Precision exercise medicine: predicting unfavourable status and development in the 20-m shuttle run test performance...

Area of Science:

• Exercise physiology
• Machine learning in sports science
• Adolescent health

Background:

• Cardiorespiratory fitness is crucial for adolescent health.
• Predicting future declines in fitness is important for targeted interventions.
• Machine learning offers potential for personalized health predictions.

Purpose of the Study:

• To evaluate machine learning (random forest classifier) for predicting unfavorable future 20m shuttle run test (20MSRT) status and development in adolescents.
• To identify key baseline characteristics influencing 20MSRT performance over time.

Main Methods:

• A 2-year observational study involving 633 adolescents (12.4±1.3 years).
• Collected 48 baseline characteristics including demographics, lifestyle, physical, psychological, and academic factors.
• Utilized a random forest classifier to predict 20MSRT status and development.

Main Results:

• The random forest classifier achieved 83% AUC for predicting unfavorable 20MSRT status in girls and 76% in boys.
• Fitness, physical activity, academic scores, and psychosocial factors were significant predictors.
• Prediction of 20MSRT development was less accurate, particularly in boys.

Conclusions:

• The random forest classifier effectively predicts future unfavorable 20MSRT status using a holistic profile of baseline characteristics.
• This approach can identify adolescents at risk for declining cardiorespiratory fitness, enabling targeted interventions.
• The study provides a valuable tool for precision exercise medicine research.