Abstract
Objective
To examine whether racial and ethnic differences exist in the frequency of and indications for cesarean delivery and to assess whether application of labor management strategies intended to reduce cesarean delivery rates is associated with patient’s race and ethnicity.
Methods
This is a secondary analysis of a multicenter observational obstetric cohort. Trained research personnel abstracted maternal and neonatal records of >115,000 pregnant women from 25 hospitals (2008–2011). Term women with singleton, nonanomalous, vertex, liveborn neonates were included in two cohorts: 1) nulliparous women (n=35,529) and 2) multiparous women with prior vaginal deliveries only (n=39,871). Women were grouped as non-Hispanic black, non-Hispanic white, Hispanic, and Asian. Multivariable logistic regression was used to evaluate the following outcomes: overall cesarean delivery frequency, indications for cesarean delivery, and utilization of labor management strategies intended to safely reduce cesarean delivery.
Results
A total of 75,400 women were eligible for inclusion, of whom 47% (N=35,529) were in the nulliparous cohort and 53% (N=39,871) in the multiparous cohort. The frequencies of cesarean delivery were 25.8% among nulliparous women and 6.0% among multiparous women. For nulliparous women, the unadjusted cesarean frequencies were 25.0%, 28.3%, 28.7%, and 24.0% for non-Hispanic white, non-Hispanic black, Asian, and Hispanic women, respectively. Among nulliparous women, the adjusted odds of cesarean were higher in all racial and ethnic groups compared with non-Hispanic white women (non-Hispanic black aOR 1.47, 95% CI 1.36–1.59; Asian aOR 1.26, 95% CI 1.14–1.40; Hispanic aOR 1.17, 95% CI 1.07–1.27), due to greater odds of cesarean delivery both for nonreassuring fetal status and labor dystocia. Nonapplication of labor management strategies regarding failed induction, arrest of dilation, arrest of descent, or cervical ripening did not contribute to increased odds of cesarean delivery for non-Hispanic black and Hispanic women. Compared with non-Hispanic white, Hispanic women were actually less likely to experience elective cesarean (aOR 0.60, 95% CI 0.42–0.87) or cesarean for arrest of dilation prior to 4 hours (aOR 0.67, 95% CI 0.49–0.92). Additionally, compared with non-Hispanic white, Asian women were more likely to experience cesarean for nonreassuring fetal status (aOR 1.29, 95% CI 1.09–1.53) and to have had that cesarean be performed in the setting of a 1-minute Apgar score ≥7 (aOR 1.79, 95% CI 1.07–3.00). A similar trend was seen among multiparous women with prior vaginal deliveries.
Conclusions
Although racial and ethnic disparities exist in the frequency of cesarean delivery, differential use of labor management strategies intended to reduce the cesarean delivery rate does not appear to be associated with these racial–ethnic disparities.
INTRODUCTION
In 2012, the Society for Maternal-Fetal Medicine (SMFM), Eunice Kennedy Shriver National Institute of Child Health and Human Development, and American College of Obstetricians and Gynecologists (the College) convened to review evidence for labor management strategies designed to safely reduce the rate of cesarean delivery.1 This was followed by publication of Obstetric Care Consensus on prevention of primary cesarean delivery, which included guidelines regarding labor management best practices.2
Although the evidence for labor management best practices should be applied similarly to women of different races and ethnicities, disparities in obstetric care exist.3–6 Health disparities, or systematic differences in health status that “negatively impact less advantaged groups”7, exist in many aspects of maternal and neonatal morbidity and mortality.3–6 For example, non-Hispanic black women experience greater cesarean frequency than non-Hispanic white women, even after adjusting for risk factors.3,8–12 Health care disparities are differences in quality of care received by individuals in particular groups even when factors such as preferences and access are similar.13 Several racial and ethnic disparities in obstetrical care have been identified, including for labor induction, delayed pushing, and analgesia use.3,14,15 However, it is unknown whether care disparities explain the racial and ethnic disparity in cesarean delivery frequency.3,16
Achieving health equity requires understanding determinants of disparities.17,18 As reasons for the cesarean delivery disparity among minority women remain poorly characterized, we evaluated whether evidence regarding the safe promotion of vaginal delivery was applied differently based on maternal race and ethnicity. Our objectives were to examine whether racial and ethnic differences exist in the frequency of and indications for cesarean delivery, and determine whether these differences may be explained by disparities in the application of labor management strategies intended to reduce the frequency of cesarean deliveries.2
MATERIALS AND METHODS
This is a secondary analysis of the Assessment of Perinatal Excellence (APEX) study, a large, multicenter observational study performed at 25 hospitals in the United States. APEX methodology has been previously reported.19,20 The APEX cohort consisted of 115,502 women and their neonates who delivered at participating hospitals from 2008 to 2011. Study patients represented a diverse sample of sociodemographic characteristics and were sampled according to plans designed to avoid overrepresentation of patients from larger hospitals. Accoucheurs included general obstetricians, family medicine physicians, midwives, and maternal-fetal medicine physicians.19,20 Clinical data for patients who were at least 23 weeks of gestation and who had arrived at the hospital with a live fetus were collected by trained research personnel at each site. Maternal characteristics, details of the medical and obstetric history, intrapartum care, obstetric outcome, and indications for operative procedures were collected. Maternal race and ethnicity as reported in the chart were abstracted from the medical record. Institutional review board approval was obtained from all participating centers under a waiver of informed consent.
For this analysis, we analyzed two cohorts comprised of term, singleton, nonanomalous, vertex, liveborn neonates. The first cohort consisted of nulliparous women. The second cohort was composed of multiparous women with prior vaginal deliveries only. In both cohorts, women who were ineligible for vaginal deliveries (placenta previa, placenta accreta, prior cavity-entering myomectomy or classical cesarean delivery, active herpes infection) were excluded.
We first evaluated the frequencies of, and indications for, cesarean delivery. The indications assessed included non-reassuring fetal status, labor dystocia (inclusive of arrest of dilation, arrest of descent, or failed induction of labor), suspected macrosomia, elective, and “other” reasons (such as abruption or maternal medical complications).
