Prediction bands and intervals for the scapulo-humeral coordination based on the Bootstrap and two Gaussian methods

Area of Science:

• Biomechanics
• Orthopedics
• Sports Medicine

Background:

• Quantitative motion analysis protocols exist for scapula-humerus coordination.
• Lack of reference prediction intervals and bands hinders assessment of
• normal
• scapular resting position or coordination patterns.

Purpose of the Study:

• To establish reference prediction intervals and bands for the "ISEO" (Integrated Scapular Exercise Outcome) protocol.
• To compare non-parametric Bootstrap and parametric Gaussian methods for generating these references.
• To investigate age-related differences in scapulo-humeral coordination patterns.

Main Methods:

• Utilized the non-parametric Bootstrap approach and two parametric Gaussian methods.
• Analyzed 111 asymptomatic subjects divided into age groups (18-30, 31-50, 51-70).
• Computed "monolateral" and "differential" prediction bands and intervals, comparing methods via cross-validation.

Main Results:

• Bootstrap method achieved 86-90% mean coverage, outperforming parametric methods (67-70% for bands, 87-88% for intervals).
• Bootstrap prediction bands revealed age-related changes in scapular retraction, lateral rotation, and posterior tilt.
• Age-stratified references are crucial for accurate patient classification.

Conclusions:

• The Bootstrap method offers optimal coverage and is recommended over parametric approaches for scapulo-humeral coordination analysis.
• Age-specific prediction intervals and bands are essential for reliable clinical assessment and patient classification.