Between 1970 and 2013, the cesarean delivery rate in the United States increased from 5.5% to 32.7%.1,2 The optimal rate remains controversial, but consensus exists that the current rate is too high.3–5 Unfortunately, reducing cesarean deliveries has proven challenging. Recent recommendations have targeted modifiable obstetric practices thought to influence the cesarean rate.4,5
Admission at less than 3–4 cm cervical dilation in term spontaneous labor is one such practice.4–7 Observational studies have reported admission at less than 4 cm cervical dilation to be associated with increased epidural use,8,9 oxytocin augmentation,8–10 and cesarean delivery7–9,11,12 compared with later admission (4 cm or greater cervical dilation). The reasons for this are unclear.
In August 2012, the legislatively mandated Washington State initiative, the Bree Collaborative, published “The Bree Collaborative Obstetrics Care Report and Recommendations.”13 One of the recommendations was: “Admit only those spontaneously laboring women at term who present with no fetal or maternal compromise when the cervix is 4 centimeters or more dilated.” The presence or absence of maternal or fetal compromise and counseling and reassessment of those patients not admitted were left to clinician discretion. Adherence to this Bree recommendation was voluntary, and its effect on obstetric care is unclear.
The aim of this study was to examine the relationship between cervical dilation on admission and both maternal and newborn outcomes. The release of the Bree Report allowed changes in admission practices and any concurrent changes in maternal and newborn outcomes, potentially attributable to the Bree Report, to be described.
MATERIALS AND METHODS
This retrospective cohort study used data from the Foundation for Health Care Quality's Obstetrics Clinical Outcomes Assessment Program, an ongoing multicenter, clinician-led, voluntary, collaborative quality initiative. Fourteen Washington State hospitals participated during all or part of the study period. The hospitals included urban and rural settings; with and without onsite perinatology, midwifery, family practice, and obstetrics services; supported by levels I, II, and III neonatal care; and performed 100–7,000 deliveries per year.
The Obstetrics Clinical Outcomes Assessment Program abstracts all consecutive deliveries at each institution, capturing more than 100 data points for each delivery related to prepregnancy health, pregnancy complications, labor course, delivery, and postnatal outcomes for mothers and newborns.
Individually trained abstractors use the medical record as the source of variables defined by the Obstetrics Clinical Outcomes Assessment Program and, where applicable, by American College of Obstetricians and Gynecologists obstetric data definitions. The information is entered into a cloud-based, standardized data tool. Abstractors include obstetric providers, nurses, and health care data and quality improvement specialists. Data undergo real-time quality checks performed both at the site and aggregate level. Monthly web meetings and unlimited access to Obstetrics Clinical Outcomes Assessment Program staff for education and support are available. Audit of Obstetrics Clinical Outcomes Assessment Program data against billing records with a minimum of 90% agreement is required. Audits for the time period represented demonstrated greater than 95% agreement between cases submitted for analysis and billing records. Institutions are identified with a unique Obstetrics Clinical Outcomes Assessment Program number. Personal health information is encrypted in the database. This allows sites access to the identifiers for their own institution for quality improvement purposes, but only a unique system-generated patient identification code is available in the aggregate database.
The analysis included singleton, term (37–43 6/7 weeks of gestation) deliveries of spontaneously laboring women between 2012 and 2014. History of cesarean delivery, cesarean delivery without labor, and ruptured membranes on presentation resulted in exclusion. Breech deliveries were included because the Bree recommendation did not specify presentation. All cesarean deliveries reported are primary cesarean deliveries (the first cesarean delivery in a nulliparous or multiparous woman). Women with more than one delivery during the study period were included.
The Western institutional review board has determined that the Obstetrics Clinical Outcomes Assessment Program is not engaged in human subjects research and therefore is exempt from institutional review board review.