In the nulliparous cohort, utilization of strategies intended to promote vaginal delivery was then evaluated using the College and SMFM Consensus2 as a framework to guide the labor management strategies analyzed. Although the Consensus2 was published after the APEX data accrual period, the evidence for these guidelines was accumulated before and during the APEX period. First, for failed induction of labor, we assessed the proportion of women who underwent cesarean performed for a diagnosis of failed induction prior to 5cm cervical dilation and prior to 12 hours of oxytocin and rupture of membranes (whichever came last).2,21–23 We additionally assessed the proportion of women with unfavorable cervical exam (modified Bishop score <5) who underwent labor induction without cervical ripening compared with those undergoing induction with cervical ripening.2,24,25 For arrest of dilation, we assessed the proportion of women whose cesarean for an indication of arrest of labor was performed prior to 4 hours of labor dystocia after reaching 5cm cervical dilation with ruptured membranes.2,26–28 Determination of labor dystocia with regard to uterine activity was made by individual providers; APEX does not have information about uterine activity. For both failed induction and arrest of dilation, 5cm was chosen for this analysis because of availability of this landmark in the APEX database and prior use of 5cm in the diagnosis of failed labor induction.22 For arrest of descent, we assessed the proportion of women undergoing cesarean delivery for an indication of arrest of descent prior to 3 hours in the second stage.2,29–31 Finally, given that there is no standard to know whether the diagnosis of non-reassuring status (NRFS) was differentially applied, we examined the proportion of women who underwent cesarean for an indication of NRFS with a 1-minute Apgar score of ≥7. Due to smaller sample sizes, we were not able to evaluate utilization of all these different labor management strategies in the multiparous cohort.
The outcomes that incorporate time were also evaluated as continuous variables: duration of induction prior to diagnosis of a failed induction (as defined above); duration of active labor prior to a diagnosis of arrest of dilation (length of time from 5cm until decision for cesarean delivery); and duration of second stage prior to diagnosis of arrest of descent (length of time from complete dilation until decision for cesarean delivery).
The Kruskal-Wallis test for continuous variables and the chi-square test for categorical variables were used to describe baseline demographic and clinical characteristics by maternal race and ethnicity. Multivariable logistic regression was used to evaluate binomial outcomes (cesarean delivery, adherence to failed induction guideline, adherence to cervical ripening guideline, adherence to arrest of dilation guideline, adherence to arrest of descent guideline, cesarean performed for NRFS with 1-minute Apgar ≥7, and the composite). Multivariable multinomial logistic regression was used to evaluate indication for cesarean delivery, an outcome with more than two categories. The continuous outcomes were evaluated using comparisons of least squares means based on multivariable linear regression. All models were adjusted for age, body mass index, gestational age at delivery, diabetes, smoking, chorioamnionitis, and gestational hypertension or preeclampsia. Due to equipoise regarding whether payer status was in the causal pathway between maternal race and cesarean delivery rates, a secondary analysis was performed with insurance status included in the regression. Hospital was also included in the models as a fixed effect. Imputation for missing data was not performed. SAS software (SAS Institute, Cary, NC) was used for the analyses. All tests were two-tailed and a p-value of <0.05 was used to define statistical significance.
RESULTS
A total of 75,400 women were eligible for inclusion, of whom 47% (N=35,529) were in the nulliparous cohort and 53% (N=39,871) in the multiparous cohort. The majority of hospitals were teaching hospitals (22/25, 88%). Of the 35,529 women eligible for analysis in the nulliparous cohort, 25.8% (9,156) underwent cesarean delivery (Table 1). There were a number of differences in demographic and clinical characteristics between racial and ethnic groups. Non-Hispanic black and Hispanic women were more likely to have government-assisted insurance and non-Hispanic white women were more likely to smoke. Asian women were more likely to have gestational diabetes mellitus, whereas non-Hispanic black women were more likely to have pre-gestational diabetes mellitus, chronic hypertension, or gestational hypertension or preeclampsia.
Table 1.
Demographic and clinical characteristics of nulliparous cohort
| Non-Hispanic white n=18,377 |
Non-Hispanic black n=7,583 |
Asian n=2,570 |
Hispanic n=6,999 |
P | |
|---|---|---|---|---|---|
| Age (years) | 28 (24–32) | 21 (19–25) | 29 (27–32) | 22 (19–26) | <.001 |
|
| |||||
| Insurance | <.001 | ||||
| Private | 14317 (78.6) | 2078 (27.5) | 2064 (80.8) | 1398 (20.2) | |
| Government assisted | 3520 (19.3) | 5127 (67.9) | 354 (13.9) | 3528 (50.9) | |
| Uninsured | 388 (2.1) | 342 (4.5) | 136 (5.3) | 2003 (28.9) | |
|
| |||||
| BMI at delivery (kg/m2) | 29.5 (26.6–33.4) | 30.9 (27.2–35.9) | 27.3 (25.1–30.0) | 29.9 (26.8–33.7) | <.001 |
