Research Day


Document Type





CONTEXT: Within Kalamazoo County, as across the United States, white infants die at markedly lower rates than infants of color. Not only they die at different rates, but they die from different causes: leading cause of death among higher-income white infants is congenital anomalies compared to prematurity, which is the leading cause of death among higher-income infants of color. This study is unique in examining whether infants from different racial and socioeconomic groups have different factors which impact prenatal health, birth outcomes and infant survival.

OBJECTIVE: The goal of the current study was to examine whether variations in infant health by race and by income were associated with different sets of predictors, including maternal demographics, health behaviors, obstetric history, maternal health condition, prenatal care.

STUDY DESIGN: Population-based cross-sectional study using secondary analysis of infant birth and death records. Sampling method was census, and the study sample consisted all infants born to Kalamazoo County residents during the study period, 2008 through 2014 (N=21,858). The study outcome was infant health, as defined by full-term gestation (>37 weeks), adequate birthweight (>2,500 grams) and infant survival to first birthday. Predictors included maternal demographics, health factors and prenatal care. Logistic regression models, stratified by race (of color, white) and income (Medicaid, private insurance), were conducted with two-sided statistical significance set at α<.05. Each predictor was tested through Baron and Kenny’s process of mediation.

RESULTS: Eighty seven percent (87.2%, 18,783 of 21,858 births) resulted in a full term, adequate weight infant who survived its first year. Regarding health risk and protective factors, white infants were 2.7 times more likely to be higher-income than infants of color (62.3% and 23.1%, respectively). Higher-income women had fewer risk factors (adolescent pregnancy, prior poor birth outcome, chronic disease, late prenatal care, STI, prenatal smoking), except for infertility treatment. Stratification revealed important differences regarding determinants of infant health based upon race and income. To begin with, higher income status was protective for white infants but not for infants of color (income: white aOR 1.34 (1.18, 1.52) / of color aOR 1.09 (0.87, 1.38). Additionally:

  • Having a prior poor birth outcome, the greatest risk to infant health overall, was more prevalent and was associated with greater risk to infants of color and to low-income infants
  • First trimester prenatal care, a protective factor, was more prevalent and brought significant health gains to white infants and to higher income infants but not to infants of color or low-income infants

Multivariable logistic regression confirmed that, even after accounting for key contributors, race and income were significant independent predictors of infant health: white infants and higher-income infants had better health than infants of color and lower-income infants (race aOR 1.41 (1.27, 1.55) and income aOR 1.25 (1.12, 1.39)). Goodness of fit statistics were 0.6853 to 0.6972 (AUC) for the final four stratified models. A race-by-income interaction was tested, but did not improve model fit.

CONCLUSION: Infants of color face different health risks than white infants; risks that vary not just in magnitude but in the character. The same is true for low- and high-income infants. This points to the need for tailored approaches to risk assessment, clinical care and public health interventions; such as patient centered clinical care models at the individual level and targeted population-level interventions which structure resources to meet the nature and degree of risk specific to a group.

This document is currently not available here.