative tool for rehabilitation.
• The software is validated for interchangeable use with traditional goniometric methods in clinical practice.