Does gestational age misclassification explain the difference in birthweights for Australian aborigines and whites?

Abstract

After 34 weeks gestation, summary measures of location for birthweight (e.g means and centiles) increase more slowly for Australian Aborigines than for whites. A similar pattern has been observed for blacks in the US. This study tests whether the reported pattern is due to differential misclassification of gestational age. Simulation was used to measure the potential effect of differential misclassification of gestational age. Reported gestational age data were obtained from Queensland Perinatal Data Collection (QPDC). Estimates of the true distributions of gestational age were obtained by assuming various (plausible) types of misclassification and applying these to the reported distributions. Previous studies and data from the QPDC were used to help specify the birthweight distributions used in the simulations. At full term, the parameters of the birthweight distributions were robust to gestational age misclassification. At preterm, the 10th centiles were robust to misclassification. In contrast, the 90th centiles were sensitive to even minor misclassification. Extreme types of misclassification were required to remove the divergence in median birthweights for Aborigines and whites. Gestational age misclassification is an unlikely explanation for the reported divergence in average birth-weights for Aborigines and whites. The results might help with the interpretation of other between-population comparisons.