|
| |||||
| Diabetes | <.001 | ||||
| None | 17470 (95.1) | 7228 (95.3) | 2310 (89.9) | 6601 (94.4) | |
| Pre-gestational | 142 (0.8) | 84 (1.1) | 22 (0.9) | 54 (0.8) | |
| Gestational | 762 (4.2) | 269 (3.6) | 237 (9.2) | 341 (4.9) | |
|
| |||||
| Smoking | 2039 (11.1) | 694 (9.2) | 34 (1.3) | 220 (3.2) | <.001 |
|
| |||||
| Chronic hypertension | 356 (1.9) | 233 (3.1) | 26 (1.0) | 77 (1.1) | <.001 |
|
| |||||
| Gestational hypertension or preeclampsia | 1997 (10.9) | 993 (13.1) | 101 (3.9) | 824 (11.8) | <.001 |
|
| |||||
| >2 prenatal care visits | 17545 (99.4) | 7108 (98.5) | 2426 (99.4) | 6642 (98.8) | <.001 |
|
| |||||
| Type of labor | <.001 | ||||
| None | 564 (3.1) | 122 (1.6) | 65 (2.5) | 95 (1.4) | |
| Spontaneous | 4433 (24.1) | 1840 (24.3) | 573 (22.3) | 1739 (24.9) | |
| Spontaneous, augmented | 6322 (34.4) | 2771 (36.5) | 1133 (44.1) | 2732 (39.0) | |
| Induced | 7058 (38.4) | 2850 (37.6) | 799 (31.1) | 2433 (34.8) | |
|
| |||||
| Chorioamnionitis (as reason for intrapartum antibiotics administration) | 1103 (6.0) | 442 (5.8) | 224 (8.7) | 851 (12.2) | <.001 |
|
| |||||
| Analgesia | <.001 | ||||
| None or IV or IM narcotic only | 1278 (7.0) | 696 (9.2) | 207 (8.1) | 1160 (16.6) | |
| Regional epidural anesthesia only | 9728 (53.0) | 3851 (50.8) | 1223 (47.6) | 2820 (40.3) | |
| Other anesthesia | 7357 (40.1) | 3030 (40.0) | 1140 (44.4) | 3011 (43.1) | |
|
| |||||
| Gestational age at delivery (weeks) | 39.9 (39.0–40.6) | 39.7 (38.9–40.4) | 39.6 (38.9–40.3) | 39.7 (38.9–40.6) | <.001 |
BMI, body mass index; IV, intravenous; IM, intramuscular
Number of missing values: insurance (274); BMI (486); diabetes (9); smoking (30); chronic hypertension (3); gestational hypertension or preeclampsia (2); number of prenatal visits (1505); chorioamnionitis (4); analgesia (28)
Data displayed as N (%) or median (IQR). The Kruskal-Wallis test was used for continuous variables and the chi-square test for categorical variables. P-value represents comparison across all racial and ethnic groups.
In the nulliparous cohort, the cesarean delivery frequencies were 25.0%, 28.3%, 28.7%, and 24.0% for non-Hispanic white, non-Hispanic black, Asian, and Hispanic women, respectively (Table 2). When adjusting for patient characteristics and hospital, the odds of cesarean delivery were higher in all racial and ethnic minority groups compared with non-Hispanic white women. This finding was related to both a greater odds of cesarean delivery for NRFS and labor dystocia (inclusive of arrest of dilation, arrest of descent and failed induction) in all racial and ethnic groups compared with non-Hispanic white women. Notably, non-Hispanic black women had a two-fold increase in odds of cesarean delivery for NRFS (aOR 2.39, 95% CI 2.13–2.69) compared with non-Hispanic white women. Conversely, non-Hispanic black (aOR 0.49, 95% CI 0.34–0.72) and Asian (aOR 0.48, 95% CI 0.28–0.81) women experienced lower odds of cesarean for suspected macrosomia, and Hispanic women experienced lower odds of elective cesarean delivery (aOR 0.60, 95% CI 0.42–0.87), compared with non-Hispanic white women.
Table 2.
Frequency and odds of cesarean delivery by maternal race and ethnicity in nulliparous cohort
| Non- Hispanic white |
Non-Hispanic black | Asian | Hispanic | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||
| N (%, 95%CI) |
N (%, 95%CI) |
Unadjusted OR (95%CI) |
aOR* (95% CI) |
N (%, 95%CI) |
Unadjusted OR (95%CI) |
aOR* (95% CI) |
N (%, 95%CI) |
Unadjusted OR (95%CI) |
aOR* (95% CI) |
|
| Cesarean delivery (vs. vaginal delivery)† | 4594 (25.0, 24.4–25.6) | 2148 (28.3, 27.3–29.3) | 1.19 (1.12–1.26) | 1.47 (1.36–1.59) | 737 (28.7, 26.9–30.4) | 1.21 (1.10–1.32) | 1.26 (1.14–1.40) | 1677 (24.0, 23.0–25.0) | 0.95 (0.89–1.01) | 1.17 (1.07–1.27) |
|
| ||||||||||
| Indication (vs. vaginal delivery)‡ | ||||||||||
| Non-reassuring fetal status | 1106 (6.0, 5.7–6.4) | 945 (12.5, 11.7–13.2) | 2.17 (1.98–2.38) | 2.39 (2.13–2.69) | 196 (7.6, 6.6–8.7) | 1.33 (1.14–1.56) | 1.29 (1.09–1.53) | 492 (7.0, 6.4–7.6) | 1.15 (1.03–1.29) | 1.22 (1.07–1.40) |
| Labor dystocia or failed induction | 2807 (15.3, 14.8–15.8) | 1043 (13.8, 13.0–14.5) | 0.94 (0.87–1.02) | 1.21 (1.09–1.33) | 460 (17.9, 16.4–19.4) | 1.23 (1.10–1.38) | 1.39 (1.23–1.57) | 1052 (15.1, 14.2–15.9) | 0.97 (0.90–1.05) | 1.24 (1.13–1.38) |
| Suspected macrosomia | 244 (1.3, 1.2–1.5) | 45 (0.6, 0.4–0.8) | 0.47 (0.34–0.64) | 0.49 (0.34–0.72) | 17 (0.7, 0.4–1.0) | 0.52 (0.32–0.86) | 0.48 (0.28–0.81) | 43 (0.6, 0.4–0.8) | 0.46 (0.33–0.63) | 0.76 (0.52–1.12) |
| Elective | 291 (1.6, 1.4–1.8) | 55 (0.7, 0.5–0.9) | 0.48 (0.36–0.64) | 0.84 (0.60–1.16) | 45 (1.8, 1.2–2.3) | 1.16 (0.85–1.60) | 1.02 (0.72–1.44) | 43 (0.6, 0.4–0.8) | 0.38 (0.28–0.53) | 0.60 (0.42–0.87) |
| Other | 138 (0.8, 0.6–0.9) | 54 (0.7, 0.5–0.9) | 0.99 (0.72–1.36) | 1.23 (0.85–1.79) | 19 (0.7, 0.4–1.1) | 1.04 (0.64–1.68) | 0.98 (0.58–1.65) | 38 (0.5, 0.4–0.7) | 0.71 (0.50–1.02) | 0.90 (0.57–1.42) |
Multivariable logistic regression (cesarean delivery) and multivariable multinomial logistic regression (indication). Compared with non-Hispanic white women. Adjusted for age, BMI, diabetes, smoking, chorioamnionitis, gestational hypertension or preeclampsia, gestational age at delivery, and hospital.
n=35,529; n=35,005 with complete covariable data.
n=35,529; n=34,982 with complete covariable data.
