New method for predicting adult height

Canva
Date of publication:
Researchers from Jožef Stefan Institute and Faculty of Sport, University of Ljubljana, developed a new method for predicting adult height of children and adolescents. The method uses large population data collected in the framework of the SLOfit program over decades. It compares the growth curve of a child with those of the most similar individuals and substantially outperforms the existing height predicting methods.
“How tall will my children be when they grow up?“ This is a question of several parents. On one hand, the reason can be medical. If a child is not growing as fast as they should, according to the percentile tables that pediatricians use, perhaps a hormonal treatment is needed. On the other hand, height plays an important role in sport talent identification. A kid that is talented for basketball stands next to zero chance of becoming the next Luka Dončić with the adult height of 170 cm but may instead become Primož Roglič or Timi Zajc.
The growth pattern of a child can often be inferred from the growth curves of both parents, but these are typically not readily available. The most common approach the pediatricians use are the abovementioned percentile tables where the doctor checks whether the child is staying the same percentile range through the years. However, these tables do not take into account the fact that the puberty-related growth spurt does not take place for everyone at the same time. Other methods for adult height estimation have been developed by several groups but they rely on additional body size measurements, including invasive radiological scans, or the researchers used only about a hundred individuals in the analysis.
Taking advantage of the massive dataset collected within the SLOfit program and containing annual records of height, weight, and several fitness-related parameters of schoolchildren, Slovenian researchers developed a new algorithm for adult height prediction. The SLOfit records span from the time children enter the primary school at the age of 6 or 7 until they finish the secondary school at the age of 18 or 19, and has been covering the entire generation of Slovenian children since the 1980s. The new algorithm compares the growth curve of a child with the most similar individuals in the dataset, based on height, and uses this information to predict the future growth curve and adult height. The method substantially outperforms the existing methods in prediction strength, partially owing to the fact that it was developed using data of over 16 000 schoolchildren.
The method has now been successfully integrated in the SLOfit website where it is publicly available, with the study being published in the scientific journal PLOS One. The research was carried out in the framework of a European Horizon 2020 project CrowdHEALTH. In 2023, the follow-up European project SmartCHANGE started, where the forecasting of fitness parameters based on the SLOfit data will be used to improve the physical fitness of schoolchildren and adolescents.
Paper: Mlakar, M., Gradišek, A., Luštrek, M., Jurak, G., Sorić, M., Leskošek, B., & Starc, G. (2023). Adult height prediction using the growth curve comparison method. PloS one, 18(2), e0281960. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0281960