A multivariate growth curve model for pregnancy.

A model is proposed for consecutive multivariate observations of pregnancy-related quantities for a particular mother assuming proportionally with an underlying univariate growth process. Apart from the longitudinal aspect of the multivariate time series, the cross-sectional distribution of growth curves is also discussed. The model is examined as regards problems of estimation and model checking, and the results are applied to data consisting of bivariate time series of symphyseal fundal distances and fetal weight estimates from 91 mothers. We find some evidence in the data on departures from the model.