Research ajog.org OBSTETRICS Early and total neonatal mortality in relation to birth setting in the United States, 2006-2009 Amos Grünebaum, MD; Laurence B. McCullough, PhD; Katherine J. Sapra, MPH; Robert L. Brent, MD, PhD, DSc (Hon); Malcolm I. Levene, MD, FRCP, FRCPH, F Med Sc; Birgit Arabin, MD; Frank A. Chervenak, MD OBJECTIVE: We examined neonatal mortality in relation to birth set- tings and birth attendants in the United States from 2006 through 2009. STUDY DESIGN: Data from the Centers for Disease Control and Pre- ventionelinked birth and infant death dataset in the United States from 2006 through 2009 were used to assess early and total neonatal mortality for singleton, vertex, and term births without congenital malformations delivered by midwives and physicians in the hospital and midwives and others out of the hospital. Deliveries by hospital midwives served as the reference. RESULTS: Midwife home births had a significantly higher total neonatal mortality risk than deliveries by hospital midwives (1.26 per 1000 births; relative risk [RR], 3.87 vs 0.32 per 1000; P < .001). Midwife home births of 41 weeks or longer (1.84 per 1000; RR, 6.76 vs 0.27 per 1000; P < .001) and midwife home births of women with a first birth (2.19 per 1000; RR, 6.74 vs 0.33 per 1000; P < .001) had significantly higher risks of total neonatal mortality than deliveries by hospital midwives. In midwife home births, neonatal mortality for first births was twice that of subsequent births (2.19 vs 0.96 per 1000; P < .001). Similar results were observed for early neonatal mortality. The excess total neonatal mortality for midwife home births compared with midwife hospital births was 9.32 per 10,000 births, and the excess early neonatal mortality was 7.89 per 10,000 births. CONCLUSION: Our study shows a significantly increased total and early neonatal mortality for home births and even higher risks for women of 41 weeks or longer and women having a first birth. These significantly increased risks of neonatal mortality in home births must be disclosed by all obstetric practitioners to all pregnant women who express an interest in such births. Key words: birth attendants, birth settings, home births, midwives, neonatal mortality, physicians Cite this article as: Grünebaum A, McCullough LB, Sapra KJ, et al. Early and total neonatal mortality in relation to birth setting in the United States, 2006-2009. Am J Obstet Gynecol 2014;211:390.e1-7. D espite the increase in home births in the United States over the last decade,1 the safety of home births has remained controversial. In our previous publication using the US natality data,2 we reported that home birth has an increased relative risk of 5 minute Apgar scores of zero and of seizures and other adverse neurological outcomes. Although a 5 minute Apgar score of zero is related to neonatal mortality,3 the linked birth/infant datasets (for live births and infant deaths) allow for the direct assessment of neonatal mortality relative to birth setting and attendant.4 From the Department of Obstetrics and Gynecology, Weill Medical College of Cornell University (Drs Grünebaum, Brent, and Chervenak), and Department of Epidemiology, Mailman School of Public Health, Columbia University (Ms Sapra), New York, NY; Center for Medical Ethics and Health Policy, Baylor College of Medicine, Houston, TX (Dr McCullough); Alfred I. DuPont Hospital for Children, Thomas Jefferson University, Wilmington, DE (Dr Brent); Division of Pediatrics and Child Health, University of Leeds, Leeds, England, UK (Dr Levene); and Center for Mother and Child, Philipps University, Marburg, and Clara Angela Foundation, Berlin, Germany (Dr Arabin). Received Jan. 19, 2014; revised Feb. 12, 2014; accepted March 19, 2014. The authors report no conflict of interest. Presented at the 34th annual meeting of the Society for Maternal-Fetal Medicine, New Orleans, LA, Feb. 3-8, 2014. Reprints: Amos Grünebaum, MD, Department of Obstetrics and Gynecology, Weill Medical College of Cornell University, New York, NY 10021. amosgrune@gmail.com 0002-9378/$36.00 ª 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2014.03.047 390.e1 American Journal of Obstetrics & Gynecology OCTOBER 2014 The purpose of this study therefore was to examine early, total, and excess neonatal mortality rates for singleton term births without congenital malformations by birth setting and birth attendant (hospital physician, hospital midwife, freestanding birth center, midwife, home midwife, and other for home births). M ATERIALS AND M ETHODS The 1989 revision of the US Standard Certificate of Live Birth provides additional detail for out-of-hospital births and makes it possible to distinguish among out-of-hospital births at home, in a birthing center, or other specified location.1 In contrast to the birth certificate files, which provide information on delivery, it is necessary to go to the Centers for Disease Control and Prevention (CDC)elinked birth/infant death dataset (for live births and infant deaths) to analyze neonatal mortality. Obstetrics ajog.org Research TABLE 1 Characteristics Total Hospital physician Hospital MW Freestanding BC MW Home all MW Home other Total 12,709,881 1,096,555 39,523 61,993 28,119 Parity total 12,658,411 1,090,290 39,254 61,051 27,643 Para 0 5,193,419 (41) 432,018 (39.6) 14,036 (35.8) 13,884 (22.7) 5024 (18.2) Para 1 7,464,992 (59) 658,272 (60.4) 25,218 (64.2) 47,167 (77.3) 22,619 (81.8) 28,119 GA total, wks 39,523 61,993 41 2,006,179 (15.8) 223,329 (20.4) 10,419 (26.4) 17,572 (28.3) 7693 (27.4) 42 810,809 (6.4) 84,512 (7.7) 3425 (8.7) 5913 (9.5) 3023 (10.8) BW total 12,709,881 12,709,881 1,096,555 1,096,555 39,523 4000 g 1,120,028 (8.8) 97,893 (8.9) 6626 (16.8) 4500 g 151,128 (1.2) 11,093 (1.0) 1171 (3) Maternal age total, y 12,709,881 1,096,555 <25 4,392,994 (34.6) 449,782 (41) 25-29 3,610,725 (28.4) 30-34 2,920,352 (23) 35 1,785,860 (14.1) R/E total 12,622,924 39,523 61,993 13,653 (22) 2821 (4.6) 61,993 28,119 5387 (19.2) 1256 (4.5) 28,119 9296 (23.5) 10,102 (16.3) 6097 (21.7) 317,099 (28.9) 13,385 (33.9) 19,292 (31.1) 8315 (29.6) 218,075 (19.9) 10,864 (27.5) 18,916 (30.5) 7602 (27) 111,599 (10.2) 5978 (15.1) 13,683 (22.1) 6105 (21.7) 1,089,006 39,298 61,097 27,666 NH white 6,939,531 (55) 572,702 (52.6) 31,552 (80.3) 55,466 (90.8) 22,269 (80.5) NH black 1,710,594 (13.6) 143,371 (13.2) 1835 (4.7) 1132 (1.9) 2316 (8.4) NH other 846,850 (6.7) 75,083 (6.9) 1042 (2.7) 1263 (2.1) 809 (2.9) Hispanic 3,125,949 (24.8) 297,850 (27.4) 4869 (12.4) 3236 (5.3) 2272 (8.2) BC, birthing center; BW, birthweight; GA, gestational age; MW, midwife; NH, non-Hispanic; R/E, race/ethnicity. Grünebaum. Total neonatal mortality in relation to birth setting. Am J Obstet Gynecol 2014. This dataset (linked file) is generally the preferred source for infant and neonatal mortality in the United States.4 It contains detailed information for the approximately 4 million births in the United States each year, including birth setting, birth attendant, and neonatal mortality.5 Period-linked files use all births in a year as the denominator and all deaths in a year as the numerator, regardless of when the birth occurred (eg, if the birth was in late 2008, then neonatal death could have been 2008 or 2009 but counted in the 2008 numerator only if the death occurred in 2008). The 2006-2009 period-linked birth/ infant deaths dataset was analyzed to examine early (deaths <7 days of life) and total (deaths <28 days of life) neonatal mortality in term singleton births ( 37 weeks and newborn weight of 2500 g) without documented congenital malformations by birth setting (hospital, birthing center, home) and provider: hospital midwife (certified nurse midwives [CNMs] and other midwife [MW]; hospital MW), hospital physician (MD or DO), free-standing birthing center midwife (CNM and other MW), home midwife (CNM and other MW, home MW), home other (including emergency situations, such as unattended births and “any other person delivering the baby, such as a husband or family member, emergency medical technician, or taxi driver”).1 Total neonatal mortality (tNNM) is defined as the death of a live-born neonate before 28 days of life, and early neonatal mortality (eNNM) is defined as neonatal death before 7 days of life. We also examined the relative risks associated with delivery by provider and setting compared with hospital midwives. Excess neonatal mortality is defined as the increased number of neonatal deaths per 10,000 births by provider and setting, using hospital-based midwife deliveries as the reference group. Data on patient characteristics included parity, race and ethnicity, maternal age, and clinical factors such as neonatal weight and weeks of gestation at delivery. We excluded infants if they met any of the following criteria: birth attendant type was not recorded; birth place was anywhere else but the hospital, home, or freestanding birthing center, or not recorded; gestational age was less than 37 weeks or not recorded; birthweight was less than 2500 g or not recorded; multiple gestations; any congenital anomaly, OCTOBER 2014 American Journal of Obstetrics & Gynecology 390.e2 Research Obstetrics ajog.org TABLE 2 Term neonatal mortality (0-27 days) by birth setting, birth attendant, and parity and postdates Neonatal mortality Per 1000 (n/total) RR (95% CI) Hospital midwife 0.32 (356/1,096,555) 1 Hospital physician 0.55 (6977/12,709,881) 1.69 (1.52e1.88) Freestanding BC midwife 0.59 (23/39,523) 1.81 (1.19e2.75) Home midwife 1.26 (78/61,993) 3.87 (3.03e4.95) Home other 1.87 (52/28,119) 5.75 (4.31e7.68) Total 0.54 (7486/13,936,071) P value Neonatal mortality (para ¼ 0) Hospital midwife 0.33 (141/432,018) 1 Hospital physician 0.57 (2946/5,193,19) 1.74 (1.47e2.06) Freestanding BC midwife 1.01 (14/14,036) 3.1 (1.8e5.36) Home midwife 2.19 (30/13,884) 6.74 (4.55e9.96) Home other 3.01 (15/5024) 9.26 (5.45e15.72) Total 0.56 (3146/5,658,381) Neonatal mortality (para >0) Hospital midwife 0.32 (213/658,272) 1 Hospital physician 0.53 (3981/7,464,992) 1.65 (1.43e1.89) Freestanding BC midwife 0.36 (9/25,218) 1.10 (0.57e2.15) Home midwife 0.96 (45/47,167) 2.97 (2.16e4.09) Home other 1.43 (32/22,619) 4.41 (3.05e6.38) Total 0.52 (4280/8,218,268) NS Neonatal mortality (<41 wks) Hospital midwife 0.34 (295/873,226) 1 Hospital physician 0.55 (5862/10,703,702) 1.62 (1.44e1.82) Freestanding BC midwife 0.48 (14/29,104) 1.44 (0.85e2.46) Home midwife 1.02 (45/44,421) 3.03 (2.22e4.14) Home other 2.12 (43/20,426) 6.29 (4.57e8.64) Total 0.54 (6259/11,670,879) NS Neonatal mortality ( 41 wks) Hospital midwife 0.27 (61/223,329) 1 Hospital physician 0.56 (1116/2,006,179) 2.04 (1.58e2.64) Freestanding BC midwife 0.86 (9/10,419) 3.17 (1.58e6.38) Home midwife 1.84 (32/17,572) 6.76 (4.42e10.36) Home other 1.19 (9/7693) 4.35 (2.17e8.72) Total 0.54 (1227/2,265,192) BC, birthing center; CI, confidence interval; NS, not significant; RR, relative risk. Grünebaum. Total neonatal mortality in relation to birth setting. Am J Obstet Gynecol 2014. Down syndrome, or other chromosomal disorder was confirmed or pending; and a resident of a foreign country. Because nonidentifiable data from a publicly available dataset were used, our study was not considered human 390.e3 American Journal of Obstetrics & Gynecology OCTOBER 2014 subjects research and did not require review by the Weill Medical College of Cornell University Institutional Review Board. Analysis of data We analyzed tNNM (deaths <28 days of age) and eNNM (deaths <7 days of age). We computed relative risks (RRs) for all patients with a first birth (para ¼ 0) and with a second or higher order birth (para of 