Agreement Between Novice Visual Assessment and Classifications Derived from Markerless Motion Capture During...

Area of Science:

• Rehabilitation Science
• Biomechanics
• Movement Analysis

Background:

• Visual assessment is standard for evaluating movement quality in rehabilitation, particularly for sit-to-stand (STS) transfers.
• The alignment between observational ratings and objective motion capture data is not well understood.
• This study investigates agreement between novice visual assessments and markerless motion capture classifications during STS.

Purpose of the Study:

• To compare novice observational ratings of sit-to-stand (STS) performance with classifications derived from a markerless motion capture system.
• To assess inter-rater reliability among novice observers.
• To evaluate agreement between visual assessment and motion capture during STS across different seating conditions.

Main Methods:

• Fifty healthy adults performed STS transfers from three different seating surfaces.
• Two physical therapy students provided observational ratings using a standardized rubric.
• A portable markerless motion capture system recorded kinematics, which were classified by severity.

Main Results:

• High exact agreement (82.3-82.9%) was found between novice raters, but low chance-corrected agreement (Krippendorff's alpha = 0.250-0.323) indicated restricted scale use.
• Agreement between observational and motion capture classifications was low, with negative alpha values (α = -0.224 to -0.561).
• Observational raters tended to assign lower severity classifications than motion capture.

Conclusions:

• Novice raters exhibit low chance-corrected agreement and systematic disagreement with motion capture during STS performance.
• Differences in classification approaches and operational definitions contribute to the observed disagreements.
• Further research is needed to explore these patterns in clinical populations and with varied measurement techniques.