The contribution of cesarean delivery to maternal morbidity and mortality in the United States placed decreasing the primary cesarean delivery rate at the top of the national public health agenda.1–3 Rates of cesarean delivery vary 10-fold with no discernible association between a hospital's cesarean delivery rate and rates of adverse obstetric outcomes.4–9 Consensus guidelines outline an evidence-based approach to prevent primary cesarean delivery, yet integration of these recommendations into practice is challenging.4,10
Prior interventions targeted low-risk women and emphasized the nulliparous term singleton vertex cesarean delivery rate.11–14 The diverse indications for cesarean delivery in even a low-risk subset of women may explain the modest effect of these quality improvement initiatives. Focusing on a specific indication for cesarean delivery such as twin pregnancy may be a more feasible alternative. Rates of cesarean delivery for multiples rose from 52.9% in 1995 to 73.2% in 2008, with many twin cesarean deliveries occurring without medical indication.10,15–18 Although evidence-based guidelines discourage elective cesarean deliveries for twins, the influence of patient and health care provider preferences in this high-risk group may yield changes in public policy meaningless in the context of clinical practice.17–19
The aim of the present study is to test the ability of a targeted intervention to decrease rates of primary cesarean delivery for twins. We used available evidence to implement an educational and clinical support program to promote twin vaginal delivery at our center.4,20–22 We hypothesized this multifaceted quality improvement intervention in a contemporary national climate supporting vaginal delivery would improve institutional rates of twin vaginal birth.
MATERIALS AND METHODS
This retrospective cohort study assessed mode of delivery for twins at a single urban academic tertiary care center during the 6 years surrounding the implementation of the twin vaginal delivery promotion in 2013. The study considered deliveries between January 2010 and December 2012 compared with those between January 2013 and December 2015. The program was informed by the results of a 25-item needs assessment to faculty practicing obstetrics at our institution to survey current knowledge and approach to mode of delivery for twins.18 From our pilot work we created a cohesive intervention launched as a single entity in 2013 (Box 1).
Components of Twin Vaginal Birth Intervention Cited Here...
• Health care provider needs assessment
• Expert lecture on approach to twin vaginal birth
• Health care provider simulation on twin vaginal birth23
• Launch of multiple pregnancy website encouraging vaginal birth24
• Expansion of dedicated twin clinic
• Creation of health care provider backup system for twin deliveries
Educational components of the initiative included lecture- and simulation-based health care provider education on twin vaginal birth.23 The expert lecture focused on the evidence supporting the safety of twin vaginal birth and approach to counseling and delivery. The 2-hour simulation consisted of a standardized patient counseling session, a breech extraction skills station, and a high-fidelity simulation emphasizing options for expedited delivery of a vertex-presenting second twin after vaginal birth of the first.23 An institutional grant supported the development of a dedicated multiple pregnancy website for patients and a complementary version for health care providers encouraging twin vaginal birth.24 The website outlines general aspects of twin pregnancy and emphasizes twin delivery through online videos including addressing medical indications for cesarean delivery for twins and familiarizing patients with methods of twin delivery including internal podalic version and breech extraction. To further support health care providers, we expanded and promoted a specialized multiple pregnancy clinic consisting of two maternal–fetal medicine providers and a dedicated nurse with expertise in multiple gestations. Within this clinic, we used available data to provide concise and objective counseling about mode of delivery for twins. Simply put, if the presenting twin was breech, we recommended cesarean delivery. If the presenting fetus was head down or vertex, we recommended vaginal birth based on the absence of demonstrated improved outcomes with cesarean delivery.21,22 If a patient declined an attempt at vaginal birth, we supported her autonomy and performed an elective cesarean delivery.17
Our institutional obstetric staff consisted of a diverse group of health care providers with varying levels of experience. We therefore dedicated renewed energy into a health care provider backup system to support obstetricians with less comfort and experience with twin vaginal birth. For this backup system, the weekly supervising maternal–fetal medicine provider for the hospital was available to provide in-house support for doctors requesting assistance with twin vaginal deliveries. The two health care providers spearheading the multiples clinic (J.N.R., C.B.) committed to serving as an additional level of support in the event that the attending of the week had limited experience or comfort with twin vaginal birth. Backup physicians provided in-room support to the primary clinician performing the delivery to facilitate a safe vaginal birth and promote development of health care provider skill.