1), and for term and postterm ( 41 weeks) pregnancies. Hospital midwives (hospital MW) included both CNMs and other midwives and served as the reference group for the estimation of early, total, and excess neonatal mortality. A freestanding birthing center midwife (CNM and other MW) and home midwives (home MW) include both CNMs and other midwives. Home ‘others’ includes others identified by the CDC database as attending home births, including family members, emergency medical service, or police, and taxi drivers as well as unattended births. Data were extracted using SAS version 9.3 (SAS Institute, Cary, NC) and compiled in Excel (Microsoft, Redmond, WA). The RRs and 95% confidence intervals were computed in SAS version 9.3 (SAS Institute). Excess mortality was computed in OpenEpi.6 R ESULTS Table 1 shows the characteristics of the study population. There were 13,936,071 deliveries between 2006 and 2009 that met study criteria. The majority of deliveries were by physicians in the hospital (91.2%) followed by hospital midwives (7.78%), home midwives (0.44%), midwives in freestanding birthing centers (0.28%), and home deliveries by others (0.2%). When compared with hospital births, home births were more likely to have a postdate pregnancy of 41 or more weeks: 28.3% for home births midwives vs 20.4% for hospital midwives and 15.7% for hospital physicians (P < .001); and 42 or more weeks: 9.5% in home births midwives vs 7.7% for hospital midwives and 6.4% for hospital physicians (P < .001). Women delivered Obstetrics ajog.org TABLE 3 Term early neonatal mortality (0-6 days) by birth setting, birth attendant, parity, and postdates Early neonatal mortality Per 1000 (n/total) RR (95% CI) Hospital midwife 0.14 (155/1,096,555) 1 Hospital physician 0.29 (3648/12,709,881) 2.04 (1.73e2.39) Freestanding BC midwife 0.46 (18/39,523) 3.26 (2.01e5.31) Home midwife 0.93 (58/61,993) 6.6 (4.88e8.93) Home other 1.65 (46/28,119) 11.73 (8.45e16.28) Total 0.28 (3925/13,936,071) Early neonatal mortality (P ¼ 0) Hospital midwife 0.13 (58/432,018) 1 Hospital physician 0.31 (1634/5,193,419) 2.35 (1.81e3.05) Freestanding BC midwife 0.8 (11/14,036) 5.94 (3.13e11.27) Home midwife 1.82 (25/13,884) 13.62 (8.54e21.72) Home other 2.61 (13/5024) 19.5 (10.71e35.48) Total 0.31 (1741/5,658,381) Early neonatal mortality (P >0) Hospital midwife 0.14 (95/658,272) 1 Hospital physician 0.27 (1980/7,464,992) 1.84 (1.5e2.27) Freestanding BC midwife 0.28 (7/25,218) 1.93 (0.9e4.16) Home midwife 0.66 (31/47,167) 4.62 (3.09e6.91) Home other 1.25 (28/22,619) 8.71 (5.73e13.25) Total 0.26 (2141/8,218,268) Early neonatal mortality (<41 wks) Hospital midwife 0.15 (127/873,226) 1 Hospital physician 0.29 (3066/10,703,702) 1.97 (1.65e2.35) Freestanding BC midwife 0.35 (10/29,104) 2.4 (1.27e4.55) Home midwife 0.8 (35/44,421) Home other 1.88 (38/20,426) Total 0.28 (3276/11,670,879) 5.48 (3.78e7.96) 12.9 (9e18.51) Early neonatal mortality ( 41 wks) Hospital midwife 0.12 (27/223,329) 1 Hospital physician 0.29 (583/2,006,179) 2.36 (1.61e3.47) Freestanding BC midwife 0.77 (8/10,419) 6.25 (2.85e13.74) 10.28 (5.88e17.98) Home midwife 1.26 (22/17,572) Home other 1.06 (8/7693) Total 0.29 (648/2,265192) 8.59 (3.93e18.79) BC, birthing center; CI, confidence interval; RR, relative risk. Grünebaum. Total neonatal mortality in relation to birth setting. Am J Obstet Gynecol 2014. at home by midwives were more likely to be 35 years old or older and more likely to have macrosomic infants. Women delivered by midwives at home were more likely to be non-Hispanic white when compared with hospital births. Research Table 2 shows the total neonatal mortality (prior to 28 days) and relative risks by parity and weeks before and after 41 weeks gestation by the 5 groups of settings and attendants. Midwife home births had a significantly higher nearly 4-fold total neonatal mortality risk when compared with those delivered by hospital midwives (1.26 per 1000 births; RR, 3.87 vs 0.32 per 1000; P < .001). Midwife home births of women with a first birth had a significantly higher nearly 7-fold risk of total neonatal mortality than those by hospital midwives (2.19 per 1000; RR, 6.74 vs 0.33 per 1000; P < .001) and a neonatal mortality more than twice that of those with a subsequent birth (2.19 vs 0.96 per 1000; P < .001). Midwife home births of 41 or more weeks had a significantly higher nearly 7-fold risk of total neonatal mortality than those delivered by hospital midwives (1.84 per 1000; RR, 6.76 vs 0.27 per 1000; P < .001). Table 3 shows the term early neonatal mortality (0-6 days) and relative risks by parity and weeks before and after 41 weeks gestation by the 5 groups of settings and attendants. Midwife home births had a significantly higher nearly 7-fold early neonatal mortality risk when compared with those delivered by hospital midwives (0.93 per 1000 births; RR, 6.6 vs 0.14 per 1000; P < .001). Midwife home births of women with a first birth had a significantly higher 13- to 14-fold risk of early neonatal mortality than those by hospital midwives (1.82 per 1000; RR, 13.62 vs 0.13 per 1000; P < .001) and an early neonatal mortality nearly 3 times that of those with a subsequent birth (1.82 vs 0.66 per 1000; P < .001). Midwife home births of 41 or more weeks had a significantly higher nearly 10-fold risk of early neonatal mortality than those delivered by hospital midwives (1.26 per 1000; RR, 10.28 vs 0.12 per 1000; P < .001). Table 4 shows the excess early and total neonatal mortality per 10,000 births for the 5 groups with hospital MW serving as the reference group. Home births by others had an excess total OCTOBER 2014 American Journal of Obstetrics & Gynecology 390.e4 Research Obstetrics ajog.org TABLE 4 Term excess early and total neonatal mortality NNM Birth setting and provider eNNM Per 10,000 births Excess tNNM Excess eNNM Per 10,000 births (95% CI) Per 10,000 births (95% CI) Hospital midwife 3.2 1.4 0 0 Hospital physician 5.5 2.9 2.24 (1.88e2.6) 1.46 (1.22e1.7) Freestanding BC midwife 5.9 4.6 2.62 (0.21e5.03) 3.19 (1.07e5.32) Home midwife 12.6 9.3 9.32 (6.51e12.1) 7.89 (5.48e10.30) Home other 18.7 16.5 15.42 (10.37e20.48) 15.12 (10.37e19.87) BC, birthing center; CI, confidence interval; eNNM, excess neonatal mortality; tNNM, total neonatal mortality. Grünebaum. Total neonatal mortality in relation to birth setting. Am J Obstet Gynecol 2014. neonatal mortality of 15.42 per 10,000, whereas neonates delivered by midwives at home had an excess of 9.32 per 10,000 births when compared with midwife hospital births. C OMMENT There has been an increase in home births in the United States over the last decade.1 Studies have shown purported advantages of home births including fewer interventions, lower cesarean delivery rates, and less use of medications or analgesia.7-10 The decrease in obstetric interventions in home births should be balanced against the increased neonatal risks.10,11 This study on early and total neonatal mortality utilized the largest and most reliable dataset on neonatal mortality for live births in the United States, which uses “...the many additional variables available from the birth certificate to conduct more detailed analyses of infant mortality patterns.”4 Our analysis shows a substantially increased risk of neonatal deaths when delivery occurred outside the hospital. There is a clear pattern in our study: total and early neonatal mortality is significantly increased in home births. Nulliparous patients and patients at 41 or more weeks’ gestation have even higher neonatal mortality risks in the home setting when compared with the hospital. The higher neonatal mortality rate for hospital physicians when compared with hospital midwives almost certainly reflects the fact that hospital physicians deliver a higher-risk population than hospital midwives and deliver patients with complications transferred from the hospital midwifery service to the hospital physician service. Our study reports on the largest population to date comparing neonatal mortality among different birth settings and providers. Other studies have found similar patterns of adverse neonatal outcomes in home births such as an increase in Apgar scores of zero, low Apgar scores, higher neonatal mortality, and an increase in hypoxic ischemic encephalopathy.2,10,12-14 Some studies conducted outside the United States reported similar7-9,15,16 or better17-19 outcomes in home births when compared with hospital births, whereas other studies from outside and within the United States have shown increased neonatal morbidity and mortality in home births.2,10,12-14,17,20,21 Home birth studies from outside the United States such as Australia, The Netherlands, and the United Kingdom are of limited comparability with those in the United States because, contrary to the United States, in these countries this birth option is integrated more fully into the medical care system.10 Patients with advanced maternal age have worse outcomes and have a higher risk of reaching 41 and 42 weeks.22,23 Advanced maternal and gestational age, as well as macrosomia, has been shown to increase neonatal mortality and 390.e5 American Journal of Obstetrics & Gynecology OCTOBER 2014 morbidity,22-26 especially in nulliparous patients,27 and meconium aspiration syndrome.28 Those induced at an earlier gestational age had better neonatal outcomes.29 Pregnancies beyond 41 weeks’ gestation, nulliparous patients, and mothers who are 35 years and older have an increased risk of neonatal mortality.24,26,27,30 This may partially explain the increased neonatal mortality among home births, in which there are more older mothers and pregnancies who deliver beyond 41 weeks. Patients delivering at home have no access to electronic fetal monitoring, which has been found to decrease neonatal mortality.31 The American Academy of Pediatrics in their home birth policy32 recommends that planned home births should not exceed 41 weeks. However, 28.3% of midwife home births in our study exceeded that threshold, and according to the American Academy of Pediatrics statement, these home births should have been performed only in the hospital. Because of the increased risks of neonatal mortality in births beyond 41 weeks, midwives should not plan to deliver patients beyond 41 weeks at home and instead they should immediately transfer their patients. Malloy33 reported an increased term vaginal delivery neonatal mortality rate of 1.60 (89 of 55,634) for home midwives when compared with 0.5 of 1000 (614 of 1,237,129) for hospital certified midwives (RR, 3.2). Accordingly, we observed a similarly higher total neonatal mortality of 1.26 of 1000 (RR, 3.87) in midwife home deliveries. Ananth et al34 reported that electronic fetal monitoring appears to be associated with a modest decline in neonatal mortality. Considering that electronic fetal monitoring is not available in home births, this may explain in part the increase in neonatal mortality in home births. In our study, hospital births included about 40% of black or Hispanic patients as compared with about 7% blacks and Hispanic in home births. Mathews and MacDorman35 have shown that neonatal mortality is significantly higher in black Obstetrics ajog.org and Hispanic patients. These data suggest that the increased mortality at home births that we have documented may be understated. The strength of our study is that we used the linked birth/infant death dataset (period-linked file), which is generally the preferred source for infant and neonatal mortality in the United States.4 According to CDC data,35 “almost all the home births attended by certified nursemidwives/certified midwives (98%) or ‘other’ midwives (99%) were planned,”36 and therefore, it is appropriate to use midwife-attended home births as proxy for planned home births. There are some limitations in our study. Criticism has been expressed about certain data collected in birth and death certificates,37 although others believe that the data are reliable.38-40 As in our previous study,2 our results likely underestimate the actual neonatal mortality rates in home births because the higher adverse neonatal outcomes for patients transferred from home to the hospital are counted in the CDC-linked data as hospital and not home birth neonatal outcomes. Hildingsson et al41 previously recommended for Swedish records that adding information on whether this was a planned home birth for women transferred before birth could produce better statistics with the opportunity to follow up women who choose to give birth outside a hospital. On the 2003 revised US birth certificate, information on planned and unplanned home birth is collected, but information on whether a birth in the hospital is the result of a transferred home delivery is not collected. We believe that US birth certificate data would be improved by using a new revision that specifies those who originally planned a home birth and then were transferred to the hospital. Conclusions Our study shows that home births are at increased risk for early and total neonatal mortality, a risk that further increases for women with a first birth and pregnancies of 41 or more weeks’ gestation. We emphasize that this increased risk is a function of the out-ofhospital setting rather than the provider. Patients considering a home birth should appreciate that home births are associated with not only increased neonatal deaths but also other increased neonatal risks such as low Apgar scores2 and an increased risk of neonatal hypoxic ischemic encephalopathy.12 As part of the informed consent process, obstetric providers should recommend strongly for hospital births and against planned home births with evidence-based recommendations. They should explain that these recommendations are based on the documented increased risk of neonatal mortality and morbidity in home births. Doing so is essential for obstetric providers to fulfill their professional responsibility and to empower the autonomy of pregnant women in the informed consent process by providing clinically important information.42 Physicians and other health care providers have a professional responsibility to understand, identify, and address the root cause motivating patients’ desire for out-of-hospital birth by providing evidence-based compassionate hospital care, improve hospital settings, address obstetric interventions,43,44 and provide excellent, supportive, and nonjudgmental hospital care to women transported from a planned home birth.45,46 REFERENCES 1. MacDorman MF, Mathews TJ, Declerq E. Home births in the United States, 1990-2009. NCHS Data Brief 2012;84:1-8. 2. Grünebaum A, McCullough LB, Sapra KJ, et al. Apgar score of zero at five minutes and neonatal seizures or serious neurologic dysfunction in relation to birth setting. Am J Obstet Gynecol 2013;209:323.e1-6. 3. Li F, Wu T, Lei X, et al. The Apgar score and infant mortality. PLoS One 2013;8:e69072. 4. Mathews TJ, MacDorman MF. Infant mortality statistics from the 2009 period linked birth/ infant death data set, National vital statistics reports, vol. 61, no. 8. Hyattsville, MD: National Center for Health Statistics; 2013. 5. National Center for Health Statistics. Vital statistics data: birth data files. Available at: http://www.cdc.gov/nchs/data_access/VitalSt atsOnline.htm. Accessed Feb. 21, 2014. 6. Dean AG, Sullivan KM, Soe MM. OpenEpi: Open source epidemiologic statistics for public health, version 2.3.1. Updated June 23, 2011. Research Available at: www.OpenEpi.com. Accessed March 10, 2013. 7. Woodcock HC, Read AW, Bower C, Stanley FJ, Moore DJ. A matched cohort study of planned home and hospital births in Western Australia 1981-1987. Midwifery 1994;10: 125-35. 8. Janssen PA, Lee SK, Ryan EM, et al. Outcomes of planned home births versus planned hospital births after regulation of midwifery in British Columbia. CMAJ 2002;166:315-23. 9. Lindgren HE, Radestad IJ, Christensson K, Hildingsson IM. Outcome of planned home births compared to hospital births in Sweden between 1992 and 2004. A population based register study. Acta Obstet Gynecol 2008;87: 751-9. 10. Cheng YW, Snowden JM, King TL, Caughey AB. Selected perinatal outcomes associated with planned home births in the United States. Am J Obstet Gynecol 2013;209: 325.e1-8. 11. Grünebaum A, McCullough LB, Chervenak FA. Interventions at home births. Am J Obstet Gynecol 2014;210:487-8. 12. Wasden S, Perlman J, Chasen S, Lipkind H. 506: Home birth and risk of neonatal hypoxic ischemic encephalopathy. Am J Obstet Gynecol 2014;210(Suppl):S25. 13. Wax JR, Lucas FL, Lamont M, Cartin A, Blackstone J. Maternal and newborn outcomes in planned home birth vs planned hospital births: a metaanalysis. Am J Obstet Gynecol 2010;203: 243.e1-8. 14. Wax JR, Pinette MG, Cartin A, Blackstone J. Maternal and newborn morbidity by birth facility among selected United States 2006 low-risk births. Am J Obstet Gynecol 2010;202:152. 15. Hutton EK, Reitsma AH, Kaufman K. Outcomes associated with planned home and planned hospital births in low-risk women attended by midwives in Ontario, Canada, 20032006: a retrospective cohort study. Birth 2009;36:180-9. 16. Mori R, Dougherty M, Whittle M. An estimation of intrapartum-related perinatal mortality rates for booked home births in England and Wales between 1994 and 2003. BJOG 2008;115:554-9. 17. Pang JW, Heffelfinger JD, Huang GJ, Benedetti T, Weiss NS. Outcomes of planned home births in Washington State: 1989-1996. Obstet Gynecol 2002;100:253-9. 18. Janssen PA, Saxell L, Page LA, Klein MC, Liston RM, Lee SK. Outcomes of planned home birth with registered midwife versus planned hospital birth with midwife or physician. CMAJ 2009;181:377-83. 19. Kennare RM, Keirse MJ, Tucker GR, Chan AC. Planned home and hospital births in South Australia, 1991-2006: differences in outcomes. Med J Aust 2010;192:76-80. 20. Bastian H, Keirse MJ, Lancaster PA. Perinatal death associated with planned home birth in Australia: population based study. BMJ 1998;317:384-8. OCTOBER 2014 American Journal of Obstetrics & Gynecology 390.e6 Research Obstetrics 21. Mehl-Madrona L, Mehl-Madrona M. Physician- and midwife-attended home births. Effects of breech, twin, and post-dates outcome data on mortality rates. J Nurse Midwifery 1997;42:91-8. 22. Caughey AB, Stotland NE, Washington AE, Escobar GJ. Who is at risk for prolonged and postterm pregnancy? Am J Obstet Gynecol 2009;200:683.e1-5. 23. Caughey AB, Stotland NE, Washington AE, Escobar GJ. Maternal and obstetric complications of pregnancy are associated with increasing gestational age at term. Am J Obstet Gynecol 2007;196:155.e1-6. 24. Reddy UM, Bettegowda VR, Dias T, YamadaKushnir T, Ko CW, Willinger M. Term pregnancy: a period of heterogeneous risk for infant mortality. Obstet Gynecol 2011;117:1279-87. 25. De Los Santos-Garate AM, Villa-Guillen M, Villanueva-García D, Vallejos-Ruíz ML, MurguíaPeniche MT; NEOSANO’s Network. Perinatal morbidity and mortality in late-term and postterm pregnancy. NEOSANO Perinatal Network’s experience in Mexico. J Perinatol 2011;31:789-93. 26. Hilder L, Costeloe K, Thilaganathan B. Prolonged pregnancy: evaluating gestation-specific risks of fetal and infant mortality. Br J Obstet Gynaecol 1998;105:169-73. 27. Hilder L, Sairam S, Thilaganathan B. Influence of parity on fetal mortality in prolonged pregnancy. Eur J Obstet Gynecol Reprod Biol 2007;132:167-70. 28. Kaimal AJ, Little SE, Odibo AO, et al. Costeffectiveness of elective induction of labor at 41 weeks in nulliparous women. Am J Obstet Gynecol 2011;204:137.e1-9. 29. Darney BG, Snowden JM, Cheng YW, et al. Elective induction of labor at term compared ajog.org with expectant management: maternal and neonatal outcomes. Obstet Gynecol 2013;122: 761-9. 30. Shapiro H, Lyons E. Late maternal age and postdate pregnancy. Am J Obstet Gynecol 1989;160:909-12. 31. Chen H-Y, Chauhan SP, Ananth CV, Vintzileos AM, Abuhamad AZ. Electronic fetal heart rate monitoring and its relationship to neonatal and infant mortality in the United States. Am J Obstet Gynecol 2011;204:491. e1-10. 32. American Academy of Pediatrics. Policy statement: planned home birth. Pediatrics 2013;131:1016-20. 33. Malloy MH. Infant outcomes of certified nurse midwife attended home births: United States 2000 to 2004. J Perinatol 2010;30:622-7. 34. Ananth CV, Chauhan SP, Chen HY, D’Alton ME, Vintzileos AM. Electronic fetal monitoring in the United States: temporal trends and adverse perinatal outcomes. Obstet Gynecol 2013;121:927-33. 35. Mathews TJ, MacDorman MF. Infant mortality statistics from the 2007 period linked birth/ infant death data set. Natl Vital Stat Rep 2011;59:1-30. 36. MacDorman MF, Declerq E, Mathews TJ. United States home births increase 20 percent from 2004 to 2008. Birth 2011;38:185-90. 37. Vinikoor LC, Messer LC, Laraia BA, Kaufman JS. Reliability of variables on the North Carolina birth certificate: a comparison with directly queried values from a cohort study. Paediatr Perinat Epidemiol 2010;24:102-12. 38. DiGiuseppe DL, Aron DC, Ranbom L, Harper DL, Rosenthal GE. Reliability of birth certificate data: a multi-hospital comparison to 390.e7 American Journal of Obstetrics & Gynecology OCTOBER 2014 medical records information. Mat Child Health J 2002;6:169-79. 39. Zollinger TW, Przybylski MJ, Gamache RE. Reliability of Indiana birth certificate data compared to medical records. Ann Epidemiol 2006;16:1-10. 40. Northam S, Knapp TR. The reliability and validity of birth certificates. J Obstet Gynecol Neonatal Nurs 2006;35:3-12. 41. Hildingsson IM, Lindgren HE, Haglund B, Rådestad IJ. Characteristics of women giving birth at home in Sweden: a national register study. Am J Obstet Gynecol 2006;195: 1366-72. 42. Chervenak FA, McCullough LB, Grunebaum A, Arabin B, Levene MI, Brent RL. Planned home birth in the United States and professionalism: a critical assessment. J Clin Ethics 2013;24:184-91. 43. Grunebaum A, Chervenak F, Skupski D. Effect of a comprehensive obstetric patient safety program on compensation payments and sentinel events. Am J Obstet Gynecol 2011;204: 97-105. 44. Grunebaum A, Dudenhausen J, Chervenak FA, Skupski D. Reduction of cesarean delivery rates after implementation of a comprehensive patient safety program. J Perinat Med 2013;41:51-5. 45. Chervenak FA, McCullough LB, Brent RL, Levene MI, Arabin B. Planned home birth: the professional responsibility response. Am J Obstet Gynecol 2013;208:31-8. 