Update: The paper is now published here.
The working paper I wrote about here has now been accepted for publication in Sociological Science. Although the results haven’t changed substantially, I revised it since the last post, so you should use this copy instead. Here’s the abstract:
Maternal Age and Infant Mortality for White, Black, and Mexican Mothers in the United States
This paper assesses the pattern of infant mortality by maternal age for White, Black, and Mexican mothers, using 2013 Period Linked Birth/Infant Death Public Use File from the Centers for Disease Control. The results are consistent with the “weathering” hypothesis, which suggests that White women benefit from delayed childbearing while for Black women early childbearing is adaptive because of deteriorating health status through the childbearing years. For White women, the risk (adjusted for covariates) of infant death is U-shaped – lowest in the early thirties – while for Black women the risk increases linearly with age. Mexican-origin women show a J-shape, with highest risk at the oldest ages. The results underscore the need for understanding the relationship between maternal age and infant mortality in the context of unequal health unequal health experiences across race/ethnic groups in the U.S.
Family Inequality 
Is paternal age an issue here, and, if so, did you control for it? I don’t see a reference to fathers, though I might have missed it.
LikeLike
Two points ( I am asking to understand, and not to argue):
1. What is the mean, standard deviation of the black maternal samples, vis-à-vis Hispanic and white samples. Based on eyeballing, black maternal death samples seem to lie within 1 standard deviation for all ages, but the white and Hispanic samples do not. How do you calculate the multivariable risk in Appendix Table 1? Statistically, how does one say “the risk increases linearly with age”. More precisely, what would the % change in risk in the appendix Table 1 that will be acceptable, 15%?
2. The preterm-related infant mortality rate among Non-Hispanic black infants (6.01 per 1,000 live births) more than triple the preterm-related infant mortality rate among Non-Hispanic white infants (1.79 per 1,000). Preterm, or premature, delivery is the most frequent cause of infant mortality, accounting for more than one third of all infant deaths during the first year of life. as compared with age differences, the pre-term birth->infant mortality dominates racial differences in infant mortality[“http://www.cdc.gov/media/subtopic/matte/pdf/CDCMatteReleaseInfantMortality.pdf”]
This result is directly correlated to genetics (Too numerous to mention, e.g. see below), and to a lesser extent, environmental effects such as lead (Burris et al). and I was surprised that there is no reference to the cause.
A large number of genetic studies sponsored by NIH on preterm birth->infant mortality of African American children. See the page under title “CANDIDATE GENES AND INCREASED RISK FOR PREMATURE DELIVERY IN AFRICAN AMERICANS” in Pediatric Research (2009) 65, 1–9.
Sorry, and not to waste time, but the impact of genetics and environment dominate age differences in mortality in child birth for black mothers, and policy recommendations should be focused on full-term child birth for all ages for black mothers, and not be focused on showing that age differences upon infant mortality by race.
LikeLike
Lots of good things to research – no reason not to study genetics and maternal age. The multivariate risk is calculated using logistic regression with all variables included. The linear effect of age was from an identical model using age as a continuous variable instead of the categories.
LikeLike
Thanks but I still think you should collaborate with someone in pediatrics or neonatology, rather than stay within sociology.
LikeLike
There is work on this question using African immigrant populations that suggests that social factors are more important than genetic factors: e.g. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1913086/
LikeLiked by 1 person
The immigrant health paradox has been discussed by Dr. Cohen in https://familyinequality.wordpress.com/2014/08/31/immigrant-health-paradox-update/. The immigrants are a self selected population, and are not representative samples of native populations. The conclusion was “…there is no simple explanation for Latinos’ perplexing health outcomes, such as simply that healthier people migrate. Rather, migrants are positively selected in some health aspects, negatively selected in others, and in yet other health outcomes, there is no selection effect. In sum, selective migration plays a role in explaining some of U.S. Latinos’ health outcomes,——-“. Social vs.genetic does not adequately describe the immigrant health paradox.
LikeLike