For this retrospective study, we used hospital birth records to identify all women with twin pregnancies of at least 24 0/7 weeks of gestation delivering between January 2010 and December 2015. We excluded women with a contraindication to vaginal birth including those with placenta previa, a nonvertex presenting twin, prior transmural uterine surgery (including cesarean delivery), abdominal cerclage, monoamniotic gestation, category 3 fetal heart rate tracing at admission, or fetal anomalies precluding labor or vaginal birth such as an open neural tube defect or large omphalocele. We also excluded patients with an intrauterine fetal demise or lethal fetal anomaly of either twin. Patients were then classified according to year of delivery. Obstetric records were reviewed to obtain data regarding maternal, obstetric, fetal, and health care provider factors as previously described.25 Labor management and choice of delivery method throughout the study period were at the discretion of the delivering provider.
Over the study period, our institutional obstetric staff consisted of eight health care provider groups with 97 clinicians. The largest of these health care provider groups was the academic faculty practice consisting of 62 maternal–fetal medicine subspecialists and specialists in obstetrics and gynecology supervising the resident physicians present to perform all faculty deliveries. The remainder of the groups consisted of private practice providers performing deliveries at the hospital but seeing patients at other locations in the community. Resident physicians were involved in the deliveries of all of the faculty patients and present for the patients of private physicians at the discretion of the delivering obstetrician. Private physicians typically perform deliveries independently but can be supported by the in-house faculty physician on request. The intervention was developed within the infrastructure of the academic faculty practice. However, patients and health care providers in the private practice groups had access to these educational opportunities, online resources, and health care provider backup system throughout the study period.
Our primary outcome of interest was vaginal birth of both twins and included operative vaginal delivery for one or both twins or breech extraction of the second twin. Secondary outcomes of interest included rates of cesarean delivery overall and rates of elective cesarean delivery. Because we excluded patients with a contraindication to labor, all patients who deferred an attempt at a trial of labor in favor of cesarean delivery were classified as undergoing elective cesarean delivery. All admission notes were reviewed by the primary author (S.R.E.), with complete ascertainment as to the intended mode of delivery. Patients undergoing vaginal delivery followed by cesarean delivery of the nonpresenting twin, also referred to as combined delivery, were classified as undergoing cesarean delivery.
Additional secondary outcomes of interest include rates of maternal or neonatal morbidity. The composite maternal outcome was based on the definition of maternal morbidity used by Barrett et al and consisted of one or more of the following: death, hemorrhage (estimated blood loss greater than or equal to 1,500 mL or need for transfusion), infection, major procedure (hysterectomy, uterine artery embolization, intraperitoneal drain placement), readmission for infection or reoperation, need for dilation and evacuation for hemorrhage or infection, venous thromboembolism, small bowel obstruction or ileus, or intensive care unit admission.21,25 Infection was defined as a clinical endometritis (based on maternal fever greater than 38°C occurring more than 24 hours after delivery requiring antibiotics), significant wound cellulitis (wound infection prolonging hospitalization for more than 3 days with vaginal birth and 5 days with cesarean delivery or requiring readmission), intraabdominal or pelvic abscess, bacteremia, pneumonia, or Clostridium difficile colitis. We analyzed composite neonatal morbidity for the pregnancy overall and separately for presenting and nonpresenting twins. The composite neonatal outcome included one or more of the following: intubation for 2 or more days, birth trauma (including cephalohematoma, clavicular fracture, or long-bone fracture), need for blood transfusion, infection requiring antibiotics, intraventricular hemorrhage, and death.