46. Chervenak FA, McCullough LB, Grunebaum A, Arabin B, Levene MI, Brent RL. Planned homebirth: a violation of the best interests of the child standards? Pediatrics 2013;132:921-3. Research www.AJOG .org OBSTETRICS Apgar score of 0 at 5 minutes and neonatal seizures or serious neurologic dysfunction in relation to birth setting Amos Grünebaum, MD; Laurence B. McCullough, PhD; Katherine J. Sapra, MPH; Robert L. Brent, MD, PhD, DSc (Hon); Malcolm I. Levene, MD, FRCP, FRCPH, FMedSc; Birgit Arabin, MD; Frank A. Chervenak, MD OBJECTIVE: To examine the occurrence of 5-minute Apgar scores of 0 and seizures or serious neurologic dysfunction for 4 groups by birth setting and birth attendant (hospital physician, hospital midwife, freestanding birth center midwife, and home midwife) in the United States from 2007-2010. significantly higher risk of a 5-minute Apgar score of 0 (P < .0001) than hospital births attended by physicians or midwives. Home births (RR, 3.80) and births in freestanding birth centers attended by midwives (RR, 1.88) had a significantly higher risk of neonatal seizures or serious neurologic dysfunction (P < .0001) than hospital births attended by physicians or midwives. METHODS: Data from the United States Centers for Disease Control’s National Center for Health Statistics birth certificate data files were used to assess deliveries by physicians and midwives in and out of the hospital for the 4-year period from 2007-2010 for singleton term births ( 37 weeks’ gestation) and 2500 g. Five-minute Apgar scores of 0 and neonatal seizures or serious neurologic dysfunction were analyzed for 4 groups by birth setting and birth attendant (hospital physician, hospital midwife, freestanding birth center midwife, and home midwife). CONCLUSION: The increased risk of 5-minute Apgar score of 0 and seizures or serious neurologic dysfunction of out-of-hospital births should be disclosed by obstetric practitioners to women who express an interest in out-of-hospital birth. Physicians should address patients’ motivations for out-of-hospital delivery by continuously improving safe and compassionate care of pregnant, fetal, and neonatal patients in the hospital setting. RESULTS: Home births (relative risk [RR], 10.55) and births in freestanding birth centers (RR, 3.56) attended by midwives had a Key words: Apgar score, birth center, homebirth, hospital birth, neonatal seizures, patient safety Cite this article as: Grünebaum A, McCullough LB, Sapra KJ, et al. Apgar score of 0 at 5 minutes and neonatal seizures or serious neurologic dysfunction in relation to birth setting. Am J Obstet Gynecol 2013;209:323.e1-6. H ome births have increased in the United States in the last decade,1 allthough home births in the Netherlands, the country with the greatest experience with home births, have decreased.2 The 2011 American College of Obstetricians and Gynecologists’ Committee Opinion, “Planned Home Birth,” provides a useful review of the literature.3 The Cochrane Collaboration has published 2 reviews, one of clinical trials comparing planned hospital birth with planned home birth4 and another of trials comparing institutional and alternative birth settings.5 The safety of out-of-hospital birth remains controversial. The purpose of this study was to examine the occurrence of 5-minute Apgar scores of zero and seizures or serious neurologic dysfunction for 4 groups by birth setting and birth attendant (hospital physician, hospital From the Department of Obstetrics and Gynecology (Drs Grünebaum, Brent, and Chervenak), Weill Medical College of Cornell University, and the Department of Epidemiology (Ms Sapra), College of Physicians and Surgeons, Columbia University, New York, NY; Center for Medical Ethics and Health Policy (Dr McCullough), Baylor College of Medicine, Houston, TX; Alfred I. DuPont Hospital for Children (Dr Brent), Thomas Jefferson University, Wilmington, DE; the Division of Pediatrics and Child Health (Dr Levene), University of Leeds, Leeds, England, UK; and the Center for Mother and Child (Dr Arabin), Philipps University, Marburg, and the Clara Angela Foundation, Berlin, Germany. Received March 21, 2013; revised April 30, 2013; accepted June 17, 2013. The authors report no conflicts of interest. Presented at the workshop Research Issues in the Assessment of Birth Settings held by the Institute of Medicine, National Academy of Sciences, March 6-7, 2013. Reprints not available from the authors. 0002-9378/$36.00 ª 2013 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2013.06.025 midwife, freestanding birth center midwife, and home midwife) in the United States from 2007-2010. M ATERIALS AND M ETHODS Data were obtained from the National Center for Health Statistics (NCHS) of the US Centers for Disease Control (CDC) birth certificate data for 20072010, the most recent data available. The CDC files contain detailed information on each of the approximately 4 million births in the United States each year. Data on patient characteristics include setting and method of delivery as well as birth attendant as reported on birth certificates filed each year with the states of the United States and compiled by NCHS. These data are publicly accessible on the internet (http://205.207.175.93/ vitalstats/ReportFolders/ReportFolders. aspx), where detailed tables can be created and downloaded for further evaluation. The data that we report in this study are for the 2007-2010 period. We excluded OCTOBER 2013 American Journal of Obstetrics & Gynecology 323.e1 Research Obstetrics preterm births (<37 weeks), infants weighing under 2500 g, and multiple gestations. This study therefore includes only singleton term births (deliveries 37 weeks) and infants weighing 2500 g. Data on patients’ characteristics included parity, race and ethnicity, maternal age, and clinical factors including neonatal weight and weeks of gestation. We included patients in the 4 CDC categories that are described by birth setting and birth attendant: hospital-based physician; hospital-based midwife; freestanding birth center midwife; and homebased midwife. Missing data were excluded for each parameter before percentages were computed. Differences noted as higher or lower were statistically significant at the P < .05 level. Apgar scores6,7 are well reported on birth certificates.1,8-11 We included outcome data on 5-minute Apgar scores of 0, the clinical and prognostic utility of which is well established.1,8-11 We also included outcome data on neonatal seizures or serious neurologic dysfunction, the category used by the CDC. Since the introduction of the 2003 revised US Standard Certificate of Live Birth, outcome data such as seizures or serious neurologic dysfunction have been documented in 21 states in 2007, 27 states in 2008 and 2009, and 35 states in 2010. We calculated the risk ratios for seizures or serious neurologic dysfunction only for those states that had these data on their birth certificates, which included about 56% of all US births. Five-minute Apgar scores of 0 and data on seizure or serious neurologic dysfunction were placed into the same 4 groups by birth place and attendant. Five-minute Apgar scores of 0 and seizures or serious neurologic dysfunction are reported, as well as by parity (0, 1). Bivariable analyses were conducted to determine whether characteristics of mothers and infants differed by the 4 groups. Hospital births attended by physicians served as the reference group in this analysis. For characteristics that had multiple levels (eg, age and race), a reference group was selected (<25-yearsold for age and non-Hispanic white for race). All levels of the characteristic were www.AJOG.org individually compared with the reference group. c2 statistics were calculated for each bivariable analysis. Risk ratios and 95% confidence intervals (CIs) were calculated for each outcome in the 4 groups. Risks of the other 3 groups were individually compared with risks for hospital births attended by physicians. To account for confounding by parity, stratified analyses were conducted for parity ¼ 0 and parity >0 for 5-minute Apgar scores of zero and seizures or serious neurologic dysfunction. In addition, stratum-specific estimates were calculated for maternal age <35-years-old, maternal age 35-year-old, gestational age 37-40 weeks, and gestational age 41 weeks for 5-minute Apgar scores of zero. All statistical analyses were conducted in OpenEpi (Open Source Epidemiologic Statistics for Public Health, Atlanta, GA).12 Because nonidentifiable data from a publicly available dataset were used, our study was not considered human subjects research and did not require review by the institutional review board of Weill Medical College of Cornell University. R ESULTS From 2007 to 2010, there were a total of 16,693,978 births in the United States. Our study population consisted of 13,891,274 singleton deliveries, 37 weeks, with a birthweight 2500 g who were delivered in the hospital, a birthing center, or at home by either a physician or a midwife. In our study population, 5minute Apgar scores were available for 98.8% of all states and for neonatal seizures or serious neurologic dysfunction in 97.5% of those states that had collected presence or absence of neonatal seizures or serious neurologic dysfunction in their birth certificates. Table 1 shows patient characteristics and the distribution of the 4 groups of settings and birth attendants of our study population. There were a total of 13,891,274 births by physicians or midwives in the hospital, a freestanding birthing center, or at home between 2007 and 2010. The majority of term singleton births (91.16%; n ¼ 12,663,051) were physician hospital births; midwife hospital births constituted 8.05% of birth (n ¼ 1,118,678), and 0.49% (n ¼ 67,429) were 323.e2 American Journal of Obstetrics & Gynecology OCTOBER 2013 midwife home deliveries. Patients delivering at home attended by midwives were significantly more likely to be multiparous, non-Hispanic white, 30 years of age, delivering beyond 41 and 42 weeks, and having macrosomic infants over 4000 and 4500 g (P < .0001). Table 2 shows the outcomes and relative risks (RRs) by the 4 groups of settings and attendants for 5-minute Apgar scores of 0, by parity. The RR of a 5-minute Apgar score of 0 for midwife home deliveries was 10.55 (95% CI, 8.62e 12.93). The RR of a 5-minute Apgar score of 0 for midwife home deliveries further increased to 14.24 (95% CI, 10.16e19.96) for nulliparous patients. The RR for freestanding birth center midwife deliveries was less than home deliveries (3.56 vs 10.55) but it was increased relative to hospital deliveries by physicians or midwives. Within the hospital, midwifeattended deliveries had a lower RR (0.55; 95% CI, 0.45e0.68) compared with physicians. When we analyzed 5-minute Apgar scores of 0 for women <35 years of age, we found that the RR for midwife home deliveries was 8.76 (95% CI, 6.85e11.21) and for freestanding birth center midwife deliveries the RR was 4.28 (95% CI, 2.81e6.52). The RR for women 35 years of age for midwife home deliveries was 15.86 (95% CI, 10.97e22.92). When we analyzed 5-minute Apgar scores of 0 for women 41 weeks’ gestation we found that the RR for midwife deliveries was 6.5 (96% CI, 4.09e10.33) and 11.7 (95% CI, 9.33e14.68) for deliveries between 37-40 weeks. Table 3 shows the outcomes by the 4 groups of settings and attendants for seizures or serious neurologic dysfunction and by parity. The RR of seizures or serious neurologic dysfunction for midwife home deliveries was 3.80 (95% CI, 2.80e5.16), and the RR of seizures or serious neurologic dysfunction for midwife home deliveries further increased to 6.28 (95% CI, 4.08e9.67) for nulliparous patients. Freestanding birthing centers midwife deliveries showed an increased risk of 1.88 (95% CI, 1.11e3.17) for seizures or serious neurologic dysfunction and an increased risk of 2.77 (95% CI, 1.48e5.15) for Obstetrics www.AJOG.org Research TABLE 1 Characteristics of study population Characteristic Hospital physician Hospital midwife Freestanding birth center midwife Home midwife TOTAL N ¼ 13,891,274 12,663,051 (91.16) 1,118,578 (8.05) 42,216 (0.30) 67,429 (0.49) Para ¼ 0 n ¼ 12,615,994 n ¼ 1,115,794 n ¼ 42,000 n ¼ 60,296 Yes 5,155,779 (40.9) 44,0642 (39.5) 15,228 (36.3) 14,801 (24.5) No 7,460,215 (59.1) 675,152 (60.5) 26,772 (63.7) 45,495 ( 75.4) n ¼ 41,992 n ¼ 66,314 Non-Hispanic white 6,894,312 (54.8) 585,553 (52.7) 34,270 (81.6) 60,017 (90.45) Non-Hispanic black 1,719,347 (13.7) 145,442 (13.1) 1865 (4.4) 1314 (1.98) Hispanic 3,100,313 (24.7) 301,223 (27.1) 4759 (11.3) 3533 (5.3) 862,493 (6.9) 78,785 (7.1) 1098 (2.6) 1490 (2.2) n ¼ 12,553,246 n ¼ 1,118,578 n ¼ 42,216 n ¼ 67,429 Ethnicity Non-Hispanic other Mother’s age n ¼ 12,576,465 n ¼ 1,111,003 <25 y 4,307,508 (34.3) 449,318 (40.2) 9338 (22.1) 10,336 (15.3) 25-29 y 3,505,877 (27.9) 325,607 (29.1) 14,432 (34.2) 20,899 (31.0) 30-34 y 2,957,460 (23.6) 228,962 (20.5) 12,119 (28.7) 21,331 (31.6) 35 y 1,782,401 (14.2) 114,691 (10.3) 6327 (15.0) 14,863 (22.0) Post EDD n ¼ 12,701,519 n ¼ 1,118,936 n ¼ 42,229 n ¼ 67,504 41 wk 1,982,383 (15.61) 227,607 (20.34) 11,184 (26.48) 19,286 (28.57) 42 wk 798,882 (6.29) 85,375 (7.63) 3711 (8.79) 6449 (9.55) Macrosomia n ¼ 12,663,051 n ¼ 1,118,578 n ¼ 312,586 n ¼ 61,684 12,831 (20.80) 4000 g 1,104,459 (8.72) 98,644 (8.82) 29,899 (9.57) 4500 g 148,509 (1.17) 11,114 (0.99) 3699 (1.18) 2538 (4.11) EDD, estimated due date. Grunebaum. Apgar score of 0 at 5 minutes and neonatal seizures. Am J Obstet Gynecol 2013. seizures or serious neurologic dysfunction for nulliparous patients. Within the hospital, midwife-attended deliveries had a lower RR compared with physicians (0.74; 95% CI, 0.62e0.89). C OMMENT Principal findings There is an identifiable pattern in these data for the outcomes of singleton term births: home birth is associated with a significantly increased risk of 5-minute Apgar scores of 0 and neonatal seizures or serious neurologic dysfunction compared with hospital birth. When it comes to home births vs hospital births, home births are strongly associated with worse outcomes. The increased rate of adverse outcomes of home births exists despite the reported lower risk profile of home birth.13 The pattern for freestanding birth centers is also identifiable: this setting is associated with increased risk compared with hospital delivery, though not as high risk as home birth. When it comes to births at a freestanding birth center vs a hospital, births at a freestanding birthing center are strongly associated with worse outcomes. It is essential to note that these significantly increased risks of adverse outcomes from the setting of home and from the setting of freestanding birth centers reported here may be serious underestimations of clinical complications. A substantial number of the adverse outcomes attributed to hospital births result from transfers from home births.14 In the Birthplace in Britain study, up to 45% of nulliparous patients were transferred to the hospital.15 In the CDC dataset the outcomes for patients whose care began out of the hospital but were transferred to the hospital are counted as outcomes of care in the hospital. They are not reported as outcomes of the original out-of-hospital setting. Obviously, correction of this factor would further negatively impact the RR of all adverse outcomes for births out of the hospital. We emphasize that the increased risks of poor outcomes from the setting of home birth, regardless of attendant, are virtually impossible to solve by transport. This is because total time for transport from home to hospital cannot realistically be reduced to clinically satisfactory times to optimize outcome OCTOBER 2013 American Journal of Obstetrics & Gynecology 323.e3 Research Obstetrics www.AJOG.org TABLE 2 5-minute Apgar scores [ 0 by birth setting, birth attendant, and parity Outcome/Birth setting n/Total (per 1000) RR (95% CI) 5-minute Apgar 0 (all) Hospital MD 1,943/12,615,994 (0.16) 1.00 Hospital midwife 95/1,115,794 (0.09) 0.55 (0.45e0.68) Freestanding BC midwife 23/42,000 (0.55) 3.56 (2.36e5.36) Home midwife 98/60,296 (1.63) 10.55 (8.62e12.93) 5-minute Apgar 0 (P ¼ 0) Hospital MD 856/5,155,779 (0.17) 1.00 Hospital midwife 37/440,642 (0.84) 0.51 (0.36e0.70) Freestanding BC midwife 11/15,226 (7.22) 4.35 (2.40e7.89) Home midwife 35/14,801 (2.36) 14.25 (10.16e19.96) 5-minute Apgar 0 (P > 0) Hospital MD 1087/7,460,215 (0.15) 1.00 Hospital midwife 58/675,152 (0.09) 0.59 (0.45e0.77) Freestanding BC midwife 12/26,772 (0.45) 3.08 (1.74e5.43) Home midwife 63/45,495 (1.35) 9.5 (7.37e12.25) Hospital MD is the reference group. BC, birth center; CI, confidence interval; MD, doctor; RR, relative risk. Grunebaum. Apgar score of 0 at 5 minutes and neonatal seizures. Am J Obstet Gynecol 2013. when time is of the essence when unexpected deterioration of the condition of either the fetal patient or pregnant patient occurs. Clinical implications Our data have important implications for the informed consent process for planned out-of-hospital birth. In the ethics and law of informed consent, obstetricians have the professional responsibility to identify medically reasonable alternatives for the management of pregnancy and their benefits and risks.16 The data reported here strongly support the clinical judgment that home delivery and birth in freestanding centers are not medically reasonable, given their preventable, clinically significant absolute and RRs of adverse perinatal outcomes. Physicians therefore should not offer and should recommend against birth settings outside the hospital.17,18 We emphasize that this stance should be accompanied by effective efforts to reduce unnecessary interventions and to improve the institutional setting of hospital delivery to make it more home like,17,18 as well as continuously improve its quality and safety.19 Implications for research Initiation of clinical trials at any phase requires protection of human subjects from preventable adverse events in the study design. In our judgment, the principal findings of our study document increased, preventable harms of out-of-hospital settings that should rule out as ethically unacceptable randomized controlled clinical trials of hospital vs out-of-hospital birth settings.17,18 Findings in other studies Some studies that reported on low-risk home births showed decreased perinatal and neonatal mortality rates,20,21 although other studies reported increased mortality rates.22-24 In a comparison of midwife-attended hospital vs midwifeattended home birth, Malloy reported an increased risk of neonatal mortality and 5-minute Apgar scores <4 for the home vs the hospital setting.25 Our 323.e4 American Journal of Obstetrics & Gynecology OCTOBER 2013 analysis is more comprehensive than Malloy’s both by setting and by attendant. There is a pattern related to the incidence of 5-minute Apgar scores of 0 in our analysis: nulliparous patients have a many-fold significantly higher risk of 5-minute Apgar scores of 0 in the home setting, when compared with multiparous women. Others have observed this pattern26 and have called for discouraging women from having their first birth at home.27 We emphasize that, despite these differences, lower risk conditions such as multiparity or term births below 41 weeks do not provide acceptable protection from adverse outcomes in the home setting. Strengths and limitations The major strength of our analysis is the large sample size for both hospital and home birth over a 4-year period from the most comprehensive and reliable dataset available in the United States. Our data are also consistent with those of others who found increased neonatal morbidity and mortality25 in home births, especially in nulliparous women.26 Our study has several limitations. The quality of data reported in birth certificates can vary.11,28 Although information on setting, birth attendant, and Apgar scores are reliable in the CDC dataset, data on seizures or serious neurologic dysfunction are less so.1,8-11 Another limitation is that our data for seizures or serious neurologic dysfunction included about 60% of the US births between 2007 and 2010 for those states that have been using the 2003 US Standard Certificate of Live Birth. Because of this sample, results about neonatal seizures or serious neurologic dysfunction may not be generalizable for the whole country. Nevertheless, for the states reporting, there was a 97.5% compliance rate for indicating presence or absence of seizures or serious neurologic dysfunction. The CDC data on seizures or serious neurologic dysfunction include those of genetic, prenatal, intrapartum, and neonatal origin that might not be related to birth setting. Another limitation is that it is not possible to know from the CDC data whether a 5-minute Apgar score of 0 was Obstetrics www.AJOG.org TABLE 3 Neonatal seizures or serious neurologic dysfunction by birth setting, birth attendant, and parity Variable N/Total (per 1000) RR (95% CI) 1823/8,102,337 (0.22) 1.00 Seizures (All) Hospital MD 121/727,395 (0.17) 0.74 (0.62e0.89) Freestanding BC midwife Hospital midwife 14/33,188 (0.42) 1.88 (1.11e3.17) Home midwife 42/49,091 (0.86) 3.8 (2.80e5.16) Seizures (P ¼ 0) Hospital MD 981/3,297,301 (0.30) 1.00 Hospital midwife 77/286,920 (0.27) 0.90 (0.72e1.14) Freestanding BC midwife 10/12,155 (0.83) 2.77 (1.48e5.15) Home midwife 21/11,239 (1.87) 6.28 (4.08e9.67) Seizures (P > 0) Hospital MD Hospital midwife Freestanding BC midwife Home midwife 842/4,805,036 (0.18) 1.00 44/440,475 (0.10) 0.57 (0.42e0.77) 4/21,073 (0.19) 1.08 (0.41e2.89) 21/37,853 (0.55) 3.17 (2.05e4.88) Hospital MD is the reference group. BC, birth center; CI, confidence interval; MD, doctor; RR, relative risk. Grunebaum. Apgar score of 0 at 5 minutes and neonatal seizures. Am J Obstet Gynecol 2013. effectively a stillbirth that occurred antepartum or intrapartum. We do not believe that this limitation changes our major findings. This is because the vast majority of stillbirths delivered in the hospital are known to be antepartum and not intrapartum.29-31 On the other hand, in out-of-hospital settings, most antepartum deaths in planned home births would be transferred to the hospital. Moreover, in out-of-hospital settings, there is likely less antepartum testing and no continuous electronic intrapartum fetal monitoring, both of which may have affected adverse outcomes. Data on long-term follow-up of neonates would be optimal, but the CDC database does not include such information. An Apgar score of 0 indicates that there are no signs of life (no heartbeat, no breathing or movements). Infants with a 5-minute Apgar score of 0 have a significantly increased risk of mortality and if they survive an increased risk of significant morbidity.32,33 Survival relates directly to the effectiveness of neonatal resuscitation that is severely limited in home births. Head cooling may improve outcomes but there is still significant mortality and morbidity.34 Most importantly, the CDC does not categorize as out-of-hospital births those hospital births that resulted from transfer from out-of-hospital settings where there was an intention for out-of-hospital birth. A midwife-attended delivery at home or at a birth center, however, is an appropriate proxy for intended or planned out-of-hospital delivery. There is no way to assess from these data when intended out-of-hospital deliveries are transferred to the hospital, making an intention-to-treat analysis impossible. Conclusion The increased risk of 5-minute Apgar score of 0 and increased rates of seizures or serious neurologic dysfunction of out-of-hospital birth must be acknowledged by all obstetric practitioners and should be disclosed to all pregnant women who express an interest Research in out-of-hospital birth. In addition, physicians have the professional responsibility to recommend against planned out-of-hospital births to women who express an interest in it and not to refer their patients to randomized controlled clinical trials of hospital vs out-of-hospital birth as ethically unacceptable.17,18 Physicians also have the professional responsibility to address the root cause of patients’ motivations for out-of-hospital delivery through continuous efforts to address patient concerns about interventions,35 and to improve compassionate and safe care of pregnant, fetal, and neonatal patients in the hospital setting.17-19 REFERENCES 1. MacDorman MF, Mathews TJ, Declercq E. Home births in the United States, 1990-2009. NCHS Data Brief 2012;84:1-8. 