I: Meta-Analyses and Systematic Reviews

Each paper is linked to either its original abstract on PubMed or the full-text article if available; click on the title to be taken to the abstract/ article. 

A) Olsen, O., Clausen, J. A. (2012). Planned hospital birth versus planned home birth. Cochrane Database of Systematic Reviews, doi: 10.1002/14651858.CD000352.pub2.

An updated systematic review of randomized controlled trials (RCTs) comparing planned home births to planned hospital births among women with uncomplicated pregnancies. The selection criteria were rigorous; only one trial met the inclusion criteria (n=11). The authors report a continued dearth of evidence from RCTs about the safety of home compared to hospital birth. Authors also conclude that evidence from increasingly well-designed observational studies suggests that low-risk women who plan a home birth experience significantly fewer interventions and complications than low-risk women who deliver in hospital. They provide a detailed discourse analysis of differing approaches to risk assessment, including the ethical application of clinically meaningful evidence, and the interaction of model of care with access to choice of birth place. They recommend that all countries facilitate evidence-based integration of home birth services into the health care system and inform all low-risk women of the option of planned home birth.

B) Leslie, M.S., Romano, A. (2007). Appendix: Birth can safely take place at home and in birthing centers. J Perinat Educ, 16(Suppl 1):81S-88S, doi: 10.1624/105812407X173236.  (full text)

A systematic review of home birth and birth center safety studies. The authors followed standard systematic review methods, including reporting levels of evidence, disclosure of inclusion and exclusion criteria and search strategies (detailed in a Methods article by Goer in same journal issue). Drawing on data from numerous studies, the authors compare incidence of interventions and perinatal outcomes between hospital births and home births and between hospital births and birth center births. The evidence for each outcome is graded for quality, quantity and consistency. This review reported that out-of-hospital births had similar perinatal outcomes to hospital births and fewer interventions.

C) Olsen O. (1997). Meta-analysis of the safety of home birth. Birth, 24(1):4-13.

Meta-analysis of observational, comparative, original studies that met criteria for rigorous methodology and investigated differences in perinatal mortality and morbidity between planned home births and planned hospital births. Multivariate statistical analysis controlled for obstetrical background and perinatal factors. Analysis revealed no statistical difference in mortality between planned home and planned hospital birth and the confidence interval did not allow for extreme excess risks in any of the groups (OR=0.87, 95% CI=0.54-1.41). There were significantly fewer medical interventions, fewer severe lacerations, fewer operative births, and fewer low Apgar scores in the home birth groups.

II: Randomized Controlled Trials

A) Hendrix, M., Van Horck, M., Moreta, D., Nieman, F., Nieuwenhuijze, M., Severens, J., Nijhuis, J. (2009). Why women do not accept randomisation for place of birth: feasibility of a RCT in The Netherlands. BJOG 116(4):537-544, doi: 10.1111/j.1471-0528.2008.02103.x.

Based on Dowswell’s findings the authors designed an RCT to compare home and home-like hospital births in the Netherlands for the following outcomes: interventions, satisfaction, referral to obstetricians, and costs. After 6 months, only one woman had enrolled in the study, therefore the trial was discontinued for lack of feasibility. The research team then re-designed their study to investigate the reasons women declined to participate in the RCT. The four main reasons that women indicated were: 1) they had already decided where to give birth prior to learning about the study, 2) they wished to choose their own place of birth 3) they wished to avoid delivering in the ‘wrong’ place for their first child, and 4) they were concerned about receiving an undesired treatment.