Early compared with late admission was based on recorded cervical dilation on admission: early admission was defined as less than 4 cm cervical dilation and late admission was defined as 4 cm or greater cervical dilation. The 4-cm threshold was based on the Bree Report recommendation.13
Maternal outcomes included epidural use for pain relief in labor, augmentation of labor with oxytocin, hospital admission-to-delivery time of less than 12 hours, primary cesarean delivery, first-stage cesarean delivery, second-stage cesarean delivery, operative vaginal delivery, third- or fourth-degree laceration, maternal blood transfusion, maternal intensive care unit admission, and maternal readmission to the delivering hospital within 30 days of delivery. Newborn outcomes included resuscitation efforts at delivery, 5-minute Apgar score less than 7, neonatal intensive care unit admission, and neonatal readmission to the delivering hospital within 30 days of birth. Exclusive breastfeeding was evaluated and defined, according to the Perinatal Core Measure by the Joint Commission,14 as nutrition with only breast milk (including donor milk) throughout the hospital stay. Women who were ineligible to breastfeed were excluded from analyses in which breastfeeding was the outcome.
Maternal demographics, prepregnancy health, and pregnancy characteristics were measured. Maternal demographics included age and race and ethnicity. Prepregnancy health characteristics included prepregnancy body mass index (BMI, calculated as weight (kg)/[height (m)]2) calculated from height and recorded prepregnancy weight, prepregnancy diagnosis of diabetes, and prepregnancy diagnosis of hypertension. Pregnancy characteristics included year of delivery, diagnosis of gestational diabetes, diagnosis of preeclampsia, and parity (nulliparous or multiparous).
Maternal demographics, prepregnancy, and pregnancy characteristics are presented by cervical dilation on admission for the entire analytic sample, stratifying by parity.
There were two parts to the analysis. In the first part, the relationship between cervical dilation on admission and maternal and newborn outcomes was examined for the total study population (2012–2014). The second part of the analysis examined practice and outcomes for each of the 3 years (2012–2014) to determine whether admission practices and outcomes had changed over the study period.
The association of timing of admission (early compared with late) with maternal and newborn outcomes was examined separately by parity. General linear models with a log-link were used to directly calculate the relative risk (RR) for the associations of cervical dilation on admission with maternal and newborn outcomes.15 Adjustment was made for lack of independence of outcomes within sites. Potential confounders were variables considered a priori to be associated with the decision to admit and the risk of outcomes and were not thought to be consequences of timing of admission.16 Adjustment variables included maternal age, prepregnancy BMI, and year of delivery. All statistical tests were two-sided with an α level of 0.05.
To examine whether admission practices changed over the 3-year study period, the percentage admitted early was compared by year and parity using a χ2 test. General linear models with a log-link, adjusted to account for lack of independence of outcomes within sites, were used to examine whether the prevalence of maternal and newborn outcomes changed over the course of the study period.15 Nulliparous and multiparous women were combined for this analysis. To account for changes in the study population over time, which may have influenced outcome prevalence, models were adjusted for race and ethnicity, age, parity, and prepregnancy BMI. The adjusted prevalence of maternal and newborn outcomes for each year was estimated from these models. Statistical significance was assessed through the Wald test.17
Although data were complete for all maternal and newborn outcomes, a total of 7,193 deliveries were excluded as a result of missing data on membrane status on admission and 14% of deliveries (n=1,558) were missing data on maternal race and ethnicity. Consequently we conducted a sensitivity analysis using multiple imputation to examine whether exclusion of deliveries with missing data for inclusion or exclusion criteria (eg, membrane status on admission) or missing covariate information (eg, race and ethnicity) biased our findings. Assuming that data were missing at random, such an analysis increases efficiency and reduces bias.18 To impute the missing values, we specified individual equations based on the selection rules recommended by van Buuren et al.19 Among the variables considered for each imputation equation were all demographic characteristics, prepregnancy health characteristics, pregnancy characteristics, and the maternal and newborn outcomes. After imputing 10 complete data sets (m=10), the regression coefficients from the analysis in each data set were combined using the rules described by Little and Rubin.18
The Obstetrics Clinical Outcomes Assessment Program database included 52,700 deliveries from January 1, 2012, through December 31, 2014. Exclusions are shown in Figure 1, resulting in a final analytic sample of 11,368 women. Of these 29% (n=3,284) were admitted early (less than 4 cm cervical dilation), and 71% (n=8,084) were admitted late (4 cm or greater cervical dilation).
