Infant Mortality by Month of Birth: An Analysis of Contemporary Cohorts

Date

2011-08-08

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

There is a well-established connection between adult mortality and the conditions an individual is exposed to while in utero. There is a wealth of research that connects conditions such as asthma and allergies, mortality due to heart disease and diagnoses of schizophrenia to conditions during an individual?s early life and even their time in utero. The aim of this dissertation is to see if this same connection can be made to infant mortality, and further will there be any connection in contemporary cohorts? I use the Linked Birth/Infant Death dataset available from the Centers for Disease Control (CDC) for the years 2000 to 2004. This dissertation specifically uses the dependent variable "cause specific infant death" with various measures of the time the infant was born or was in utero. I undertake three multinomial logistic regression models with the dependent variable "cause specific infant death." I then proceed to a multilevel multinomial logistic regression model using state-level climate measures at the second level. I conclude with the construction of maps displaying the spatial relationship between infant mortality and climate. The first analysis uses the independent variable of interest "month of birth," the second analysis uses the independent variable of interest "months of first trimester," and the last level-one analysis uses the independent variable of interest "months of third trimester." After running all three models, I determined that the most effective independent variable of interest is "month of birth," which I use in a multilevel logistic regression model. The multilevel model uses the month of birth variable at level-one and incorporates state level measures of climate at the second level. I find that the humidity index and the temperature index are negatively associated with the month of birth variable and cause specific infant death variables, meaning that the higher these indices, the more the benefit to an infant's chances of survival. The wind index is consistently positive, meaning that the interaction of wind with cause specific infant death and month of birth is detrimental to an infant's survival. The last methods chapter shows the spatial relationship between infant mortality and climate. In this chapter I find that infant mortality in the United States is concentrated in the Southern U.S., which is also where there is a concentration of high temperature states. The connections between wind and humidity with the infant mortality rate are less consistent.

Description

Citation