B) Dowswell, T., Thornton, J. G., Hewison, J., Lilford, R. J., Raisler, J., Macfarlane, A., Young, G., Newburn, M., Dodds, R. & Settatree, R. S. (1996). Should there be a trial of home versus hospital delivery in the United Kingdom? Measuring outcomes other than safety is feasible. BMJ 312(7033): 753-757.  (full text)

The authors of this small study (n=11) suggested that conducting a trial to assess birth outcomes by birth place (home versus hospital) would be feasible. Eleven subjects were recruited from a pool of 71 women who met the eligibility criteria for a home birth. This ratio suggested that a larger scale trial may be possible. The following outcomes were measured, following an intention to treat analysis: mode of delivery, obstetrical interventions, complications, and infant feeding (breastfeeding versus bottle feeding). However, the authors note that mortality is not an appropriate outcome variable to assess the safety of home birth with a randomized controlled trial because of the extremely large number of subjects required to compare such rare outcomes.

III: Cohort and Population-Based Observational Studies:

North America

A) Janssen, P. A., Saxell, L., Page, L. A., Klein, M. C., Liston, R. M., & Lee, S. K. (2009). Outcomes of planned home births with registered midwife versus planned hospital birth with midwife or physician. CMAJ, 181(6-7):377-83, doi: 10.1503/cmaj.081869.  (full text)

Prospective, five-year long cohort study comparing outcomes among midwife-attended planned home births (n=2802), midwife-attended planned hospital births (n=5984), and physician-attended hospital births (n=5985). Women in all three groups of the study met eligibility criteria for home birth, and thus had comparable maternal and fetal risk profiles. Women in the home birth group who needed intrapartum transfer to the hospital were retained in their original cohort. This study reported similarly low rates of perinatal death in all three cohorts, and similar or reduced rates of adverse outcomes in the planned home birth group. Women in the planned home birth group had significantly fewer intrapartum interventions, including narcotic or epidural analgesia, augmentation or induction of labour, and assisted vaginal or caesarean delivery. In addition, women in the home birth group were less likely to suffer from postpartum hemorrhage, pyrexia, and 3rd or 4th degree tears. Babies of women planning a home birth were less likely to have Apgar scores of < 5 at one minute and the babies were less likely to need drugs for resuscitation. These differences were associated with planned place of birth and persisted regardless of actual place of birth.

B) Hutton, E. K., Reitsma, A. H., & Kaufman, K. (2009). Outcomes associated with planned home and planned hospital births in low-risk women attended by midwives in Ontario, Canada, 2003-2006: A retrospective cohort study.Birth, 36(3):180-89, doi: 10.1111/j.1523-536X.2009.00322.x.

Hutton et al. used the Ontario Ministry of Health Midwifery Program (OMP) database to compare outcomes of all women planning home births from 2003-2006 (n=6692) with a matched sample of women planning a hospital birth (n=6692.) Women with contraindications for home birth were excluded from the hospital sample. The primary outcome was a composite measures of perinatal and neonatal mortality or serious morbidity, i.e. the presence of one or more of the following: death (stillbirth or neonatal death 0–27 days, excluding lethal anomalies and fetal demise before the onset of labor); Apgar score of less than 4 at 5 minutes of age; neonatal resuscitation requiring both positive pressure ventilations and cardiac compressions; admission to a neonatal or pediatric intensive care unit with a length of stay greater than 4 days; or birthweight less than 2,500 g. The home birth group had lower rates of caesarean section (RR 0.64), and neonatal morbidity/mortality (RR 0.84) compared to low risk women who planned a hospital birth. Results suggest that Ontario midwives provide adequate screening and safe care for women planning home births.

C) Johnson, K. C., Daviss, B. A. (2005). Outcomes of planned home birth with certified professional midwives: large prospective study in North America. BMJ, 330:1416.  (full text)

A prospective study of 5418 planned home births in a single year of mandatory data collection for all Certified Professional Midwives (CPMs) in 2000. The authors describe the design as a cohort study; however, the comparison group for rates of intervention was a composite of low risk term hospital births as reported by the National Center for Health Statistics in 2000, and intrapartum and neonatal death rates were compared with those in other North American studies of at least 500 births that were either planned out of hospital or low risk hospital births. In their sample of planned home births attended by CPMs, the transfer rate was 12%, the caesarean section rate was 3.7%, the neonatal mortality rate was 1.7/1000, and the intervention rates were lower among women who planned a home birth than low risk women who delivered at hospital in the US.

