Predictors of Static Postural Loading in Primary-School-Aged Children: Comparing Elastic Net and Multiple Regression...

Area of Science:

• Pediatric Health
• Biomechanical Engineering
• Data Science

Background:

• Sedentary lifestyles pose significant health risks to children.
• Establishing good postural hygiene is crucial for musculoskeletal development and long-term well-being in primary school children.
• Understanding and mitigating static postural loading (SPL) is essential for pediatric health.

Purpose of the Study:

• To develop and internally validate a regularized regression model for predicting static postural loading (SPL) in primary school children.
• To identify key predictors of SPL in this age group.
• To compare the performance of regularized elastic net (EN) with multiple linear regression (MLR).

Main Methods:

• Systematic literature review and expert panels to identify SPL predictors.
• Data collected from 258 primary school children.
• Developed regularized elastic net (EN) and multiple linear regression (MLR) models, validated using five-fold cross-validation and permutation importance analysis.

Main Results:

• Both EN and MLR models showed good predictive fit, with EN marginally outperforming MLR.
• Key predictors of SPL identified: postural risk, sedentary behavior, task duration, and Body Mass Index (BMI).
• Permutation importance analysis provided predictor rankings for both models.

Conclusions:

• Sedentary lifestyles directly and alarmingly impact children's overall health.
• Identifying and managing factors contributing to SPL in children is critical.
• Regularized regression methods demonstrate strong performance for forecasting outcomes in clinical settings.