Improvement in the Roche-Wainer-Thissen stature prediction model: A comparative study

Area of Science:

• Pediatrics
• Anthropometry
• Biostatistics

Background:

• The Roche-Wainer-Thissen (RWT) model, established in 1975, is a standard for predicting adult stature.
• While generally effective, the RWT model's procedural steps present opportunities for enhancement.
• Accurate stature prediction is crucial for monitoring child growth and development.

Purpose of the Study:

• To investigate and compare seven variations of the RWT prediction model.
• To identify an improved method for predicting adult stature in Caucasian American children.
• To enhance the accuracy and reliability of stature prediction models.

Main Methods:

• Evaluation of seven modified versions of the RWT prediction model.
• Comparative analysis based on prediction accuracy and reliability.
• Implementation of multivariate cubic spline smoothing (MCS²) as a novel approach.

Main Results:

• The multivariate cubic spline smoothing (MCS²) method demonstrated superior performance.
• MCS² offers a simplified procedure compared to the existing multivariate semi-metric smoothing (MS²) method.
• The recommended MCS² method resulted in smaller maximum deviations between predicted and actual adult statures.

Conclusions:

• The multivariate cubic spline smoothing (MCS²) method is recommended for predicting adult stature in Caucasian American children.
• This new method enhances prediction accuracy and simplifies the prediction process.
• The findings contribute to more reliable pediatric growth assessment tools.