Among the 39,871 multiparous women with prior vaginal deliveries only, the overall frequency of cesarean delivery was 6.0%. The frequencies of cesarean delivery were 4.9%, 8.4%, 5.8%, and 6.0%, for non-Hispanic white, non-Hispanic black, Asian, and Hispanic women, respectively (Table 3). Non-Hispanic black women experienced greater odds of cesarean delivery overall (aOR 1.51, 95% CI 1.33–1.70), mainly attributed to increased odds of cesarean for NRFS and labor dystocia compared with non-Hispanic white women. Hispanic women had higher odds of having cesarean for NRFS and labor dystocia compared with non-Hispanic white women. In contrast, non-Hispanic black and Asian women experienced lower odds of cesarean for suspected macrosomia, and non-Hispanic black and Hispanic women experienced lower odds of elective cesarean.
Table 3.
Frequency and odds of cesarean delivery by maternal race and ethnicity in multiparous (prior vaginal deliveries only) cohort
| Non- Hispanic white |
Non-Hispanic black | Asian | Hispanic | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||
| N (%, 95%CI) |
N (%, 95%CI) |
Unadjusted OR (95%CI) |
aOR* (95% CI) |
N (%, 95%CI) |
Unadjusted OR (95%CI) |
aOR* (95% CI) |
N (%, 95%CI) |
Unadjusted OR (95%CI) |
aOR* (95% CI) |
|
| Cesarean delivery (vs. vaginal delivery)† | 873 (4.9, 4.6–5.2) | 712 (8.4, 7.9–9.0) | 1.78 (1.61–1.97) | 1.51 (1.33–1.70) | 105 (5.8, 4.8–6.9) | 1.20 (0.97–1.48) | 1.00 (0.80–1.25) | 718 (6.0, 5.6–6.5) | 1.24 (1.12–1.37) | 1.12 (0.98–1.27) |
|
| ||||||||||
| Indication (vs. vaginal delivery)‡ | ||||||||||
| Non-reassuring fetal status | 210 (1.2, 1.0–1.3) | 361 (4.3, 3.9–4.7) | 3.75 (3.16–4.46) | 3.12 (2.54–3.84) | 27 (1.5, 0.9–2.1) | 1.28 (0.85–1.92) | 1.24 (0.81–1.90) | 282 (2.4, 2.1–2.6) | 2.02 (1.69–2.42) | 1.48 (1.18–1.85) |
| Labor dystocia or failed induction | 233 (1.3, 1.2–1.5) | 210 (2.5, 2.2–2.8) | 1.97 (1.63–2.38) | 1.77 (1.42–2.21) | 30 (1.7, 1.1–2.3) | 1.28 (0.87–1.88) | 1.38 (0.91–2.08) | 232 (2.0, 1.7–2.2) | 1.50 (1.25–1.80) | 1.44 (1.14–1.82) |
| Suspected macrosomia | 100 (0.6, 0.5–0.7) | 34 (0.4, 0.3–0.5) | 0.74 (0.50–1.10) | 0.51 (0.32–0.80) | 6 (0.3, 0.1–0.6) | 0.60 (0.26–1.36) | 0.33 (0.14–0.78) | 72 (0.6, 0.5–0.7) | 1.09 (0.80–1.47) | 1.06 (0.72–1.56) |
| Elective | 167 (0.9, 0.8–1.1) | 31 (0.4, 0.2–0.5) | 0.41 (0.28–0.60) | 0.43 (0.28–0.66) | 21 (1.2, 0.7–1.7) | 1.25 0.79–1.98) | 0.76 (0.47–1.26) | 46 (0.4, 0.3–0.5) | 0.42 (0.30–0.58) | 0.51 (0.34–0.75) |
| Other | 163 (0.9, 0.8–1.1) | 74 (0.9, 0.7–1.1) | 0.99 (0.75–1.31) | 0.82 (0.60–1.14) | 20 (1.1, 0.6–1.6) | 1.22 (0.77–1.95) | 0.84 (0.51–1.40) | 86 (0.7, 0.6–0.9) | 0.80 (0.61–1.03) | 0.86 (0.62–1.20) |
Multivariable logistic regression (cesarean delivery) and multivariable multinomial logistic regression (indication). Compared with non-Hispanic white women. Adjusted for age, BMI, diabetes, smoking, chorioamnionitis, gestational hypertension or preeclampsia, gestational age at delivery, and hospital.
n=39,871; n=38,831 with complete covariable data.
n=39,871; n=38,828 with complete covariable data.
Nevertheless, women with racial and ethnic minority status did not experience higher frequencies of cesarean delivery secondary to differential application of labor management strategies or adherence to professional societies’ guidelines for preventing cesarean delivery (Table 4). Specifically, over three-quarters of women who had a cesarean performed for failed induction had that cesarean performed prior to 12 hours of oxytocin and ruptured membranes, approximately 20% of women undergoing labor induction with an unfavorable cervix did not receive cervical ripening, and over half of women who underwent cesarean delivery for arrest of descent were delivered prior to 3 hours in the second stage. Yet, the odds of these occurrences were similar across all racial and ethnic groups. Also, approximately one quarter of all women who underwent cesarean delivery for arrest of dilation were delivered prior to 4 hours of labor dystocia, although Hispanic women actually had decreased odds compared with non-Hispanic white women (aOR 0.67, 95% CI 0.49–0.92). Finally, in all racial and ethnic groups, over three-quarters of women experiencing cesarean delivery for NRFS had a neonate with a 1-minute Apgar score ≥7; however, Asian women were at increased odds of this scenario (aOR 1.79, 95% CI 1.07–3.00). The composite outcome (comprised of each of the above scenarios) did not differ by race and ethnicity.