Overall, the majority of women was 20–34 years old (n=8,877 [78.1%]; Table 1). Approximately half of deliveries was among racial or ethnic minorities with Hispanic women comprising the largest group (n=2,986 [30.4%]). Fifty seven percent (n=6,205) had a prepregnancy BMI less than 25. Prepregnancy diabetes and prepregnancy hypertension were each present in less than 1% of the study population (0.3% and 0.8%, respectively). Gestational diabetes mellitus was recorded for 5.8% (n=652) and preeclampsia for 2.9% (n=333). The majority of deliveries was among multiparous women (n=6,399 [56.3%]). Compared with those admitted early, women admitted late tended to be slightly older, were more likely to have a prepregnancy BMI less than 25, and more likely to be multiparous.
Among nulliparous women, after adjusting for maternal age, prepregnancy BMI, and year of delivery, early admission was associated with increased likelihood of epidural use and augmentation with oxytocin and with decreased likelihood of delivering within 12 hours of admission (Table 2). After adjustment, early admission was also associated with increased likelihood of primary cesarean delivery and increased risk of first-stage but not second-stage cesarean delivery. A decreased risk of third- or fourth-degree lacerations was also observed. Because lacerations only occur to women who reach the second stage of labor, we conducted a sensitivity analysis restricted to women who reached the second stage. The direction of the association remained consistent but the sample size was smaller and the results were no longer statistically significant (RR 0.73, confidence interval 0.52–1.03). Postdelivery complications were rare in this group (less than 1%) and did not appear to differ by early compared with late admission. Among newborns of nulliparous women, after adjustment, early admission was associated with increased likelihood of neonatal intensive care unit admission and a slight decrease in the likelihood of exclusive breastfeeding.
Among multiparous women, after adjusting for maternal age, prepregnancy BMI, and year of delivery, early admission was associated with increased likelihood of epidural use and augmentation with oxytocin and with decreased likelihood of delivering within 12 hours of admission (Table 2). After adjustment, early admission was associated with increased likelihood of primary cesarean delivery, increased risk of first-stage cesarean delivery, and an increased likelihood of operative vaginal delivery. Postdelivery complications were rare among multiparous women and did not appear to differ by early compared with late admission except for maternal readmission to the delivering hospital within 30 days of delivery, which appeared to be more likely among those admitted early. Among newborns of multiparous women, after adjustment, early admission was associated with decreased risk of readmission to the delivering hospital within 30 days of delivery.
Examining changes in admission practices across the 3 study years, the proportion of women admitted late increased over time (Fig. 2). Among nulliparous women, late admission increased from 53.6% (n=565) in 2012 to 68.2% in 2014 (n=1,631) (P<.001), a 14.6% increase. Similarly for multiparous women, late admission increased from 72.3% (n=1,075) in 2012 to 82.4% in 2014 (n=2,422) (P<.001), a 10.1% increase.
After adjusting for maternal nonwhite race and ethnicity, age, parity, and prepregnancy BMI, primary cesarean delivery rates decreased from 10.5% to 7.9% between 2012 and 2014 (P<.001), and prevalence of third- or fourth-degree lacerations decreased from 3.2% to 2.7% (P<.001) (Table 3). The prevalence of evaluated adverse newborn outcomes may also have decreased between 2012 and 2014, but these changes were small (mostly less than 1%) and did not reach statistical significance. Exclusive nutrition with breast milk during the hospital stay increased from 77.6% in 2012 to 86.1% in 2014 (P<.001).
