Date of Award


Document Type

Open Access Thesis


Epidemiology and Biostatistics


The Norman J. Arnold School of Public Health

First Advisor

Jihong Liu


Objective: Excessive and inadequate gestational weight gain (GWG) are associated with a myriad of adverse pregnancy outcomes. The objective of this study was to examine secular trends in GWG in South Carolina (SC) from 2004 to 2015. We hypothesized that there was a trend of higher GWG z scores (GWGZ) over this 11-year time period after adjusting for changes in population characteristics. We also hypothesize that a trend of higher mean GWGZ at higher percentiles is more apparent for African Americans, rural women, and women who are overweight or obese before pregnancy.

Data and Methods: Data came from SC 2004-2015 birth certificates (n = 525,411). In this study, we used gestational-age-standardized GWGZ, which were calculated using smoothed reference values for GWG to account for gestational age and pre-pregnancy BMI. Quantile regression models were used to understand GWGZ trends over time at different percentiles of GWGZ, adjusting for important maternal characteristics. We further evaluated the modifying effects of maternal race, maternal pregnancy BMI and rural/urban residence on the GWG trends.

Results: SC women had an overall mean GWGZ of -0.40. We saw an overall increase in GWGZ in the 5th and 10th percentiles and an overall decrease in the 25th, 50th, 75th, 90th, and 95th percentiles (2004-2015). White women had a higher increase in GWGZ in the 5th percentile compared to African American women. In the 95th percentile, White women had a higher

decrease in GWGZ compared to African American women. Rural residents had significantly lower GWGZ in the 5th percentile and modest increases in GWGZ compared to urban residents. Lastly, underweight women showed decreases in the 5th percentile of GWGZ and overweight and obese women showed negligible changes in GWGZ in the 90th and 95th percentiles.

Implication and significance: Knowledge about the trends of GWG in SC women and its correlates is helpful for addressing health concerns of high risk populations, such as racial minority groups and women who are underweight or overweight before pregnancy. GWGZ scores as a GWG measure and quantile regression models are feasible for trend analyses of GWG.


© 2017, Marilyn Elizabeth Wende

Included in

Epidemiology Commons