D) Janssen, P. A., Lee, S. K., Ryan, E. M., Etches, D. J., Farquharson, D. F., Peacock, D., & Klein, M. C. (2002). Outcomes of planned home births versus planned hospital births after regulation of midwifery in British Columbia.CMAJ, 166(3):315-23.  (full text)

This study compared outcomes of 862 planned home births attended by midwives with hospital births attended by either midwives (n=571) or physicians (n=743). Women in the home birth group were matched with women in the physician- and midwife-attended hospital groups who met eligibility criteria for home birth. Women were matched according to age, partner status, parity, and hospital where midwives had privileges. Transfers from home to hospital were tracked, and subjects were retained in their original study groups for analysis. The study reports reasons for transfer, methods of transfer, and time spent in transfer. To assess similarity of groups, investigators also collected data on the process of midwifery care, on prenatal and obstetric history, and rates and indications for consultation or referral. Women in the home birth group were less likely to have epidural analgesia, experience induction or augmentation of labour compared to women in the physician attended group. Women in both midwife-attended groups had similar rates of obstetric procedures. There were no significant differences between home and hospital groups for the following outcomes: perinatal mortality, 5-minute APGAR scores, meconium aspiration syndrome, and need for specialized newborn care.

E) Schlenzka PF. (1999). Safety of alternative approaches to childbirth [Unpublished Dissertation]. Palo Alto, CA: Department of Sociology, Stanford University.  (full text)

In order to account for errors associated with relying solely on birth certificate data, Schlenzka merged birth certificate and hospital discharge data for California for 1989 and 1990, and by applying a comprehensive risk profile to cases, isolated a cohort of nearly 816,000 low risk births. Outcomes are reported according to planned and actual birth setting. Perinatal mortality was compared with two statistical approaches: indirect standardization using only birth weight, sex, race, age, education, and insurance as risk adjusters, and logistic regression controlling for all risk factors available in the database. No differences in perinatal mortality were found across birth sites, with lower rates of obstetric interventions in out of hospital groups.

IV: Cohort & Population-Based Observational Studies: International

A) Birthplace in England Collaborative Group. (2011). 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 343:d7400.  (full text)

A prospective cohort study in England from April 2008-April 2010 compared perinatal and maternal outcomes and interventions by planned place of birth at the onset of care during labour (planned home birth, freestanding midwifery birth centers, alongside midwifery units and obstetric units). The study included 64,538 low-risk women with a singleton pregnancy at term. The primary study outcome was a Composite Index combining intrapartum stillbirth, early neonatal death, neonatal encephalopathy, meconium aspiration syndrome, and birth related injuries including brachial plexus injury, fractured humerus or clavicle. Stillbirths before onset of labour were excluded. The researchers found that the incidence of the composite outcome measure was low for the entire sample (4.3/1000 births). In the overall sample, there were no statistically significant differences in the odds of the primary outcome in home, free-standing birth centers or alongside midwifery units when compared with planned birth in obstetric units. However, when the sample was split into nulliparous and multiparous women, the adverse outcome measures during planned home birth were higher than for hospital birth for nulliparous, but not for multiparous women. There was no evidence of a difference in adverse outcomes for freestanding or alongside midwifery units compared to obstetric units. Of women who started labour in obstetrical units, 20% had at least one complicating condition compared with less than 7% in other settings. For low-risk women birthing in an obstetric unit, the odds of receiving augmentation, epidural, spinal analgesia, general anesthesia, vacuum or forceps delivery, caesarean section, episiotomy, and active management of third stage were higher than all other settings. The study concludes that for healthy women with low risk pregnancies, the incidence of adverse perinatal outcomes is low in all settings and therefore the results support offering healthy low-risk nulliparous and multiparous women a choice of birth setting. Given the rarity of events for any of the included perinatal outcomes, and as some of them typically appear as co-morbidities, a composite index might inflate some differences in outcomes as attributable to place of birth. It is unclear how some of the items selected for inclusion in the composite index relate specifically to place of birth causality rather than skill of provider.

