Dissertations and Theses

Date of Degree


Document Type


Degree Name

Doctor of Philosophy (Ph.D.)


Epidemiology and Biostatistics


Denis Nash, PhD

Committee Members

Dr. Luisa N. Borrell

Mary Huynh, PhD

Dr. Elizabeth A Kelly

Subject Categories

Epidemiology | Public Health


Obstetric Care, Cesarean Delivery, Racial and Ethnic Disparities, New York City


BACKGROUND: While Cesarean delivery is a life-saving procedure when certain complications arise, it is associated with increased risks of maternal mortality and morbidity, as well as neonatal and childhood morbidities, and increased risks for women during subsequent pregnancies. Stark and persistent racial/ethnic disparities in Cesarean delivery that are not explained by clinical risks raise concerns about overuse of the procedure, as well as the contribution of potentially avoidable Cesareans to disparities in maternal mortality and morbidity. Understanding the extent to which disparities in Cesarean delivery may be attributable to differences in care during labor is critical for addressing these disparities. Accordingly, this dissertation focused on examining obstetric care among women with low-risk pregnancies in New York City (NYC) with the aim of identifying potentially modifiable factors contributing to racial/ethnic disparities in Cesarean delivery.

METHODS: The dissertation comprised three studies that drew on a population-based, cross-sectional birth cohort, using birth certificate records for NYC, for the period 2012-2017. The study was restricted to women with low-risk pregnancies, defined as nulliparous singleton, vertex-presentation (NTSV) pregnancies with no contraindications to vaginal delivery at term (≥37-41 weeks’ gestation) giving birth in NYC hospitals. Deliveries to women with prior births were excluded, along with non-singleton pregnancies, and stillbirths. For Aim 1, controlled interrupted time series (CITS) analysis was used to describe and quantify the effect of the 2014 consensus statement by the American College of Obstetricians and Gynecologists (ACOG) and the Society for Maternal-Fetal Medicine (SMFM) on the prevention of avoidable primary Cesareans on age-standardized Cesarean rates among women with NTSV deliveries. For Aim 2, causal mediation analysis was used to examine the extent to which racial/ethnic disparities in NTSV Cesareans were mediated through non-absolute indications for Cesarean delivery and non-medically indicated induction of labor (IO) at 39 weeks’ gestation. For Aim 3, multilevel logistic regression was used to quantify between-hospital variation in Cesarean delivery during 2015-2017, overall and by maternal race/ethnicity, and to examine the association between hospital characteristics (e.g., public vs. private ownership, delivery caseload, interventionist approach to delivery care, and staffing by licensed midwives) and Cesarean delivery, overall and by maternal race/ethnicity.

RESULTS: Our CITS analyses showed that the 2014 ACOG-SMFM recommendations on reducing unnecessary primary Cesareans led to a small but significant decrease in Cesarean delivery (risk ratio: 0.94; 95% CI: 0.92-0.97), with no change in the control series of nulliparous women with pre-term, non-singleton pregnancies or non-vertex presentation births. The effect of the ACOG-SMFM recommendations differed by maternal race/ethnicity, with decreases in age-standardized NTSV Cesarean rates observed among non-Hispanic White women (risk ratio: 0.93; 95% CI: 0.88-0.98) and Hispanic women (risk ratio: 0.92; 95% CI: 0.87-0.97), and no change among non-Hispanic Black or Asian women. This analysis also showed significant decreases in Cesareans for dystocia indications, a key focus of the ACOG-SMFM recommendations. Decreases in Cesareans were observed at minority- and non-minority-serving hospitals, however, the overall decrease at minority-serving hospitals was of marginal statistical significance (p=0.048), and at non-minority-serving hospitals, statistically significant decreases in Cesarean rates were observed only among non-Hispanic White women.

Mediation analyses (Aim 2) showed that non-Hispanic Black, Asian and Hispanic women were significantly more likely than non-Hispanic White women to be diagnosed with non-absolute Cesarean indications, and these differences explained a substantial proportion of excess Cesarean risks—from approximately 14% among non-Hispanic Black women to 50% among Asian and Hispanic women. Labor induction with no medical indication at 39 weeks was rare, and it was slightly more prevalent among non-Hispanic Black, Asian and Hispanic women compared with non-Hispanic White women. While it was not significantly associated with Cesarean delivery among non-Hispanic White women, non-medically indicated IOL at 39 weeks was associated with reduced risks of Cesarean delivery among non-White women, and our mediation analyses suggested that eliminating differences in non-medically indicated IOL would increase racial/ethnic disparities in Cesarean delivery by approximately 4% among non-Hispanic Black women and 9% among Asian and Hispanic women.

In Aim 3, among 127,449 women with NTSV deliveries between 2015 and 2017, multilevel logistic regression analyses showed that general contextual effects were small, indicating little between-hospital variation after adjustment for maternal and hospital characteristics. The variation explained by hospital of delivery and hospital characteristics associated with Cesarean delivery differed by maternal race/ethnicity, with delivery in a teaching hospital reducing odds of Cesarean delivery among non-Hispanic White and Asian women, but not among non-Hispanic Black or Hispanic women. Other hospital characteristics (e.g., public vs. private ownership, delivery caseload, interventionist approach to delivery care, and staffing by licensed midwives) were not consistently associated with Cesarean delivery.

DISCUSSION: Our findings related to decreases in age-standardized NTSV Cesarean rates, overall, and in Cesareans for dystocia indications provide encouraging evidence that primary Cesarean rates may be modifiable through changes in labor management strategies, such as those recommended by the ACOG and SMFM. The results from causal mediation analyses provide further support for the ACOG/SMFM recommendations for reducing unnecessary primary Cesareans—and particularly Cesareans for non-absolute indications, as such indications appear to contribute to excess Cesarean risks among non-Hispanic Black, Asian and Hispanic women. At the same time, however, the lack of change in overall Cesarean rates among non-Hispanic Black women after the 2014 recommendations, as well as the lack of any decrease in Cesareans for dystocia indications among these women, raise concerns about differential care by maternal race/ethnicity. Similarly, findings from our our multilevel regression analysis that delivery in a teaching hospital reduces Cesarean risks for non-Hispanic White and Asian women, but not for Black or Hispanic women suggest that the provision of obstetric care differs by maternal race/ethnicity. As provider practices during labor and delivery are potentially modifiable through training and other hospital-level interventions, further examination of racial/ethnic disparities in obstetric care is critical for advancing birth equity and redressing persistent disparities in pregnancy outcomes.

Available for download on Thursday, September 22, 2022

Included in

Epidemiology Commons