Graduation Year


Document Type




Degree Granting Department

Public Health

Major Professor

Russell S. Kirby


adverse childhood experiences, low birth weight, preterm birth, Social position, structural equation models


BACKGROUND: The social causation of preterm birth remains elusive, without an adequate explanatory framework. Thus, this study proposed and evaluated a conceptual model of the social determinants of perinatal health for the understanding of perinatal health disparities.

METHODS: A prospective cohort study was conducted with pregnant women between 20 and 35 weeks gestation who were participating in two Healthy Start programs in Central Florida, from July 2011-August 2013. Perinatal health was operationalized based on gestational age, birth weight, and healthy start infant risk screen score. The predictors were: early life adversity, social position, maternal health-related quality of life, maternal stress, racism and discrimination, lack of social support, father involvement during pregnancy, intimate partner violence, and adverse maternal behaviors. Data collection consisted of a self-administered survey and birth outcome data was obtained from Healthy Start administrative databases. The statistical framework was structural equation modeling.

RESULTS: The study sample was racially and ethnically diverse (N, Hispanics=72; N, non-Hispanic blacks=61; and N, non-Hispanic whites=48). The majority of mothers in this study were single or not married (cumulative 76%), US born (74.6%), and with English speaking preference (74.6%). The sample tended to cluster in low income groups (cumulative 58% less than $25,000 annual household income) and with education levels of less than high school (79.6%). A greater proportion of Hispanic mothers were married (66.7%) compared to non-Hispanic blacks (34.4%) and non-Hispanic whites (47.9%). Only 41.7% had completed high school, compared to 63.9% non-Hispanic blacks and 64.6% non-Hispanic whites. Nearly all non-Hispanic blacks and non-Hispanic whites were born in the US, compared to only 43.1% Hispanic mothers. Only 40% of non-Hispanic blacks reported on currently living with the baby's father at the time of the survey, compared to 66.2% for Hispanic mothers, and 58.3% for non-Hispanic whites. Furthermore, non-Hispanic blacks reported a greater proportion of discriminatory experiences in daily situations (mean = 4.74), compared to the other groups (mean for Hispanics was 2.14, and mean for non-Hispanic whites was 1.95). Non-Hispanic whites reported the greater proportion of daily alcohol use (mean 3.8 beverages per month), compared to other groups (Hispanic mean was 0.69, and non-Hispanic blacks mean was 1.68). Non-Hispanic white mothers also presented a higher mean of adverse childhood experiences before 18 years of life (mean = 3.4), compared to other groups (mean for Hispanics was 1.63, mean for non-Hispanic blacks was 2.48). With the exception of the confirmatory factor analysis for intimate partner violence (low correlations with common factor), all other confirmatory factor analyses demonstrated an acceptable Chi-square to degrees of freedom ratio (<6), and the RMSEA was less than 0.08 (minimum for acceptance). Thus, structural equation models were estimated subsequently. The first model was a model of direct effects between social position and perinatal health (hypothesis 1: direct effects), which demonstrated a good fit as indicated by X2/DF ratio of 1.4 (Chi-Square = 19, DF =13) and a RMSEA of 0.05. However, the direct effect of social position was very small and non-significant (Beta=-.02, p-value =.76), supporting the conclusion that a simple direct effect of social position on perinatal health was not found in this population. The second model explored indirect effects of social position through intermediate factors (hypothesis 2: indirect effects), which demonstrated a good fit to the data, as indicated by a Chi-square/df ratio = 1.45 and RMSEA=.05. Social support was a statistically significant mediator between social position (Beta=0.284, p<0.05) and perinatal health (Beta=0.22, p<0.05). The third model incorporated adverse childhood experiences as predictor of social position effects. Adverse childhood experiences were significantly associated with social position (Beta=.363, p<0.05) and moderated the effects of social position on social support and perinatal health. In the presence of adverse childhood experiences, the social position was significantly associated to maternal health-related quality of life (Beta=-0.226, p<0.05) and maladaptive maternal behaviors (Beta=0.654, p<0.05).

CONCLUSION: This study demonstrated synergistic effects of social determinants of health. Controlling for all factors considered, social support was significantly associated with perinatal health, which presents implications for strengthening prenatal programs that provide support to pregnant women. Findings need to be replicated in larger studies with the US general population. Policy makers and researchers need to pay greater attention to the role of early life adversity on perinatal health outcomes.