Table 4.
Use of labor management strategies to reduce cesarean deliveries in nulliparous cohort by maternal race and ethnicity
| Non- Hispanic white |
Non-Hispanic Black | Asian | Hispanic | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| N (%) | N (%) | Unadjuste d OR (95%CI) |
aOR (95% CI)* |
N (%) | Unadjusted OR (95%CI) |
aOR (95% CI)* |
N (%) | Unadjusted OR (95%CI) |
aOR (95% CI)* |
|
| Cesarean performed for failed induction prior to 5cm dilation and prior to 12 hours of oxytocin or rupture of membranes (whichever came last)† | 357 (80.6) | 150 (75.8) | 0.75 (0.50–1.13) | 1.02 (0.61–1.72) | 43 (89.6) | 2.07 (0.80–5.39) | 1.06 (0.35–3.19) | 119 (78.8) | 0.90 (0.57–1.41) | 1.35 (0.69–2.63) |
| Labor induction with an unfavorable cervix and without cervical ripening‡ | 212 (21.2) | 147 (18.5) | 1.18 (0.93–1.49) | 0.99 (0.72–1.37) | 32 (19.5) | 1.11 (0.73–1.68) | 1.08 (0.67–1.76) | 86 (18.0) | 1.23 (0.93–1.62) | 0.98 (0.68–1.40) |
| Cesarean performed for arrest of dilation after reaching 5 cm dilatation and prior to 4 hours of dystocia § | 309 (26.0) | 135 (23.4) | 0.87 (0.69–1.10) | 0.79 (0.60–1.05) | 43 (23.1) | 0.86 (0.60–1.24) | 0.76 (0.51–1.13) | 105 (19.7) | 0.70 (0.55–0.90) | 0.67 (0.49–0.92) |
| Cesarean performed for arrest of descent after reaching complete dilation and prior to 3 hours of second stage ‖ | 464 (54.9) | 71 (60.2) | 1.24 (0.84–1.84) | 1.09 (0.69–1.72) | 94 (57.3) | 1.10 (0.79–1.55) | 1.25 (0.85–1.86) | 139 (55.2) | 1.01 (0.76–1.34) | 1.26 (0.87–1.83) |
| Cesarean for NRFS and Apgar score at 1 minute ≥7¶ | 890 (80.5) | 749 (79.3) | 0.93 (0.75–1.15) | 0.88 (0.67–1.16) | 175 (89.3) | 2.01 (1.25–3.24) | 1.79 (1.07–3.00) | 394 (80.1) | 0.97 (0.74–1.27) | 1.05 (0.76–1.45) |
| Composite, any# | 2179 (49.6) | 1219 (51.7) | 1.09 (0.98–1.20) | 1.08 (0.96–1.23) | 379 (52.2) | 1.11 (0.95–1.30) | 1.04 (0.88–1.24) | 822 (46.1) | 0.87 (0.78–0.97) | 0.91 (0.80–1.04) |
aOR, adjusted odds ratio; CI, confidence interval; NRFS, non-reassuring fetal status
Multivariable logistic regression. Compared with non-Hispanic white women. Adjusted for age, BMI, diabetes, smoking, chorioamnionitis, gestational hypertension or preeclampsia, gestational age at delivery, and hospital.
In patients who were induced, received oxytocin, did not reach 5 cm dilation, and underwent cesarean for an indication of failed induction (n=908; n=831 with complete covariable data).
In patients who were induced and had an unfavorable cervix (modified Bishop score <5) (n=2436; n=2403 with complete covariable data).
In patients with who reached 5 cm dilation but did not reach complete dilation, and underwent cesarean for an indication of arrest of dilation (n=2667; n=2466 with complete covariable data).
In patients with who reached complete dilation, and underwent cesarean for an indication of arrest of descent (n=1533; n=1359 with complete covariable data).
In patients who underwent cesarean for an indication of non-reassuring fetal status (n=2739; n=2699 with complete covariable data).
Of the strategies that the patient was eligible for, patient received one or more strategies considered non-use of labor management strategies to reduce cesarean deliveries (cesarean performed for failed induction prior to 5cm dilation and prior to 12 hours of oxytocin and rupture of membranes, labor induction without cervical ripening in women with an unfavorable cervix, cesarean performed for arrest of dilation prior to 4 hours of dystocia after reaching 5 cm dilatation, cesarean performed for arrest of descent prior to 3 hours of second stage, cesarean for NRFS with Apgar score at 1 minute ≥7).
The evaluation of continuous outcomes also could not explain the racial and ethnic differences in cesarean frequencies among the specific denominators. There were no differences in adjusted mean hours from oxytocin or rupture of membranes (whichever came last) to cesarean decision for failed induction by maternal race and ethnicity. Non-Hispanic black, Hispanic, and Asian women who reached 5cm cervical dilation remained pregnant for greater number of hours after 5 cm dilation and before a decision was made to proceed with cesarean decision for arrest of dilation compared with non-Hispanic white women (Table 5). In contrast, Asian women had a lower adjusted mean number of hours in the second stage prior to cesarean decision, although this finding does not fully explain the higher rate of cesarean delivery among Asian women, given the non-significant association between cesarean delivery for arrest of descent prior to 3 hours of the second stage and Asian race in categorical analyses.
Table 5.
