Determination of the physical working capacity in children using three different regression models

Area of Science:

• Exercise Physiology
• Biostatistics
• Pediatric Cardiology

Background:

• Accurate prediction of physical working capacity (W170) is crucial for assessing cardiovascular health in adolescents.
• Traditional linear regression models may introduce systematic errors in W170 prediction.
• Evaluating alternative regression models is necessary for reliable exercise testing.

Purpose of the Study:

• To compare the predictive accuracy of linear, polynomial, and "break point" regression models for W170.
• To assess the suitability of these models using data from adolescent physical education students.
• To determine the optimal model for unbiased W170 prediction from early exercise data.

Main Methods:

• Investigated W170 prediction using data from 28 adolescents (15-16 years old).
• Applied linear, polynomial, and "break point" regression models to heart rate data during a 10-minute incremental workload test.
• Analyzed systematic deviations and prediction errors between measured and predicted W170 values.

Main Results:

• Linear regression showed systematic deviations and overestimated W170.
• Polynomial and "break point" regression models demonstrated strong agreement with heart rate time course.
• These advanced models provided unbiased W170 predictions from the initial 10 minutes of exercise.

Conclusions:

• Polynomial and "break point" regression models are superior to linear regression for predicting W170 in adolescents.
• Accurate W170 prediction is achievable using early-stage exercise test data with appropriate statistical models.
• This research informs more reliable exercise capacity assessments in young populations.