Estimating differences in birtweight outcomes using panel data approach (Umbria 2005-2008)

Background: The main part of international literature finds a significant increasing of birthweight in second-borns in comparison with first-borns. We assume a significant effect of some physical maternal characteristics that change between the first delivery and the other ones, but also the possible effect due to the change of social maternal characteristics could result in a significant impact. Objectives: We investigate about the differences in birthweight between first- and second-borns, evaluating the impact of changes in physical (e.g., gestational age, chronological age) and social (e.g, education level, marital status) maternal characteristics. Methods: All analyses are performed on data collected in Umbria (Italy) taking into account a set of 792 women who delivered twice from 2005 to 2008.Firstly, we use univariate paired t-tests for the comparison between weights of first- and second-borns. Secondly, we investigate the topic through linear and nonlinear regression approaches in order to evaluate the effect of physical and social maternal characteristicsto predict the odds-ratio of underweighted infants, respectively. We consider (i) the difference in birthweight between the first two deliveries as a quantitative response variable in a linear regression model and (ii) the birthweight of first- and second-borns as two categorical response variables in a marginal model. Results: We find that the birthweight of second-borns is significantly higher than first-borns ones (t-value = 4.83, p-value < 0.0001; association in marginal model = 2.13, p-value < 0.0001). Statistically significant effects are related with a longer gestational age, an increased number of visits during the pregnancy, and the gender of infants. More precisely, a longer gestational age at the second delivery has a positive effect on the second-born's weight: it shows an average increase of 153.72 gr (95% CI = 140.69-166.76, p-value < 0.0001) for each additional gestational week. Moreover, we do not observe any significant effect related with the mother's age (95% CI = -0.07-0.17, p-value = 0.4420) and with social characteristics. Conclusion: The analyses allow us to conclude for a significant increasing of birthweight between first- and second-borns for a representative sample of Italian population. The main part of the observed differences depends on differences in gestational ages. For further developments, we intend to extend our study to include the effect of changes in mother's weight