The associations between year of delivery and patient characteristics of interest were evaluated with the Cochran Mantel-Haenszel test for trend for categorical variables and the Kruskal-Wallis test for continuous variables. We analyzed our primary outcome of interest in three ways. We first used least-squares means methodology from linear regression to calculate adjusted rates of outcomes of interest according to year of the study period. To better understand the effect of the intervention on rates of outcomes of interest, we then examined adjusted rates of outcomes of interest for the 3 years before and the 3 years after our 2013 interventions using similar methodology. To account for secular trends influencing rates of twin vaginal birth, we performed an interrupted time-series analysis using a linear autoregression model to estimate the rate of change before and after the intervention.26
All models contained all variables that modified the adjusted rate of the primary outcome by 10% or more and necessarily included parity. We report omnibus P values to examine the presence of an association between delivery year and the outcome of interest while adjusting for covariates. Statistical significance for all analyses was defined with a two-tailed P value <.05. The attending obstetricians based in the faculty practice, the residents they supervise, and the patients for whom they cared likely had more exposure to the patient- and health care provider-focused interventions. They also cared for patients with higher rates of maternal or fetal comorbidities that may differentially affect their probability of a successful vaginal birth. As opposed to adjusting for prenatal care provider, we chose to perform a stratified analysis according to whether a patient received care in the faculty practice compared with a private practice. This study was approved by the Partners Human Research Committee (Protocol #2012P001737, approved September 12, 2012). All analyses were performed with SAS 9.4.
Of 1,574 patients delivering twins at 24 weeks of gestation or greater, 897 patients (57%) were eligible for inclusion in the present study. Table 1 compares the maternal, fetal, and health care provider characteristics according to year of delivery. There were no significant changes in rates of nulliparity, fetal presentation, chorionicity, or other maternal or fetal characteristics across the study period. There was an expected increase in rates of prenatal care by maternal–fetal medicine subspecialists in general or in the specialized twin clinic specifically across the study period. When considering patient characteristics stratified by practice type, patients in the faculty practice tended to have higher rates of monochorionicity (27.4% vs 7.8%, P<.01), spontaneous conception (46.9% vs 31.6%, P<.01), and fetal anomalies (8.7% vs 1.5%, P<.01) compared with patients seen by private practitioners (Appendix 1, available online at http://links.lww.com/AOG/B101). Faculty patients delivered at a lower median gestational age (36 weeks of gestation, interquartile range 33–37 weeks of gestation) compared with patients of private health care providers (median 37 weeks of gestation, interquartile range 35–38 weeks of gestation, P<.01).
The frequency and adjusted rates of the primary and secondary outcomes of interest according to year of delivery are presented in Table 2. In an analysis adjusted for maternal age and parity, adjusted rates of vaginal delivery increased over the study period, with a nadir in 2010 (28.2%) to a peak in 2015 (55.7%), with a statistically significant difference in rates of vaginal birth across years (P<.01). Conversely, rates of elective cesarean delivery showed a stable decrease over the study period from a peak 59.5% in 2010 to a nadir of 27.6% in 2015 (P<.01 across years). Delivery by breech extraction increased over the time period from a nadir of 4.3% in 2010 to a peak of 13.3% in 2014 (P=.03 across years). Rates of combined delivery, operative vaginal delivery, and composite maternal morbidity were similar over time. Rates of composite neonatal morbidity for the nonpresenting twin specifically decreased over time even after controlling for gestational age. Table 3 presents these outcomes of interest stratified by health care provider type. The adjusted rate of vaginal delivery increased substantially for patients seen in the faculty practice (24.8% in 2010 vs 60.9% in 2016, P<.01 across years) with no statistically significant difference in the adjusted rate of vaginal delivery (P=.21) or elective cesarean delivery (P=.39) for lower risk patients cared for in the private practices.