Results did not change appreciably in the analysis using multiple imputation (data not shown, available on request).
Our study shows that for nulliparous and multiparous women in term spontaneous labor with intact membranes, early hospital admission (less than 4 cm) is associated with increased likelihood of medical intervention (oxytocin augmentation and epidural use) and primary cesarean delivery, in agreement with earlier studies.6–9,12 Our study adds to the existing literature by including associations with additional maternal and newborn outcomes, crucial factors in consideration of any change in practice. Notably for newborns of nulliparous women, early admission was associated with an increased risk of neonatal intensive care unit admission and a decreased likelihood of exclusive breastfeeding.
The opportunity to evaluate changes in admission practices was afforded by the legislatively mandated Bree Report, released August 2012.13 Government involvement in medical decision-making is controversial and in May 2013, the Executive Board of the American College of Obstetricians and Gynecologists issued a Statement of Policy opposing government interference in physician decisions about patient care.20 Although the Bree Report was mandated by the Washington State Legislature, its recommendations were developed by obstetricians and adoption of recommendations relating to the management of spontaneous labor was voluntary. During the study period, the proportion of patients admitted early in labor decreased, suggesting the Bree recommendation had influenced obstetric care. Our study detected no obvious adverse effects in the outcomes evaluated, and some improvements were observed, including a decrease in the primary cesarean delivery rate.
Whether decreasing early admission truly contributed to the observed decrease in cesarean delivery remains unproven. Early admission may reflect an underlying risk for cesarean delivery independent of care. If this is the case, the rate of cesarean delivery should be independent of cervical dilation on admission. In our study, if women with an inherently increased risk of cesarean delivery had moved from the early (less than 4 cm) to the late (4 cm or greater) admission group, the prevalence of cesarean delivery might be expected to increase among those admitted late. The rate of cesarean delivery across the 3-year period did not differ significantly among those admitted late (Fig. 3). Thus, pre-existing risk factors for operative delivery may not fully account for the increased cesarean delivery rate observed in the early admission group.
Clearly not all term women presenting in spontaneous labor are appropriate for late admission. Only approximately 25% of women in the Obstetrics Clinical Outcomes Assessment Program database were even potential candidates for flexibility in timing of admission (term, singleton, intact membranes, and no prior cesarean delivery), thus limiting the effect of any changes in admission practices. Appropriately, the Bree Report stipulated late admission should only be considered in women who present without maternal or fetal compromise. However, the determination of this was left to clinical judgment, and we were unable to evaluate whether there was any real or perceived evidence of maternal or fetal compromise underlying early admission in our study population.
Strengths of our study include the large sample size (N=11,368 deliveries), chart-abstracted data, inclusion of all deliveries without sampling, a diverse group of participating hospitals, and measurement of maternal and newborn outcomes including mode of delivery. However, our findings must be considered in light of several limitations. This study was observational and although adjustments were made for differences in population characteristics both by timing of admission and over time, these may not fully account for the influence of other changes in obstetric practice on cesarean delivery and maternal and newborn outcomes. Small numbers of adverse outcomes also limited the power to detect differences between admission groups and over time. However, post hoc power calculations suggest that we had greater than 80% power to detect differences of 2% or greater. Additionally, maternal and newborn outcomes examined did not include all potential outcomes of interest (eg, infectious morbidity, neonatal ischemic–hypoxic encephalopathy) and readmissions were only captured when they were to the hospital where the delivery occurred. We were also unable to identify women evaluated and sent home before being admitted for delivery to report on any differences in outcomes in this group. Finally, although the study population represents a diverse group of hospitals in Washington State, whether findings are generalizable to other populations is not known.
