Estimation of standing height in spina bifida: model development and validation

Area of Science:

• Pediatric orthopedics
• Biomedical engineering
• Growth and development

Background:

• Childhood standing height estimation often relies on arm span-related (height_AS) models.
• Individuals with spina bifida (SB) may present unique challenges for standard height estimation methods.

Purpose of the Study:

• To develop and cross-validate a novel height_AS model for children with SB.
• To assess the accuracy of existing height_AS models in this population.

Main Methods:

• Developed a new model (M1) using arm span, age, sex, and SB group status.
• Cross-validated M1 and compared it with existing SB-specific (M2) and non-SB specific models (M3-M5) using sequential regression and leave-one-out cross-validation.
• Included 14 individuals with SB and 83 non-SB participants aged 7-16 years.

Main Results:

• The new model (M1) and two existing models (M2, M5) showed non-significant mean differences between actual and estimated height for groups.
• Arm span and SB group status explained 95% of height variance in M1 (R² = 0.95).
• Bland-Altman analysis revealed wide limits of agreement (7.4-10.9 cm), indicating significant individual variability; models M3 and M4 showed considerable errors.

Conclusions:

• Models M1, M2, and M5 demonstrate potential for accurate group-level standing height estimation in children with SB.
• Caution is advised when using these models for individual height estimations due to observed variability.
• Further research may be needed to refine models for precise individual height prediction in pediatric SB populations.