We next analyzed outcomes in the 3 years before and after our 2013 intervention. The adjusted rates of primary and secondary outcomes of interest for the group overall and stratified by health care provider type are presented in Table 4. Rates of vaginal birth and breech extraction increased among faculty patients but not private patients. Similarly, rates of elective cesarean delivery decreased for faculty patients while remaining stable for patients of private health care providers. Even after controlling for gestational age, adjusted rates of neonatal morbidity decreased among patients of private practitioners after the 2013 intervention. This may reflect a decrease in the number of high-risk patients such as monochorionic twins seen by private health care providers after promotion of the specialized twin clinic.
Figure 1 displays the annual adjusted rate of vaginal birth for the cohort overall and stratified by health care provider type before and after the 2013 intervention (unadjusted rates demonstrated in Appendix 2, available online at http://links.lww.com/AOG/B101). In the interrupted time-series analysis controlling for secular trend, there was no significant effect of calendar year on the rate of twin vaginal birth in the time period before (1.35% annual increase, P=.76) or after (5.8% annual increase, P=.40) the intervention. Similarly, the rate of change of twin vaginal birth was similar before and after the intervention for faculty patients (4.7%, P=.31; 3.2%, P=.58) or private patients (3.6%, P=.28; 1.6%, P=.70). There was no statistically significant difference in the slope of the regression lines before and after the intervention for the overall cohort and for faculty and private patients. The similar rates of change before and after our intervention suggest the increase in the rates of twin vaginal birth is a product of secular change and not our intervention.
The increased rates of twin vaginal birth over the 6-year study period appear to reflect secular trend as opposed to a direct effect from our quality improvement intervention. These observed changes seem to be a product of declining rates of elective cesarean delivery rather than a greater likelihood of success among women laboring with twins. Notably, this increase in attempted vaginal birth was accompanied by higher rates of breech extraction and operative vaginal delivery without a rise in morbidity or combined delivery. Although our targeted intervention had no effect on the naturally rising institutional rate of twin vaginal birth, these observed secular changes do have important implications for clinicians and policymakers interested in decreasing rates of primary cesarean delivery.
National recommendations support attempts at twin vaginal birth to decrease rates of primary cesarean delivery and increase health care provider exposure to operative obstetrics.4 Our institutional observations offer evidentiary support to current guidelines. The decrease in elective cesarean delivery decreased the overall rate of primary cesarean delivery for twins while increasing exposure to breech extraction as anticipated by the guidelines.4 With recent concerns about the ability of guideline changes to improve outcomes, our findings offer reassurance about the safety and efficacy of this recommended practice change.27 The differential findings between rates of vaginal birth over time in our faculty compared with private practitioners warrant further consideration. We planned our stratified analysis to demonstrate an association between our initiative and changes in outcomes in the faculty practice conceptualizing private practice patients as a control group unexposed to the intervention. Comparing the naturally rising rates of vaginal birth among faculty patients with the stable rates seen among the lower risk patients of our private practice colleagues suggest that spontaneous clinical practice change may be less fluid within this population. With this in mind, a targeted intervention such as ours could potentially have an effect among patients in community settings as a direction for future research.