Duration of time until cesarean decision by maternal race and ethnicity in nulliparous cohort
| Non- Hispanic white |
Non-Hispanic Black | Asian | Hispanic | ||||
|---|---|---|---|---|---|---|---|
| Adjusted mean (95%CI)* |
Adjusted mean (95%CI)* |
P | Adjusted mean (95%CI)* |
P | Adjusted mean (95%CI)* |
P | |
| Hours from oxytocin or rupture of membranes (whichever came last) to cesarean decision† | 9.50 (8.06–10.94) | 9.36 (7.74–10.98) | .82 | 10.15 (7.76–12.54) | .52 | 9.19 (7.40–10.98) | .67 |
| Hours from 5 cm to cesarean decision‡ | 7.9 (7.3–8.5) | 8.4 (7.7–9.1) | .02 | 8.6 (7.8–9.4) | .02 | 8.8 (8.1–9.5) | <.001 |
| Hours from complete dilation to cesarean decision§ | 2.96 (2.63–3.28) | 2.88 (2.47–3.29) | .63 | 2.68 (2.29–3.06) | .03 | 2.78 (2.41–3.16) | .16 |
CI, confidence interval
Comparisons of least squares means based on multivariable linear regression. Compared with non-Hispanic white women. Adjusted for age, BMI, diabetes, smoking, chorioamnionitis, gestational hypertension or preeclampsia, gestational age at delivery, and hospital.
In patients who were induced, received oxytocin, did not reach 5 cm dilation, and underwent cesarean for an indication of failed induction (n=908; n=838 with complete covariable data).
In patients who reached 5 cm dilation but did not reach complete dilation, and underwent cesarean for an indication of arrest of labor (n=2472; n=2454 with complete covariable data).
In patients with who reached complete dilation, and underwent cesarean for an indication of arrest of descent (n=1533; n=1366 with complete covariable data).
Finally, all models were performed with inclusion of insurance status in the regression. Inclusion of insurance status in the regression equations did not materially alter any results (data not shown). Specifically, inclusion of insurance status in the model did not alter the significance of any findings for nulliparous women, including indications for cesarean delivery, the application of labor management strategies for preventing cesarean delivery, or the duration of time until cesarean delivery. However, for multiparous women with prior vaginal deliveries, the inclusion of insurance status in the model resulted in no association between Hispanic ethnicity and cesarean for NRFS (aOR 1.21, 95% CI 0.94–1.55) or elective cesarean (aOR 0.73, 95% CI 0.48–1.12); no other differences were identified for women in other racial and ethnic groups.
DISCUSSION
Our data showed that non-Hispanic black, Asian, and Hispanic women all experience greater likelihood of cesarean delivery than non-Hispanic white women, even after accounting for socio-demographic and clinical differences, and that indications for cesarean delivery vary by maternal race and ethnicity. However, we did not identify differences in care provision by maternal race and ethnicity. Notably, among nulliparous women, the differential odds of cesarean could not be traced to differential application of guidelines for failed induction of labor, arrest of dilation, arrest of descent, NRFS, or cervical ripening.
Understanding sources of disparities is essential to promoting equity in health care delivery. We designed this investigation to determine whether women of racial and ethnic minority status were more likely to experience cesarean delivery due to differences in application of labor management strategies intended to promote safe vaginal delivery.1,2 Disparities in application of professional guidelines have been well documented in other areas of medicine.13 Our findings regarding disparities in cesarean indication have been seen in other cohorts. For example, Edmonds et al demonstrated non-Hispanic black women experienced greater odds of cesarean delivery for fetal distress or failure to progress9 and others have identified non-Hispanic black women were more likely to experience cesarean delivery for NRFS.10,12 Yet, the reasons for these disparities in overall rate and indication remain unknown.10,11,32 Although some have hypothesized minority women are at greater risk of cesarean delivery for “subjective” indications such as NRFS,10 our data do not support this hypothesis. This finding is consistent with data from a 1997–2002 California cohort in which, although specific labor and delivery practices could not be examined, racial differences in cesarean delivery rate were not fully explained by delivery hospital or physician.33
Identifying the roots of disparities in obstetric outcomes requires understanding differences in care provision. These findings suggest that labor management strategies to promote vaginal delivery are not differentially applied by maternal race and ethnicity. It is unclear that differences in frequency of NRFS or labor dystocia are due to true biological differences; however, it is plausible that other mediating factors may underlie these disparities. For example, the greater frequencies of hypertensive disorders and small-for-gestational-age neonates16,34 among minority women may contribute to the greater likelihood of cesarean delivery for NRFS. Others have hypothesized that differential outcomes of labor induction may be due to differential pharmacologic response to induction agents.35,36 Perhaps these findings warrant investigation into individualized responses to induction or augmentation agents, particularly in this era of precision medicine.37 Yet, any of these possibilities are speculative, and next steps must include multidisciplinary investigation of other patient- and provider-centered reasons for the identified disparities.
Strengths of this analysis include the large cohort with detailed information about patient characteristics and features of labor and delivery. In addition to having indications for cesarean delivery, we were able to take advantage of the granularity of APEX data to investigate specific labor and delivery management. Moreover, data were collected by trained research personnel at clinically and geographically diverse settings. However, several limitations warrant note. Evidence from the Consortium on Safe Labor has suggested using 6 cm as a cut-off for active labor,38 but the APEX study collected information on time of achieving 5 cm cervical dilation. Additionally, as is necessary for the nature of this question, these data are observational and cannot demonstrate causation. Additionally, we are unable to account for the role of health care provider type, as APEX collected the delivering attending as the assigned provider, which may not have been the same provider type as the individual managing the patient’s labor. Further, the majority of hospitals were teaching hospitals, which limits the generalizability of the findings. Finally, due to small numbers for some racial and ethnic groups, we were not able to evaluate specific types of care in the cohort of multiparous women with prior vaginal deliveries only.
Understanding whether and how maternal race and ethnicity influence the complex decision making processes that take place in obstetric care is critical to improving the quality of care for all women. Although women of minority racial and ethnic status experience greater likelihood of cesarean delivery as well as differences in indications for cesarean delivery, our data suggest these differences are not clearly attributed to unequal application of labor management strategies intended to promote safe vaginal deliveries. While these findings may offer some reassurance about equity in care provision, the racial and ethnic disparities in cesarean delivery and other maternal and neonatal health outcomes remain and the reasons for these differences continue to require elucidation.