Whether later admission can safely decrease cesarean delivery rates remains undetermined. How patients might feel about later admission is also unknown. Interestingly, McNiven et al9 reported more positive patient feedback from women randomized to later admission in labor. In the absence of an adequately powered prospective randomized study, which may be logistically challenging, our data may shed some light on this important issue and we hope they will stimulate a study to answer this question definitively.
1. Placek PJ, Taffel SM. Trends in cesarean section rates for the United States, 1970–78. Public Health Rep 1980;95:540–8.
2. Osterman MJ, Martin JA. Trends in low-risk cesarean delivery in the United States, 1990–2013. Natl Vital Stat Rep 2014;63:1–16.
3. WHO statement on cesarean section rates. Geneva (Switzerland): World Health Organization; 2015.
4. Spong CY, Berghella V, Wenstrom KD, Mercer BM, Saade GR. 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;120:1181–93.
5. American College of Obstetricians and Gynecologists; Society for Maternal-Fetal Medicine. Obstetric care consensus no. 1: safe prevention of the primary cesarean delivery. Obstet Gynecol 2014;123:693–711.
6. Main EK, Bloomfield L, Hunt G; Sutter Health, First Pregnancy and Delivery Clinical Initiative Committee. Development of a large-scale obstetric quality-improvement program that focused on the nulliparous patient at term. Am J Obstet Gynecol 2004;190:1747–56.
7. Main EK, Moore D, Farrell B, Schimmel LD, Altman RJ, Abrahams C, et al.. Is there a useful cesarean birth measure? Assessment of the nulliparous term singleton vertex cesarean birth rate as a tool for obstetric quality improvement. Am J Obstet Gynecol 2006;194:1644–51.
8. Holmes P, Oppenheimera LW, Wen SW. The relationship between cervical dilatation at initial presentation in labour and subsequent intervention. BJOG 2001;108:1120–4.
9. McNiven PS, Williams JI, Hodnett E, Kaufman K, Hannah ME. An early labor assessment program: a randomized, controlled trial. Birth 1998;25:5–10.
10. Malone FD, Geary M, Chelmow D, Stronge J, Boylan P, D'Alton ME. Prolonged labor in nulliparas: lessons from the active management of labor. Obstet Gynecol 1996;88:211–5.
11. Rahnama P, Ziaei S, Faghihzadeh S. Impact of early admission in labor on method of delivery. Int J Obstet Gynecol 2006;92:217–20.
12. Jackson DJ, Lang JM, Ecker J, Swartz WH, Heeren T. Impact of collaborative management and early cervical admission in labor on method of delivery. J Obstet Gynecol Neonatal Nurs 2003;32:147–57.
13. Obstetrics Care Workgroup. Obstetrics care topic report and recommendations. In: Quinn R, editor. Seattle (WA): Dr. Robert Bree Collaborative; 2012.
15. Lumley TR, Kronmal R, Ma S. Relative risk regression in medical research: models, contrasts, estimators, and algorithms. UW Biostatistics Working Paper Series. 2006. Available at: http://biostats.bepress.com/uwbiostat/paper293/
. Retrieved January 11, 2016.
16. Greenland S, Pearl J, Robbins JM. Causal diagrams for epidemiologic research. Epidemiology 1999;10:37–48.
17. Hosmer DW, Lemeshow S. Applied logistic regression. New York (NY): John Wiley & Sons, Inc; 2000.
18. Little R, Rubin D. Statistical analysis with missing data. New York (NY): John Wiley & Sons, Inc; 2002.
19. van Buuren S, Boshuizen HC, Knook DL. Multiple imputation of missing blood pressure covariates in survival analysis. Stat Med 1999;18:681–94.
20. American College of Obstetricians and Gynecologists; American Congress of Obstetricians and Gynecologists. Legislative interference with patient care, medical decisions, and the patient–physician relationship. Statement of Policy. Washington, DC: American College of Obstetricians and Gynecologists; American Congress of Obstetricians and Gynecologists; 2013.