2. Visser GHA. Obstetric care in the Netherlands: relic or example? J Obstet Gynaecol Can 2012;34:971-5. 3. American College of Obstetricians and Gynecologists. ACOG committee opinion no. 476. Committee on Obstetric Practice. Planned home birth. Obstet Gynecol 2011;117:425-8. 4. Olsen O, Clausen JA. Planned hospital birth versus planned home birth. Cochrane Database of Systematic Reviews 2012;9: CD000352. http://dx.doi.org/10.1002/14651858. CD000352.pub2. 5. Hodnett ED, Downe S, Walsh D. Alternative versus conventional institutional settings for birth. Cochrane Database of Systematic Reviews 2012;8:CD000012. http://dx.doi.org/ 10.1002/14651858.CD000012.pub4. 6. Apgar V. A proposal for a new method of evaluation of the newborn infant. Curr Res Anesth Analg 1953;32:260-7. 7. Apgar V, Holiday DA, James LS, Weisbrot IM, Berrien C. Evaluation of the newborn infant: second report. JAMA 1958;168:1985-8. 8. DiGiuseppe DL, Aron DC, Ranbom L, Harper DL, Rosenthal GE. Reliability of birth certificate data: a multi-hospital comparison to medical records information. Mat Child Health J 2002;6:169-79. 9. Zollinger TW, Przybylski MJ, Gamache RE. Reliability of Indiana birth certificate data compared to medical records. Ann Epidemiol 2006;16:1-10. 10. Northam S, Knapp TR. The reliability and validity of birth certificates. JOGNN 2006;35: 3-12. 11. Vinikoor LC, Messer LC, Laraia BA, Kaufman JS. Reliability of variables on the North Carolina birth certificate: a comparison with directly queried values from a cohort study. Paediatr Perinat Epidemiol 2010;24:102-12. OCTOBER 2013 American Journal of Obstetrics & Gynecology 323.e5 Research Obstetrics 12. Dean AG, Sullivan KM, Soe MM. OpenEpi: open source epidemiologic statistics for public health, version 2.3.1. Updated June 23, 2011. Available at: www.OpenEpi.com. Accessed March 10, 2013. 13. MacDorman MF, Mathews TJ, Declerq E. Home births in the United States, 1990-2009. NCHS Data Brief 2012;84:1-8. 14. Evers CC, Brouwers HA, Hukkelhoven CW, et al. Perinatal mortality and severe morbidity in low and high risk term pregnancies in the Netherlands: prospective cohort study. BMJ 2010;341:c 359. 15. Birthplace in England Collaborative Group. Brocklehurst P, Hardy P, Hollowell J, et al. Perinatal and maternal outcomes by planned place of birth for healthy women with low risk pregnancies: the Birthplace in England national prospective cohort study. BMJ 2011;343:d7400. 16. McCullough LB, Chervenak FA. Ethics in obstetrics and gynecology. New York: Oxford University Press; 1994. 17. Chervenak FA, McCullough LB, Arabin B. Obstetric ethics: an essential dimension of planned home birth. Obstet Gynecol 2011;117:1183-7. 18. Chervenak FA, McCullough LB, Brent RL, Levene MI, Arabin B. Planned home birth: the professional responsibility response. Am J Obstet Gynecol 2013;208:31-8. 19. Grünebaum A, Chervenak F, Skupski D. Effect of a comprehensive patient safety program on compensation payments and sentinel events. Am J Obstet Gynecol 2011;204: 97-105. www.AJOG.org 20. Janssen PA, Lee SK, Ryan EM, et al. Outcomes of planned home births versus planned hospital births after regulation of midwifery in British Columbia. CMAJ 2002;166:315-23. 21. Amelink-Verburg MP, VerlooveVanhorick SP, Hakkenberg RM, Veldhuijzen IM, Bennebroek Gravenhorst J, Buitendijk SE. Evaluation of 280,000 cases in Dutch midwifery practices: a descriptive study. BJOG 2008;115:570-8. 22. Wax JR, Lucas FL, Lamont M, Cartin A, Blackstone J. Maternal and newborn outcomes in planned home birth vs planned hospital births: a metaanalysis. Am J Obstet Gynecol 2010;203: 243.e1-8. 23. Pang JW, Heffelfinger JD, Huang GJ, Benedetti T, Weiss NS. Outcomes of planned home births in Washington State: 1989-1996. Obstet Gynecol 2002;100:253-9. 24. Bastian H, Keirse MJ, Lancaster PA. Perinatal death associated with planned home birth in Australia: population based study. BMJ 1998;317:384-8. 25. Malloy MH. Infant outcomes of certified nurse midwife attended home births: United States 2000 to 2004. J Perinatol 2010;30:622-7. 26. Cheng YW, Snowden JM, King TL, Caughey AB. Selected perinatal outcomes associated with planned home births in the United States. Am J Obstet Gynecol 2013 Jun 18. http:// dx.doi.org/10.1016/j.ajog.2013.06.022 [Epub ahead of print]. 27. Buekens P, Keirse MJ. In the literature: home birth: safe enough, but not for the first baby. Birth 2012;39:165-7. 323.e6 American Journal of Obstetrics & Gynecology OCTOBER 2013 28. Grimes DA. Epidemiologic research using administrative databases: garbage in, garbage out. Obstet Gynecol 2010;116:1018-9. 29. Getahun D, Ananth CV, Kinzler WL. Risk factors for antepartum and intrapartum stillbirth: a population-based study. Am J Obstet Gynecol 2007;196:499-507. 30. Smith GC. Life-table analysis of the risk of perinatal death at term and post term in singleton pregnancies. Am J Obstet Gynecol 2001;184:489-96. 31. Goldenberg RL, Kirby R, Culhane JF. Stillbirth: a review. J Matern Fetal Neonatal Med 2004;16:79-94. 32. Haddad B, Mercer BM, Livingston JC, Talati A, Sibai BM. Outcome after successful resuscitation of babies born with Apgar scores of 0 at both 1 and 5 minutes. Am J Obstet Gynecol 2000;182:1210-4. 33. Harrington DJ, Redman CW, Moulden M, Greenwood CE. The long-term outcome in surviving infants with Apgar zero at 10 minutes: a systematic review of the literature and hospitalbased cohort. Am J Obstet Gynecol 2007;196: 463.e1-5. 34. Sarkar S, Bhagat I, Dechert RE, Barks JD. Predicting death despite therapeutic hypothermia in infants with hypoxic-ischaemic encephalopathy. Arch Dis Child Fetal Neonatal Ed 2010;95:F423-8. 35. Glantz JC. Birth. Obstetric variation, intervention, and outcomes: doing more but accomplishing less. Birth 2012;39: 286-90. Research ajog.org OBSTETRICS Perinatal risks of planned home births in the United States Amos Grünebaum, MD; Laurence B. McCullough, PhD; Robert L. Brent, MD, PhD, DSc (Hon); Birgit Arabin, MD; Malcolm I. Levene, MD, FRCP, FRCPH; Frank A. Chervenak, MD OBJECTIVE: We analyzed the perinatal risks of midwife-attended planned home births in the United States from 2010 through 2012 and compared them with recommendations from the American College of Obstetricians and Gynecologists (ACOG) and the American Academy of Pediatrics (AAP) for planned home births. STUDY DESIGN: Data from the US Centers for Disease Control and Prevention’s National Center for Health Statistics birth certificate data files from 2010 through 2012 were utilized to analyze the frequency of certain perinatal risk factors that were associated with planned midwife-attended home births in the United States and compare them with deliveries performed in the hospital by certified nurse midwives. Home birth deliveries attended by others were excluded; only planned home births attended by midwives were included. Hospital deliveries attended by certified nurse midwives served as the reference. Perinatal risk factors were those established by ACOG and AAP. RESULTS: Midwife-attended planned home births in the United States had the following risk factors: breech presentation, 0.74% (odds ratio [OR], 3.19; 95% confidence interval [CI], 2.87e3.56); prior cesarean delivery, 4.4% (OR, 2.08; 95% CI, 2.0e2.17); twins, 0.64% (OR, 2.06; 95% CI, 1.84e2.31); and gestational age 41 weeks or longer, 28.19% (OR, 1.71; 95% CI, 1.68e1.74). All 4 perinatal risk factors were significantly higher among midwife-attended planned home births when compared with certified nurse midwiveseattended hospital births, and 3 of 4 perinatal risk factors were significantly higher in planned home births attended by noneAmerican Midwifery Certification Board (AMCB)ecertified midwives (other midwives) when compared with home births attended by certified nurse midwives. Among midwife-attended planned home births, 65.7% of midwives did not meet the ACOG and AAP recommendations for certification by the American Midwifery Certification Board. CONCLUSION: At least 30% of midwife-attended planned home births are not low risk and not within clinical criteria set by ACOG and AAP, and 65.7% of planned home births in the United States are attended by non-AMCB certified midwives, even though both AAP and ACOG state that only AMCB-certified midwives should attend home births. Key words: home birth, midwives, perinatal risks Cite this article as: Grünebaum A, McCullough LB, Brent RL, et al. Perinatal risks of planned home births in the United States. Am J Obstet Gynecol 2015;212:350.e1-6. T here has been an increase in home births in the United States over the last 10 years.1 Recent studies have shown that when compared with hospital births, planned home births by midwives are associated with an increase in adverse neonatal outcomes, such as neonatal deaths,2-4 Apgar score of 0, neonatal seizures, or serious neurological dysfunction.5 The American College of Obstetricians and Gynecologists (ACOG) and the American Academy of Pediatrics (AAP) have concluded that planned hospital births are safer than planned home births, and both professional organizations have also identified clinical criteria for selecting low-risk patients for planned home births.6,7 ACOG and AAP have also stated that midwives attending planned home births should be certified by the American Midwifery Certification Board (AMCB).6,7 From the Department of Obstetrics and Gynecology, Weill Medical College of Cornell University, New York, NY (Drs Grünebaum, Brent, and Chervenak); Center for Medical Ethics and Health Policy, Baylor College of Medicine, Houston, TX (Dr McCullough); Departments of Obstetrics and Gynecology, Thomas Jefferson University, Philadelphia, PA, and Alfred I. DuPont Hospital for Children, Wilmington, DE (Dr Brent); Center for Mother and Child, Philipps University, Marburg, and Clara Angela Foundation, Berlin, Germany (Dr Arabin); and Division of Pediatrics and Child Health, University of Leeds, Leeds, England, UK (Dr Levene). Received Aug. 29, 2014; revised Sept. 9, 2014; accepted Oct. 13, 2014. The authors report no conflict of interest. Corresponding author: Amos Grünebaum, MD. amg2002@med.cornell.edu 0002-9378/$36.00 ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2014.10.021 350.e1 American Journal of Obstetrics & Gynecology MARCH 2015 The purpose of this study was to evaluate the frequency of certain perinatal risk factors that were associated with planned midwife-attended home births in the United States from 2010 through 2012 and to compare them with clinical criteria for planned home births established by the ACOG and AAP. M ATERIALS AND M ETHODS We utilized data from the National Center for Health Statistics of the US Centers for Disease Control and Prevention (CDC) birth certificate data for 2010e2012, the most recent data available to analyze the 4 ACOG/AAP clinical criteria for planned home births. The CDC files contain detailed information on each of the approximately 4 million births in the United States each year. Data on patient characteristics including birth setting, method of delivery, birth attendant, gestational age, infant birthweight, maternal age, history of ajog.org prior cesarean delivery, and parity are reported on birth certificates filed each year with each of the states in