In a recent study, researchers examined the feasibility of generating new pediatric growth charts using a novel big data approach.
Pediatric growth charts are clinical tools used by pediatricians and general practitioners to monitor the growth of infants and children over time. The charts identify serious health problems early, or in time, to provide the necessary care and thus reduce associated risks. Growth monitoring involves the routine collection of children’s’ body measurements, such as height and weight. Physicians compare the observed values to the standardized information that is established in the charts. In many countries, growth charts are developed, based on the standards or references that are set by the World Health Organization (WHO) and national healthcare organizations. Recent studies suggest the current national and WHO growth charts are not perfectly calibrated with the growth of children in several countries.
A study, published in The Lancet Digital Health, investigated the feasibility of producing new pediatric growth charts, using a novel big data approach. The big data approach involved examining large datasets of growth measurements that were routinely collected by physicians and stored in electronic medical records databases.
In this study, researchers recruited 32 pediatricians and general practitioners from across France. The participating physicians used the same electronic medical records database from which the researchers extracted all the physical growth measurements of children. The study included only children born from 1990 and aged one month to 18 years by 2018, with a birth weight that was greater than 2.5kg. Data cleaning was applied to the datasets to find and remove any measurement or transcription errors. After data cleaning, researchers included over three million height and weight measurements from more than 230,000 children and applied modelling techniques to the data to generate the new growth charts. The new charts were compared with the existing national and WHO charts and were validated using data from national cross-sectional surveys.
The study reports the new growth charts display height and weight percentile curves that are higher than the national and WHO charts. The findings suggest that the new pediatric growth charts are better calibrated with the growth of children in comparison to the national and WHO charts. According to the researchers, the big data approach will reduce cost and time for the collection and analyses of huge datasets of physical growth measurements and allow for an optimal calibration between the growth charts and the population to which they will be applied.
The study demonstrated the feasibility of generating new pediatric growth charts, using the big data approach, with success. Based on the study findings, the researchers suggest the approach can be applied in other countries and medical fields as long as the related physical growth data are stored in electronic medical records databases from which it can be easily obtained.
Written by Ranjani Sabarinathan, MSc
Reference: Heude B, Scherdel P, Werner A, et al. (2019). A big-data approach to producing descriptive anthropometric references: a feasibility and validation study of paediatric growth charts. The Lancet Digital Health. doi: 10.1016/S2589-7500(19)30149-9
Image by sathyatripodi from Pixabay