Supplementary Material
Acknowledgments
Supported by grants from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) [HD21410, HD27869, HD27915, HD27917, HD34116, HD34208, HD36801, HD40500, HD40512, HD40544, HD40545, HD40560, HD40485, HD53097, HD53118] and the National Center for Research Resources [UL1 RR024989; 5UL1 RR025764].
The authors thank Cynthia Milluzzi, R.N. and Joan Moss, R.N.C., M.S.N. for protocol development and coordination between clinical research centers; Vinay Bhandaru, M.S. for statistical analysis; and William A. Grobman, M.D., M.B.A., Elizabeth Thom, Ph.D., Brian M. Mercer, M.D. and Catherine Y. Spong, M.D. for protocol development and oversight.
Footnotes
Presented as a poster at the 2017 Society of Maternal-Fetal Medicine 37th Annual Meeting in Las Vegas, NV (January 23–28, 2017).
Financial Disclosure
Each author has indicated that he or she has met the journal’s requirements for authorship.
The authors did not report any potential conflicts of interest.
References
- 1.Spong C, Berghella V, Wenstrom K, Mercer B, Saade G. Preventing the first cesarean delivery: Summary of a joint Eunice Kennedy Shriver National Institute of Child Health and Human Development, Society for Maternal-Fetal Medicine, and American College of Obstetricians and Gynecologists Workshop. Obstet Gynecol. 2012;120(5):1181–1193. doi: 10.1097/aog.0b013e3182704880. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Safe prevention of the primary cesarean delivery. Obstetric Care Consensus No. 1. American College of Obstetricians and Gynecologists. Obstet Gynecol. 2014;123:693–711. doi: 10.1097/01.AOG.0000444441.04111.1d. [DOI] [PubMed] [Google Scholar]
- 3.Grobman W, Bailit J, Rice M, et al. Racial and ethnic disparities in maternal morbidity and obstetric care. Obstet Gynecol. 2015;125(6):1460–1467. doi: 10.1097/AOG.0000000000000735. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Creanga A, Bateman B, Kuklina E, Callaghan W. Racial and ethnic disparities in severe maternal morbidity: a multistate analysis, 2008–10. Am J Obstet Gynecol. 2014;210(5):435.e431–438. doi: 10.1016/j.ajog.2013.11.039. [DOI] [PubMed] [Google Scholar]
- 5.Creanga A, Berg C, Syverson C, Seed K, Bruce F, Callaghan W. Pregnancy-related mortality in the United States, 2006–2010. Obstetrics and Gynecology. 2015;125(1):5–12. doi: 10.1097/AOG.0000000000000564. [DOI] [PubMed] [Google Scholar]
- 6.March of Dimes. [Accessed November 2, 2016];March of Dimes 2016 Premature Birth Report Card. 2016 www.marchofdimes.org/materials/premature-birth-report-card.pdf.
- 7.Dehlendorf C, Bryant A, Huddleston H, Jacoby V, Fujimoto V. Health disparities: definitions and measurements. Am J Obstet Gynecol. 2010;202(3):212–213. doi: 10.1016/j.ajog.2009.12.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Min C, Ehrenthal D, Strobino D. Investigating racial differences in risk factors for primary cesarean delivery. Am J Obstet Gynecol. 2015;212(6):814.e811–814. doi: 10.1016/j.ajog.2015.01.029. [DOI] [PubMed] [Google Scholar]
- 9.Edmonds J, Yehezkel R, Liao X, Moore Simas T. Racial and ethnic differences in primary, unscheduled cesarean deliveries among low-risk primiparous women at an academic medical center: a retrospective cohort study. BMC Pregnancy and Childbirth. 2013;13:1–7. doi: 10.1186/1471-2393-13-168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Washington S, Caughey A, Cheng Y, Bryant A. Racial and ethnic differences in indication for primary cesarean delivery at term: experience at one U.S. institution. Birth. 2012;39(2):128–134. doi: 10.1111/j.1523-536X.2012.00530.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bryant A, Washington S, Kuppermann M, Cheng Y, Caughey A. Quality and equality in obstetric care: Racial and ethnic differences in caesarean section delivery rate. Paediatric and Perinatal Epidemiology. 2009;23(5):454–462. doi: 10.1111/j.1365-3016.2009.01059.x. [DOI] [PubMed] [Google Scholar]
- 12.Getahun D, Strickland D, Lawrence J, Fassett M, Koebnick C, Jacobsen S. Racial and ethnic disparities in the trends in primary cesarean delivery based on indications. Am J Obstet Gynecol. 2009;201(4):422.e421–427. doi: 10.1016/j.ajog.2009.07.062. [DOI] [PubMed] [Google Scholar]
- 13.Institute of Medicine. Unequal treatment: Confronting racial and ethnic disparities in health care. Washington, DC: The National Academies Press; 2003. [PubMed] [Google Scholar]
- 14.Butwick A, Blumenfeld Y, Brookfield K, Nelson L, Weiniger C. Racial and ethnic disparities in mode of anesthesia for cesarean delivery. Anesthesia and Analgesia. 2016;122(2):472–479. doi: 10.1213/ANE.0000000000000679. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Toledo P, Sun J, Grobman W, Wong C, Feinglass J, Hasnain-Wynia R. Racial and ethnic disparities in neuraxial labor analgesia. Anesthesia and Analgesia. 2012;114(1):172–178. doi: 10.1213/ANE.0b013e318239dc7c. [DOI] [PubMed] [Google Scholar]
- 16.Hamilton BE, Martin JA, Osterman MJ, Curtin SC, Mathews T. Births: Final data for 2014. National Vital Statistics Reports. 2015;64(12):1–64. [PubMed] [Google Scholar]
