Related Experiment Video
Updated: Aug 28, 2025

Assessment of Child Anthropometry in a Large Epidemiologic Study
Published on: February 2, 2017
Assessment of height growth in Indian children using growth centiles and growth curves
Sandra Aravind Areekal1, Pranay Goel1, Anuradha Khadilkar2
1Department of Biology, Indian Institute of Science Education and Research, Pune, India.
Insights
This study compared height growth centiles and curves in Indian children using GAMLSS and SITAR methods. SITAR effectively summarized individual growth patterns, while GAMLSS showed population height distributions.
Area of Science:
- Pediatric endocrinology
- Biostatistics
- Human growth and development
Background:
- Growth centiles and curves are standard methods for presenting child anthropometry.
- These methods differ in data requirements, analytical approaches, biological parameters, and interpretation.
Purpose of the Study:
- To fit and compare height growth centiles and curves in Indian children.
- To evaluate the utility of GAMLSS and SITAR models for analyzing pediatric growth data.
Main Methods:
- Longitudinal height data from 1468 Indian children (aged 6-18 years) were analyzed.
- Generalised Additive Models for Location Scale and Shape (GAMLSS) were used for growth centiles.
- SuperImposition by Rotation and Translation (SITAR) was employed for growth curves.
Main Results:
- SITAR explained over 98% of height variance in boys and girls, accurately identifying peak height velocity.
- GAMLSS (Box-Cox Cole Green model) also identified pubertal growth spurts, but produced shallower centiles compared to SITAR.
- Boys exhibited a mid-growth spurt around age 8 years.
Conclusions:
- GAMLSS provides population-level height distributions by age and sex.
- SITAR offers a parsimonious summary of individual child height growth patterns.
- Both GAMLSS and SITAR yield distinct and valuable insights into child growth.
Background:
Growth centiles and growth curves are two ways to present child anthropometry; however, they differ in the type of data used, the method of analysis, the biological parameters fitted and the form of interpretation.
Aim:
To fit and compare height growth centiles and curves in Indian children.
Subjects And Methods:
1468 children (796 boys) from Pune India aged 6-18 years with longitudinal data on age and height (n = 7781) were analysed using GAMLSS (Generalised Additive Models for Location Scale and Shape) for growth centiles, and SITAR (SuperImposition by Rotation and Translation) for growth curves.
Results:
SITAR explained 98.7% and 98.8% of the height variance in boys and girls, with mean age at peak height velocity 13.1 and 11.0 years, and mean peak velocity 9.0 and 8.0 cm/year, respectively. GAMLSS (Box-Cox Cole Green model) also captured the pubertal growth spurt but the centiles were shallower than the SITAR mean curve. Boys showed a mid-growth spurt at age 8 years.
Conclusion:
GAMLSS displays the distribution of height in the population by age and sex, while SITAR effectively and parsimoniously summarises the pattern of height growth in individual children. The two approaches provide distinct, useful information about child growth.
Related Concept Videos
Variation: Normal Distribution, Range, and Standard Deviation
Nature and Nurture
Applications of Normal Distribution
The heights of 15 to 18-year-old males from Chile from 1984 to 1985 followed a normal distribution. The mean height is 172.36...
Polygenic Traits
Microbial Growth Measurement: Indirect Methods
Signs of Puberty