Our study is one of many to examine the effect of an intervention on rates of cesarean delivery, yet we failed to demonstrate a statistically significant effect of our intervention on the rate of change.12–14,28 These negative findings within the context of an apparently clinically significant result highlight the limitations of our study design. Our intervention was inspired by new evidence and society guidelines supporting attempts at vaginal birth.4 Implementing this intervention within the context of a natural increase in vaginal birth limits our power to change the rate of change. By comparison, the intervention by Rosenstein et al increased vaginal birth after cesarean delivery rates at an institution with a previously stable decline in vaginal birth after cesarean delivery rates.12 These authors demonstrate statistical significance by altering the direction and not just the magnitude of the rate of change. Similarly, the interventions decreasing the singleton cesarean delivery rate by Vadnais et al14 occurred over 8 years among 51,973 deliveries lending stability to the data points over time. It may be that our single-institution sample over a 6-year time period simply lacks power to demonstrate a statistically significant difference in the rate of change.26 Given its low cost and reproducibility, analyzing the effect of our intervention in an adequately powered cluster-randomized trial could overcome these potential limitations in methodology before abandoning health care provider support interventions such as ours entirely.13
The observed increase in rates of twin vaginal birth reflect secular trend as opposed to an effect of our quality improvement intervention.27 Despite these negative findings, this study addresses a challenge at the heart of the national public health debate on cesarean delivery. The majority of patients with twin pregnancies appear to prefer vaginal birth, but the increased risk of complications in twin relative to singleton pregnancies may dissuade patients from an attempted vaginal birth.25,29 These same concerns in conjunction with lack of health care provider comfort with operative vaginal delivery and breech extraction may inform health care provider preference for elective cesarean delivery for twins.18,19 Operative obstetrics offer an alternative to cesarean delivery in many settings, but may not be entertained despite appropriate clinical situations based solely on patient or health care provider preference.18,19,30 Rigorous prospective studies testing the ability of targeted patient education and health care provider support interventions to safely increase utilization of operative obstetrics in both community and academic settings may be a critical tool to curb the rising rate of cesarean delivery. Until the optimal approach is available, our findings lend evidentiary support to guidelines encouraging vaginal delivery of twins. Increasing access to twin vaginal birth appears to be an effective strategy to decrease primary cesarean delivery while ensuring obstetricians maintain the requisite obstetric skills to facilitate safe vaginal birth.19,30
1. Blanchette H. The rising cesarean delivery rate in America: what are the consequences? Obstet Gynecol 2011;118:687–90.
2. Clark SL, Belfort MA, Dildy GA, Herbst MA, Meyers JA, Hankins GD. Maternal death in the 21st century: causes, prevention, and relationship to cesarean delivery. Am J Obstet Gynecol 2008;199:36.e1–5.
3. Silver RM, Landon MB, Rouse DJ, Leveno KJ, Spong CY, Thom EA, et al. Maternal morbidity associated with multiple repeat cesarean deliveries. Obstet Gynecol 2006;107:1226–32.
4. Safe prevention of the primary cesarean delivery. Obstetric Care Consensus No. 1. American College of Obstetricians and Gynecologists. Obstet Gynecol 2014;123:693–711.
5. Kozhimannil KB, Law MR, Virnig BA. Cesarean delivery rates vary tenfold among US hospitals; reducing variation may address quality and cost issues. Health Aff (Millwood) 2013;32:527–35.
6. Wakeam E, Molina G, Shah N, Lipsitz SR, Chang DC, Gawande AA, et al. Variation in the cost of 5 common operations in the United States. Surgery 2017;162:592–604.
7. Grobman WA, Bailit JL, Rice MM, Wapner RJ, Varner MW, Thorp JM, et al. Can differences in obstetric outcomes be explained by differences in the care provided? The MFMUM Network APEX study. Am J Obstet Gynecol 2014;211:147.e1–16.
8. Molina G, Weiser TG, Lipsitz SR, Esquivel MM, Uribe-Leitz T, Azad T, et al. Relationship between cesarean delivery rate and maternal and neonatal mortality. JAMA 2015;314:2263–70.
9. Srinivas SK, Fager C, Lorch SA. Evaluating risk-adjusted cesarean delivery rate as a measure of obstetric quality. Obstet Gynecol 2010;115:1007–13.
10. Barber EL, Lundsberg LS, Belanger K, Pettker CM, Funai EF, Illuzzi JL. Indications contributing to the increasing cesarean delivery rate. Obstet Gynecol 2011;118:29–38.
11. Armstrong JC, Kozhimannil KB, McDermott P, Saade GR, Srinivas SK; Society for Maternal-Fetal Medicine Health Policy Committee. Comparing variation in hospital rates of cesarean delivery among low-risk women using 3 different measures. Am J Obstet Gynecol 2016;214:153–63.