- 17.Eichelberger KY, Doll K, Ekpo G, Zerden M. Black Lives Matter: Claiming a space for evidence-based outrage in obstetrics and gynecology. American Journal of Public Health. 2016;106(10):1771–1772. doi: 10.2105/AJPH.2016.303313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Institute of Medicine. How far have we come in reducing health disparities? Progress since 2000: Workshop summary. Washington, DC: The National Academies Press; 2012. [PubMed] [Google Scholar]
- 19.Grobman W, Bailit J, Rice M, et al. Can differences in obstetric outcomes be explained by differences in the care provided? The MFMU Network APEX study. Am J Obstet Gynecol. 2014;211(2):e1–147.e116. doi: 10.1016/j.ajog.2014.03.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Bailit J, Grobman W, Rice M, et al. Risk-adjusted models for adverse obstetric outcomes and variation in risk-adjusted outcomes across hospitals. Am J Obstet Gynecol. 2013;209:446.e441–430. doi: 10.1016/j.ajog.2013.07.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Simon C, Grobman W. When has an induction failed? Obstet Gynecol. 2005;105(4):705–709. doi: 10.1097/01.AOG.0000157437.10998.e7. [DOI] [PubMed] [Google Scholar]
- 22.Rouse D, Weiner S, Bloom S, et al. Failed labor induction: Toward an objective diagnosis. Obstet Gynecol. 2011;117(2 Pt 1):267–272. doi: 10.1097/AOG.0b013e318207887a. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Rouse D, Owen J, Hauth J. Criteria for failed labor induction: Prospective evaluation of a standardized protocol. Obstet Gynecol. 2000;96(5 Pt 1):671–677. doi: 10.1016/s0029-7844(00)01010-3. [DOI] [PubMed] [Google Scholar]
- 24.Jozwiak M, Bloemenkamp K, Kelly A, Mol B, Irion O, Boulvain M. Mechanical methods for induction of labor. Cochrane Database Syst Rev. 2012;3 doi: 10.1002/14651858.CD001233.pub2. (CD001233) [DOI] [PubMed] [Google Scholar]
- 25.Hofmeyr G, Gulmezoglu A, Pileggi C. Vaginal misoprostol for cervical ripening and induction of labor. Cochrane Database Syst Rev. 2010;10 doi: 10.1002/14651858.CD000941.pub2. (CD000941) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Rouse D, Owen J, Hauth J. Active-phase labor arrest: oxytocin augmentation for at least 4 hours. Obstet Gynecol. 1999;93(3):323–328. doi: 10.1016/s0029-7844(98)00448-7. [DOI] [PubMed] [Google Scholar]
- 27.Rouse D, Owen J, Savage K, Hauth J. Active phase labor arrest: Revisiting the 2-hour minimum. Obstet Gynecol. 2001;98(4):550–554. doi: 10.1016/s0029-7844(01)01516-2. [DOI] [PubMed] [Google Scholar]
- 28.Henry D, Cheng Y, Shaffer B, Kaimal A, Bianco K, Caughey A. Perinatal outcomes in the setting of active phase arrest of labor. Obstet Gynecol. 2008;112(5):1109–1115. doi: 10.1097/AOG.0b013e31818b46a2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Rouse D, Weiner S, Bloom S, et al. Second-stage labor duration in nulliparous women: Relationship to maternal and perinatal outcomes. American Journal of Obstetrics and Gynecology. 2009;201(4):357.e351–357. doi: 10.1016/j.ajog.2009.08.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.LeRay C, Audibert F, Goffinet F, Fraser W. When to stop pushing: Effects of duration of second-stage expulsive efforts on maternal and neonatal outcomes in nulliparous women with epidural analgesia. American Journal of Obstetrics and Gynecology. 2009;201(4):361.e361–367. doi: 10.1016/j.ajog.2009.08.002. [DOI] [PubMed] [Google Scholar]
- 31.Cheng Y, Hopkins L, Caughey A. How long is too long: Does a prolonged second stage of labor in nulliparous women affect maternal and neonatal outcomes? American Journal of Obstetrics and Gynecology. 2004;191(3):933–938. doi: 10.1016/j.ajog.2004.05.044. [DOI] [PubMed] [Google Scholar]
- 32.Kabir A, Pridjian G, Steinmann W, Herrera E, Khan M. Racial differences in cesareans: an analysis of U.S. 2001 National Inpatient Sample Data. Obstet Gynecol. 2005;105(4):710–718. doi: 10.1097/01.AOG.0000154154.02581.ce. [DOI] [PubMed] [Google Scholar]
- 33.Chung J, Garite T, Kirk A, Hollard A, Wing D, Lagrew D. Intrinsic racial differences in the risk of cesarean delivery are not explained by differences in caregivers or hospital site of delivery. Am J Obstet Gynecol. 2006;194(5):1323–1328. doi: 10.1016/j.ajog.2005.11.043. [DOI] [PubMed] [Google Scholar]
- 34.Martin J, Hamilton BE, Ventura S, et al. Births: Final data for 2009. National Vital Statistics Reports. 2011;60(1):1–70. [PubMed] [Google Scholar]
- 35.Stephenson M, Hawkins J, Pevzner L, Powers B, Wing D. Racial/ethnic differences in labor outcomes with prostaglandin vaginal inserts. Journal of Racial and Ethnic Health Disparities. 2015;2(2):149–157. doi: 10.1007/s40615-014-0058-7. [DOI] [PubMed] [Google Scholar]
- 36.Wing D, Sheibani L. Pharmacotherapy options for labor induction. Expert Opin Pharmacother. 2015;16(11):1657–1668. doi: 10.1517/14656566.2015.1060960. [DOI] [PubMed] [Google Scholar]
- 37.Precision Medicine Initiative Working Group Report to the Advisory Committee to the Director of the National Institutes of Health. The Precision Medicine Initiative Cohort Program - Building a Research Foundation for 21st Century Medicine. Washington, DC: National Institutes of Health; 2015. [Google Scholar]
- 38.Zhang J, Landy H, Branch D, et al. Contemporary patterns of spontaneous labor with normal neonatal outcomes. Obstet Gynecol. 2010;116(6):1281–1287. doi: 10.1097/AOG.0b013e3181fdef6e. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.