Relationship between neck kinematics and neck dissability index. An approach based on functional regression

Area of Science:

• Biomechanics
• Rehabilitation Medicine
• Statistical Modeling

Background:

• Traditional methods using numerical neck movement variables show limited correlation with disability levels, especially in nonspecific neck pain.
• Existing prediction models often fail to capture the complexity of neck movement in relation to patient-reported disability.

Purpose of the Study:

• To apply Functional Data Analysis (FDA), specifically scalar-on-function regression, to predict the Neck Disability Index (NDI) using complete neck movement data.
• To improve the prediction accuracy of neck disability compared to conventional scalar-based methods.

Main Methods:

• Utilized scalar-on-function regression models within FDA to analyze kinematic curves of neck movement.
• Compared functional regression models with traditional models using only scalar predictors.
• Identified angular velocity curves as the most effective predictor in the best functional model.

Main Results:

• Functional regression models demonstrated a significant improvement, doubling the multiple correlation coefficient compared to scalar predictors alone.
• The optimal model, using angular velocity curves, achieved a multiple correlation coefficient of 0.64.
• Functional models provided interpretable insights, highlighting the influence of movement's braking phases on NDI.

Conclusions:

• Functional regression models offer superior predictive power for neck disability by leveraging comprehensive kinematic data.
• FDA enables a more nuanced understanding of the relationship between neck movement dynamics and disability.
• The findings support the use of FDA in clinical settings for more accurate and interpretable assessment of neck pain severity.