Commentary: Methods for calculating growth trajectories and constructing growth centiles.

Area of Science:

• Pediatrics
• Biostatistics
• Growth Monitoring

Background:

• Accurate statistical analysis of growth data is crucial for understanding child development and health.
• Existing growth trajectory models and centile construction methods require critical evaluation for optimal application.

Purpose of the Study:

• To review and critically assess statistical methods for analyzing early life growth data.
• To evaluate growth trajectory models and centile construction techniques using real-world datasets.

Main Methods:

• Analysis of four growth trajectory models (Laird-Ware, SITAR, brokenstick, FACE) applied to length data.
• Review of centile construction methods with worked examples using birthweight and fetal head circumference data from the INTERGROWTH-21st project.
• Evaluation of the GAMLSS software for centile analysis.

Main Results:

• Brokenstick and FACE models showed better fit for length data compared to Laird-Ware and SITAR.
• Concerns raised regarding the timescale of growth faltering detection and its relevance to long-term health outcomes.
• Poor reporting quality in fetal centile studies identified; GAMLSS software demonstrated power and potential for misuse.
• Longitudinal fetal head circumference centiles were found to be biased.

Conclusions:

• The choice of growth model depends on whether short-term fluctuations or long-term growth summaries are prioritized.
• Recommendations for good practice in fetal centile studies are provided to improve data quality and reporting.
• Further research is needed to fully understand the predictive value of different growth models for later health outcomes.