12. Vadnais MA, Hacker MR, Shah NT, Jordan J, Modest AM, Siegel M, et al. Quality improvement initiatives lead to reduction in nulliparous term singleton cesarean delivery rate. Jt Comm J Qual Patient Saf 2017;43:53–61.
13. Chaillet N, Dumont A, Abrahamowicz M, Pasquier JC, Audibert F, Monnier P, et al. A cluster-randomized trial to reduce cesarean delivery rates in Quebec. N Engl J Med 2015;372:1710–21.
14. Rosenstein MG, Nijagal M, Nakagawa S, Gregorich SE, Kuppermann M. The association of expanded access to a collaborative midwifery and laborist model with cesarean delivery rates. Obstet Gynecol 2015;126:716–23.
15. Lee HC, Gould JB, Boscardin WJ, El-Sayed YY, Blumenfeld YJ. Trends in cesarean delivery for twin births in the United States: 1995–2008. Obstet Gynecol 2011;118:1095–101.
16. Ecker J. Elective cesarean delivery on maternal request. JAMA 2013;309:1930–6.
17. Bettes BA, Coleman VH, Zinberg S, Spong CY, Portnoy B, DeVoto E, et al. Cesarean delivery on maternal request: obstetrician-gynecologists' knowledge, perception, and practice patterns. Obstet Gynecol 2007;109:57–66.
18. Easter SR, Taouk L, Schulkin J, Robinson JN. Twin vaginal delivery: innovate or abdicate. Am J Obstet Gynecol 2017;216:484–8.e4.
19. Blickstein I. Delivery of vertex/nonvertex twins: did the horses already leave the barn? Am J Obstet Gynecol 2016;214:308–10.
20. Multifetal gestations: twin, triplet and higher-order multifetal pregnancies. Practice Bulletin No. 169. American College of Obstetricians and Gynecologists. Obstet Gynecol 2016;128:e131–46.
21. Barrett JF, Hannah ME, Hutton EK, Willan AR, Allen AC, Armson A, et al. A randomized trial of planned cesarean section or vaginal delivery for twin pregnancy. N Engl J Med 2013;369:1295–305.
22. Hutton EK, Hanna ME, Ross S, Joseph KS, Ohlsson A, Asztalos EV, et al. Maternal outcomes at 3 months after planned caesarean section versus planned vaginal birth for twin pregnancies in the Twin Birth Study: a randomized controlled trial. BJOG 2015;122:1653–62.
23. Easter SR, Gardner R, Barrett J, Robinson JN, Carusi D. Simulation to improve trainee knowledge and comfort about twin vaginal birth. Obstet Gynecol 2016;128(suppl 1):34–9S.
25. Easter SR, Robinson JN, Lieberman E, Carusi D. Association of intended route of delivery and maternal morbidity in twin pregnancy. Obstet Gynecol 2017;129:305–10.
26. Penfold RB, Zhang F. Use of interrupted time series analysis in evaluating health care quality improvement outcomes. Acad Pediatr 2013;13(suppl):S38–44.
27. Rosenbloom JI, Stout MJ, Tuuli MG, Woolfolk CL, Lopez JD, Macones GA, et al. New labor management guidelines and changes in cesarean delivery patterns. Am J Obstet Gynecol 2017;217:689.e1–8.
28. Plough AC, Galvin G, Li Z, Lipsizt SR, Alidina S, Henrigh NJ, et al. Relationship between labor and delivery unit management practices and maternal outcomes. Obstet Gynecol 2017;130:358–65.
29. Yee LM, Kaimal AJ, Houston KA, Wu E, Thiet MP, Nakagawa S, et al. Mode of delivery preferences in a diverse population of pregnant women. Am J Obstet Gynecol 2015;212:377.e1–24.
30. Dildy GA, Belfort MA, Clark SL. Obstetric forceps: a species on the brink of extinction. Obstet Gynecol 2016;128:436–9.
Supplemental Digital Content
© 2018 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.