Date of Award

Fall 2018

Document Type

Open Access Thesis


Epidemiology and Biostatistics

First Advisor

Jihong Liu


Physical activity is known to be beneficial to adults in general, including pregnant women, but the literature on physical activity and neonatal outcomes has been inconsistent. Few studies have differentiated between physical activity prior to pregnancy and during pregnancy to see whether the typical beneficial effect is not due to having a more active lifestyle in general. The Pregnancy Risk Assessment Monitoring System (PRAMS) is a national cross-sectional surveillance system. Our data from SC PRAMS included 6391 respondents and was weighted to be representative of population of pregnant women that delivered in the state. We excluded births that were not singleton (n=624), births that were not viable (n=240), women with pre-existing diabetes (n=213) and women who had missing responses for our covariates(n=27), leaving a total sample of 5294 mother-neonate pairs. Database management and statistical analysis were carried out using SAS software version 9.4 (SAS Institute, Cary, NC). Our main analysis was carried out using logistic regression. All results were weighted to be representative of all live births from 2009-2015 in South Carolina and were considered significant at p ≤ 0.05. We found that women who were previously inactive but became active during pregnancy had reduced odds of preterm birth (PTB) (aOR=0.54 (95% CI, 0.30-0.98)) compared with completely inactive women. We also found that women who did more exercise and for longer had lower odds of PTB. Conversely, we found women who had been active but stopped vi activity had higher odds of PTB (aOR=1.5 (95% CI, 1.07-2.10)) and low birth weight (aOR=1.30 (95% CI, 1.03-1.62) than women who were inactive. This analysis allowed a complex conceptualization of physical activity in a state representative population and demonstrated a dose response between PA and PTB but was subject to recall bias and did not have information on disability. In the future, prospective data collection should be used and exercise should be measured objectively.


© 2018, Andrew Tyler Broadway

Included in

Epidemiology Commons