the United States and compiled by National Center for Health Statistics. These data are publicly accessible on the Internet (http://205.207.175.93/vitalstats/Report Folders/ReportFolders.aspx), where detailed tables can be created and downloaded for further evaluation. According to CDC data, “almost all the home births attended by certified nurse-midwives ⁄certified midwives (98%) or “other” midwives (99%) were planned,”8,9 and therefore, we defined planned home births as births attended at home by midwives. We excluded from planned home births those performed at home by others (eg, family members, emergency medical service, or police, taxi drivers as well as unattended births). Planned US midwife home deliveries for the years 2010e2012, the most recent years available, were analyzed for ACOG and AAP perinatal risk factors that should be excluded from home births7: vaginal breech deliveries, prior cesarean delivery, twin gestations, and postdate pregnancies (gestational age 41 weeks or longer). Hospital births attended by certified nurse midwives served as a reference. Home birtheplanned midwife-attended deliveries were compared with hospital-certified nurse midwives (CNM)eattended deliveries. The CDC database separates midwives into CNM and other midwives. The AMCB certifies 2 kinds of midwives: CNMs and certified midwives (CMs), both of whom have graduated from a midwifery education program accredited by the American Commission for Midwifery Education. The total number of AMCB-certified midwives (CNMs plus CMs) includes only a small percentage of CMs because CMs are permitted to practice in only 5 states. Therefore, the CDC designation of CNMs captures nearly all of AMCBcertified midwives. In addition to CNMs, the CDC also has a designation of other midwives, which includes certified professional midwives, who are not eligible for certification by the AMCB and who have no requirement of a Bachelor’s degree or Obstetrics graduate training. In addition, the CDC designation of other midwives may include lay midwives and others without any graduate midwifery training. We performed a subanalysis and compared the frequency with which certain perinatal risk factors were associated with home births attended by CNMs with those attended by other midwives (ie, midwives not eligible to get certified by the AMCB). Data were abstracted from the US birth certificate data. Because nonidentifiable data from a publicly available data set were used, this study was not considered human subject research and did not require review by the Institutional Review Board of Weill Medical College of Cornell University. Statistical analyses were conducted for comparisons between planned midwife-attended home births and CNM-attended deliveries in the hospital. Odds ratios and 95% confidence intervals were calculated for each of the 3 provider groups (CNM-attended home birth, other midwife-attended home birth, and CNM-attended hospital birth) and 4 of the risk groups. All statistical analyses were conducted in OpenEpi.10 R ESULTS Between 2010 and 2012, there were a total of 11,905,817 deliveries in the United States, of which 736,070 were attended by CNMs in the hospital. There were 85,318 home births (0.71% of all births in the United States) and after exclusion of 29,178 home birth deliveries performed by others, we included 56,140 deliveries that were attended by midwives at home and are considered planned midwife-attended home births. CNMs attended 19,263 (34.3%) of these home births, whereas other midwives attended 36,877 (65.7%) of planned home births. Table 1 shows the comparisons of perinatal risk factors between deliveries attended in the hospital by CNMs and planned midwife-attended home births by CNMs and other midwives. Of the midwife-attended planned home births, approximately 3 in 10 were at a gestation of 41 weeks or longer, 1 in 156 were Research births with twins, approximately 1 in 23 were vaginal births after cesarean deliveries, and 1 in 135 home births were births with breech presentation. Planned home births attended by CNMs and other midwives had a significantly higher frequency of certain perinatal risks when compared with CNM-attended hospital births. Planned home births attended by noneAMCBcertified other midwives had a significantly higher frequency of perinatal risks for breech presentation, prior cesarean deliveries, and twins, when compared with planned home births attended by CNMs. C OMMENT The AAP and ACOG previously published policy statements on planned home birth with recommendations when to consider planned home birth, and they listed the use of strict selection criteria for planned home births (Tables 2 and 3).6,7 According to the ACOG, selection criteria for home births include singletons, cephalic pregnancies between 37 and 41 weeks, no prior cesarean deliveries, and certified midwives or physicians as birth attendants. This study shows that 1 in 156 of midwife-attended planned home births (0.64%) were twin pregnancies, even though the ACOG considers twins a contraindication for home births because there is no adequate fetal monitoring, no experienced team, and no ultrasound available in home births.6,7 Studies on the safety of home births from Canada, England, and The Netherlands excluded twins as candidates for home birth because of increased risks.11-14 Even within hospitals, delivery of the second twin, especially when not engaged or nonvertex, requires an experienced obstetrician to prevent perinatal morbidity or even mortality.15 We note that our data indicate that in some hospitals there were apparently CNM deliveries of twin and breech-presentation pregnancies. The data in this study show that 1 in 135 of planned home births attended by midwives (0.74%) were vaginal breech deliveries. Breech vaginal birth is associated with significantly increased risks. MARCH 2015 American Journal of Obstetrics & Gynecology 350.e2 ajog.org 1.08 (1.01e1.14) 1.03 (0.99e1.06) 1.44 (1.39e1.5) 1.73 (1.69e1.77) 1.34 (1.28e1.41) 1.67 (1.62e1.73) 1.41 (1.37e1.45) 1.71 (1.68e1.74) 9.73 (3567) 9.08 (1744) 6.91 (50,848) Postdates 42 wks 27.7 (5320) CI, confidence interval; CNM, certified nurse midwife; MW, midwife; OR, odds ratio. 18.59 (136,729) Postdates 41 wks 350.e3 American Journal of Obstetrics & Gynecology MARCH 2015 Grünebaum. Perinatal risks of planned US home births. Am J Obstet Gynecol 2015. 9.5 (5311) 28.19 (15,755) 28.45 (10,435) 1.33 (1.05e1.67) 2.25 (1.98e2.57) 1.7 (1.39e2.08) 2.06 (1.84e2.31) 0.64 (357) 0.52 (101) 0.31 (2276) Twins 0.69 (256) 1.15 (1.06e1.26) 2.25 (2.14e2.37) 1.93 (1.8e2.08) 2.08 (2.0e2.17) 4.4 (2463) Prior cesarean delivery 3.99 (767) 2.11 (15,455) Vaginal breech 4.6 (1696) 3.49 (3.08e3.94) 2.64 (2.19e3.18) 3.19 (2.87e3.56) 0.74 (416) 0.23 (1716) Risk factor 0.61 (118) CNM-attended (n [ 736,070), % (n) 0.81 (298) Other MWattended home births vs CNMattended hospital births, OR (95% CI) CNM-attended home births vs CNM-attended hospital births, OR (95% CI) MW-attended home births vs CNM-attended hospital births, OR (95% CI) All MWattended (n [ 56,140), % (n) CNMattended (n [ 19,263), % (n) Other MWattended (n [ 36,877), % (n) Home births Hospital births Perinatal risk factors: CNM-attended hospital births vs midwife-attended home births TABLE 1 1.32 (1.07e1.64) Obstetrics Other MWattended home births vs CNMattended home births, OR (95% CI) Research Since the publication of the Term Breech Trial, clinical practices changed around the world, increasing cesarean deliveries for breech births.16 The ACOG recommends that planned vaginal breech births should be done only under hospital-specific protocol guidelines.17 Azria et al18 recommended that a trial for vaginal births in breech presentations should be attempted only with continuous electronic fetal heart rate monitoring and the presence of ultrasound during labor and delivery. Neither electronic fetal heart rate monitoring nor ultrasound is available in home births. Janssen et al11 from Canada and the Home Birth in England Study12 excluded breech presentations from their home birth eligibility requirements. Therefore, it is not surprising that the Midwives Alliance of North America study of planned home births reported an intrapartum death rate of 13.51 per 1000 and a 9.16 per 1000 neonatal mortality rate in breech presentations.19 When compared with the neonatal death rates from hospital deliveries,2-4 these adverse neonatal outcomes are significantly increased. The ACOG and AAP criteria for home births specifically exclude pregnancies 41 weeks or longer from their home birth eligibility.6,7 In this study, 28.19% of home births were 41 weeks or longer. Postterm pregnancies are associated with multiple, well-known complications, such as labor dystocia, increased perinatal mortality rate, low umbilical artery pH levels at delivery, low 5 minute Apgar scores, postmaturity syndrome, fetal distress, cephalo-pelvic disproportion, postpartum hemorrhage, and an increased risk of neonatal death within the first year of life.20,21 A trial of labor after prior cesarean delivery (TOLAC) is associated with a greater perinatal risk than is elective repeat cesarean delivery without labor. TOLACs have an overall small but significantly increased risk of uterine rupture with often catastrophic consequences to mother and/or fetus.22,23 This study showed that nearly 1 in 23 midwife-attended home births (n ¼ 2463, 4.4%) had a home vaginal birth after prior cesarean delivery (VBAC) in ajog.org Obstetrics Research TABLE 2 Planned home birth: recommendations when considering planned home birth Candidates for home delivery Absence of preexisting maternal disease Absence of significant disease occurring during the pregnancy A singleton fetus estimated to be appropriate for gestational age A cephalic presentation A gestation of 37 to less than 41 completed weeks of pregnancy Labor that is spontaneous or induced as an outpatient A mother who has not been referred from another hospital Systems needed to support planned home birth The availability of a certified nurse midwife, certified midwife, or physician practicing within an integrated and regulated health system Attendance by at least 1 appropriately trained individual (see text) whose primary responsibility is the care of the newborn infant Ready access to consultation Assurance of safe and timely transport to a nearby hospital with a preexisting arrangement for such transfers Adapted from American Academy of Pediatrics.6 Grünebaum. Perinatal risks of planned US home births. Am J Obstet Gynecol 2015. spite of the fact that ACOG considers prior cesarean section an absolute contraindication to planned home birth.7 TOLACs at home births are even more worrisome, considering the increase in VBACs in home births.24 The recent Midwives Alliance of North America study showed a very high 2.85 per 1000 intrapartum fetal death rate with VBACs.19 The ACOG and AAP recommend that only midwives certified by the AMCB should attend home births.6,7 Nevertheless, this study shows that approximately 2 of 3 planned home births were performed by non-AMCB-certified midwives. Professional organizations like the ACOG and AAP should respond to these findings by continuing to support collaborative practices in the hospital between physicians and AMCBcertified midwives and strive for a hospital birth that resembles more closely a home birth environment.25 Selection of patients for home births by countries with midwife organizations, such as the Royal Dutch Organisation of Midwives, follows collaborative guidelines with strict protocols for selecting patients for home births.26 The American