Istituto di Matematica Applicata e Tecnologie Informatiche     
Tentoni S., Astolfi P., De Pasquale A., Zonta L. Adjustment of fetal ponderal growth curves according to a mixture model. Preprint ercim.cnr.ian//2001-1209, 2001.
In the distribution of birthweight by gestation age (GA) several authors have observed an excess of babies large for GA, suggesting that pregnancy length might be, in many cases, underestimated by the reported last menstrual period (LMP). In a sample of 2,723,453 single liveborns, selected from the 1990-94 newborns in Italy (ISTAT), we observed positively skewed weight distributions in GA classes <35 weeks, markedly bimodal at 30-32 weeks. This suggests that the observations by GA might be not really homogeneous, and that large babies might follow a secondary birthweight distribution shifted to the right tail. In order to define reliable fetal ponderal growth curves, we have modeled the data by a two-component normal mixture distribution, where the major component accounts for the normal process of growth while the secondary one accounts for large babies. The mixture model provides a more robust way to correct the growth centile curves for large babies. For each GA, the model parameters have been estimated, as well as the threshold values for a probabilistic clustering of the observations. The secondary component ranges from ~7% to ~27% of the total distribution, with a maximum at the 30th week, and then vanishes by the 35th week. The values of its parameters suggest that in large-for-GA babies the pregnancy duration might be underestimated of about six weeks.
Subject Fetal growth
Birth weight
Gestational age
Growth centile curves
Mixture models
Probabilistic clustering

Icona documento 1) Download Document PS

Icona documento Open access Icona documento Restricted Icona documento Private


Per ulteriori informazioni, contattare: Librarian http://puma.isti.cnr.it

Valid HTML 4.0 Transitional