College of Nurse Midwives has not established midwife-generated guidelines of patient selection for home births in the United States saying that “. guidelines would impact [midwives’] autonomy” and “guidelines might not support midwives if they choose to attend the home birth of a woman with a breech presentation or a twin gestation or a woman who desires a trial of labor after a previous cesarean.”27 It is possible that some pregnant women with risk factors may insist on home births despite the increased risks of adverse outcomes. In circumstances in which home births are contraindicated because of risk factors, physicians and midwives have the professional responsibility to strongly recommend for hospital birth, to recommend against home births, and to refuse the woman’s request to attend their home birth. This is because patients’ requests by themselves do not determine professional responsibility.28 Kennare et al29 and Bastian et al30 showed that the increase in neonatal TABLE 3 ACOG Statement on planned home birth (ACOG 2011) Recent cohort studies reporting lower perinatal mortality rates with planned home birth describe the use of strict selection criteria for appropriate candidates. These criteria include the absence of any preexisting maternal disease, the absence of significant disease arising during the pregnancy, a singleton fetus, a cephalic presentation, gestational age greater than 36 weeks and less than 41 completed weeks of pregnancy, labor that is spontaneous or induced as an outpatient, and that the patient has not been transferred from another referring hospital. Trial of labor after cesarean should be undertaken only in facilities with staff immediately available to provide emergency care. The American College of Obstetricians and Gynecologists’ Committee on Obstetric Practice considers a prior cesarean delivery to be an absolute contraindication to planned home birth. Ready access to consultation and assurance of safe and timely transport to nearby hospitals are critical to reducing perinatal mortality rates and achieving favorable home birth outcomes. Availability of a certified nurse midwife, certified midwife, or physician practicing within an integrated and regulated health system. For quality and safety reasons, the American College of Obstetricians and Gynecologists does not support the provision of care by lay midwives who are not certified by the American Midwifery Certification Board. Adapted from ACOG Committee.7 Grünebaum. Perinatal risks of planned US home births. Am J Obstet Gynecol 2015. MARCH 2015 American Journal of Obstetrics & Gynecology 350.e4 Research ajog.org Obstetrics mortality in planned home births was associated with a poor selection of candidates for home births. Our findings of increased perinatal risks among US midwife-attended planned home births may partially explain reports that show preventable increased adverse outcomes such as increased neonatal mortality rates, low Apgar scores, neonatal seizures, and serious neurological dysfunction among US midwifeattended planned home births.2-5 The strength of this study is that the CDC data are nationally comprehensive. No comparable database exists. A limitation of the results is that the actual number of patients with increased perinatal risks in home births is possibly higher than reported here because patients transferred prior to delivery from a planned home birth to the hospital are counted in the CDC birthing data as hospital births and not home births. Other limitations in this study include concerns that have been expressed about the quality of certain data collected in birth certificates, especially those that address maternal health behaviors or certain medical and obstetric conditions (eg, anemia, gestational diabetes, pregnancy-induced hypertension, concurrent illnesses, congenital anomalies, and comorbidities).31-34 These data elements were not used in our study. Our study used data elements found to be a good source of reliable information in birth certificates and that were validated such as place of births, gestational weeks, presentation, history of prior cesarean delivery, and multiple births.34,35 Conclusion This study demonstrates that many midwife-attended planned home births in the United States do not have low perinatal risks but include readily identifiable prenatal risks such as breech presentation, twins, patients with prior cesarean deliveries, and postdate pregnancies. These risks as well as other perinatal risks are known to be associated with increased adverse birth outcomes and are therefore listed by the ACOG and AAP as contraindications for planned home births. In addition, about two-thirds of planned midwife-attended home births in the United States are attended by noneAMCB-certified midwives. Our study also shows that planned home births attended by midwives not certified by the AMCB have a higher frequency of perinatal risks than planned home births attended by AMCB-certified nurse midwives. REFERENCES 1. MacDorman MF, Mathews TJ, Declercq E. Trends in out-of-hospital births in the United States, 1990e2012. NCHS Data Brief no. 144, March, 2014. Hyattsville MD: National Center for Health Statistics. 2. Grünebaum A, McCullough LB, Sapra KJ, et al. Early and total neonatal mortality in relation to birth setting in the United States, 2006e2009. Am J Obstet Gynecol 2014;211:390.e1-7. 3. Cheng YW, Snowden JM, King TL, Caughey AB. Selected perinatal outcomes associated with planned home births in the United States. Am J Obstet Gynecol 2013;209: 325.e1-8. 4. Wax JR, Lucas FL, Lamont M, Cartin A, Blackstone J. Maternal and newborn outcomes in planned home birth vs planned hospital births: a metaanalysis. Am J Obstet Gynecol 2010;203: 243.e1-8. 5. Grünebaum A, McCullough LB, Sapra KJ, et al. Apgar score of zero at five minutes and neonatal seizures or serious neurologic dysfunction in relation to birth setting. Am J Obstet Gynecol 2013;209:323.e1-6. 6. American Academy of Pediatrics. Policy statement on planned home birth. Pediatrics 2013;131:1016-20. 7. American College of Obstetricians and Gynecologists. Planned home birth. Committee Opinion no. 476. ACOG Committee on Obstetric Practice. Obstet Gynecol 2011;117(2 Pt 1): 425-8. 8. National Center for Health Statistics. Vital statistics data available online: birth data files. Available at: http://www.cdc.gov/nchs/data_ access/VitalStatsOnline.htm. Accessed Aug. 28, 2014. 9. MacDorman MF, Declerq E, Mathews TJ. United States Home Births Increase 20 Percent from 2004 to 2008. Birth 2011;38:185-90. 10. Dean AG, Sullivan KM, Soe MM. OpenEpi: open source epidemiologic statistics for public health, version 2.3.1. Updated June 23, 2011. Available at: www.OpenEpi.com. Accessed June 1, 2014. 11. Janssen PA, Saxell L, Page LA, Klein MC, Liston RM, Lee SK. Outcomes of planned home birth with registered midwife versus planned hospital birth with midwife or physician. CMAJ 2009;181:377-83. 12. Brocklehurst P, Hardy P, Hollowell J, et al. Perinatal and maternal outcomes by planned place of birth for healthy women with low risk pregnancies: the Birthplace in England national 350.e5 American Journal of Obstetrics & Gynecology MARCH 2015 prospective cohort study. Birthplace in England Collaborative Group. BMJ 2011;343: d7400. 13. Hutton EK, Reitsma AH, Kaufman K. Outcomes associated with planned home and planned hospital births in low-risk women attended by midwives in Ontario, Canada, 2003e2006: a retrospective cohort study. Birth 2009;36:180-9. 14. Amelink-Verburg MP, Verloove-Vanhorick SP, Hakkenberg RM, Veldhuijzen IM, Bennebroek Gravenhorst J, Buitendijk SE. Evaluation of 280,000 cases in Dutch midwifery practices: a descriptive study. BJOG 2008;115:570-8. 15. Arabin B, Kyvernitakis I. Vaginal delivery of the second nonvertex twin: avoiding a poor outcome when the presenting part is not engaged. Obstet Gynecol 2011;118:950-4. 16. Hannah ME, Hannah WJ, Hewson SA, Hodnett ED, Saigal S, Willan AR. Planned cesarean section versus planned vaginal birth for breech presentation at term: a randomized multicenter trial. Term Breech Trial Collaborative Group. Lancet 2000;356: 1375-83. 17. American College of Obstetricians and Gynecologists. Mode of term singleton breech delivery. Committee Opinion no. 340 (ACOG Committee on Obstetric Practice). Obstet Gynecol 2006;108:235-7. 18. Azria E, Le Meaux JP, Khoshnood B, et al. Factors associated with adverse perinatal outcomes for term breech fetuses with planned vaginal delivery. Am J Obstet Gynecol 2012;207:285.e1-9. 19. Cheyney M, Bovbjerg M, Everson C, Gordon W, Hannibal D, Vedam S. Outcomes of care for 16,924 planned home births in the United States: the Midwives Alliance of North America Statistics Project, 2004 to 2009. J Midwifery Womens Health 2014;59:17-27. 20. Caughey AB, Stotland NE, Washington AE, Escobar GJ. Who is at risk for prolonged and postterm pregnancy? Am J Obstet Gynecol 2009;200:683.e1-5. 21. Caughey AB, Stotland NE, Washington AE, Escobar GJ. Maternal and obstetric complications of pregnancy are associated with increasing gestational age at term. Am J Obstet Gynecol 2007;196:155.e1-6. 22. Landon MB, Hauth JC, Leveno KJ, et al. Maternal and perinatal outcomes associated with a trial of labor after prior cesarean delivery. N Engl J Med 2004;351:2581-9. 23. Chauhan SP, Martin JN Jr, Henrichs CE, Morrison JC, Magann EF. Maternal and perinatal complications with uterine rupture in 142,075 patients who attempted vaginal birth after cesarean delivery: a review of the literature. Am J Obstet Gynecol 2003;189:408-17. 24. MacDorman MF, Declercq E, Mathews TJ, Stotland N. Trends and characteristics of home vaginal birth after cesarean delivery in the United States and selected States. Obstet Gynecol 2012;119:737-44. ajog.org 25. Chervenak FA, McCullough LB, Brent RL, Levene MI, Arabin B. Planned home birth: the professional responsibility response. Am J Obstet Gynecol 2013;208:31-8. 26. Commissie Verloskunde van het CVZ. Verloskundig vademecum, 2003. Diemen (The Netherlands): College voor Zorgverzekeringen; 2003. 27. Cook E, Avery M, Frisvold M. Formulating evidence-based guidelines for certified nursemidwives and certified midwives attending home births. J Midwifery Womens Health 2014;59:153-9. 28. Chervenak FA, McCullough LB. The professional responsibility model of obstetrical Obstetrics ethics: avoiding the perils of clashing rights. Am J Obstet Gynecol 2011;205:315.e1-5. 29. Kennare RM, Keirse MJ, Tucker GR, Chan AC. Planned home and hospital births in South Australia, 1991e2006: differences in outcomes. Med J Aust 2010;192:76-80. 30. Bastian H, Keirse MJ, Lancaster PA. Perinatal death associated with planned home birth in Australia: population based study. BMJ 1998;317:384-8. 31. Northam S, Polanovich S, Restrepo E. Birth certificate methods in five hospitals. Public Health Nursing 2003;20:318-27. 32. Zollinger TW, Przybylski MJ, Gamache RE. Reliability of Indiana birth certificate data Research compared to medical records. Ann Epidemiol 2006;16:1-10. 33. Northam S, Knapp TR. The reliability and validity of birth certificates. J Obstet Gynecol Neonatal Nurs 2006;35:3-12. 34. DiGiuseppe DL, Aron DC, Ranbom L, Harper DL, Rosenthal GE. Reliability of birth certificate data: a multi-hospital comparison to medical records information. Matern Child Health J 2002;6:169-79. 35. Vinikoor LC, Messer LC, Laraia BA, Kaufman JS. Reliability of variables on the North Carolina birth certificate: a comparison with directly queried values from a cohort study. Paediatr Perinat Epidemiol 2010;24:102-12. MARCH 2015 American Journal of Obstetrics & Gynecology 350.e6