B) van der Kooy, J., Poeran, J., de Graff, J. P., Birnie, E., Denktass, S., Steegers, E. A. P., & Bonsel, G. J. (2011). Planned home compared with planned hospital births in the Netherlands: intrapartum and early neonatal death in low-risk pregnancies. Obstet Gynecol 118(5):1037-46.

In this retrospective cohort study, records of 679,952 low risk women from the Netherlands Perinatal Registry (2000-2007) were analyzed to compare intrapartum and early neonatal mortality rates (0-7 days after birth) of planned home versus planned hospital births attended by midwives. Outcomes for a third group of women, for which the planned place of birth was unknown, were also reported. The hospital cohort was used as the comparison group in all analyses. The authors used two methods for analyzing data: a ‘per protocol analysis’, or ‘perfect guideline approach’, which examined outcomes from only those low risk women who were eligible for planned home birth according to Dutch guidelines (n= 602,331) and a ‘natural prospective approach’, which looked at outcomes for all women who planned a home birth under the care of midwives (n=679,952) The per protocol analysis excluded midwifery clients with one or more of the following conditions: intrauterine death, prolonged rupture of membranes, gestational ages < 37 weeks and > 41 weeks. Results revealed a significantly decreased risk of intrapartum and early neonatal mortality in the home birth cohort, using the natural prospective approach (RR = 0.80; 95% CI: 0.71-0.91). When the authors calculated RRs using the perfect guideline approach, and adjusted ORs using either approach, they found no increased risk/odds of intrapartum and early neonatal death in the home versus the hospital setting. These findings align with those reported by De Jonge, et al (2009) using a similar cohort of women (2000-2006). A problematic secondary analysis of data was also reported (See review: Section B, III, A).

C) de Jonge, A., van der Goes, B. Y., Ravelli, A. C., Amelink-Verburg, M. P., Mol, B. W., Nijhuis, J. G., Bennebroek Gravenhorst, J., & Buitendijk, S. E. (2009). Perinatal mortality and morbidity in a nationwide cohort of 529,688 low-risk planned home and hospital births. BJOG, 116(9):1177-84, doi: 10.1111/j.1471-0528.2009.02175.x.

Retrospective cohort study of 529,688 low-risk women in the Netherlands who were in primary midwife-led care at labour onset. This study compared perinatal mortality and morbidity between planned home births (321,301; 60.7%), planned hospital births (163,261; 30.8%), and unknown place of birth (45,120; 8.5%), using the national perinatal and neonatal registration data from 2000-2006. The following differences between groups were controlled for using logistic regression: parity, gestational age, maternal age, ethnic background, and socio-economic status. Inclusion criteria ensured the subjects were strictly low-risk. The main outcomes were intrapartum death, intrapartum and neonatal death within 24 hours and 7 days after birth, and admission to a neonatal intensive care unit. No significant differences were found between planned home and planned hospital births for any of the main outcomes. The authors concluded that planned home birth in a low-risk population is not associated with higher perinatal mortality rates or an increased risk of admission to a NICU compared to planned hospital birth.

D) Kennare, R. M., Keirse, M. J., Tucker, G. R., & Chan, A. C. (2009). Planned home and hospital births in South Australia 1991-2006: differences in outcomes. Med J Aust, 192(2):76-80.

Retrospective population based-study of all births and perinatal deaths from 1991-2006 in South Australia. 1141 planned home births and 297,192 hospital births were included. Planned home birth was defined as any birth that was intended to occur at home at the time of antenatal booking; 30.6% of the planned home births occurred in hospital. Perinatal outcomes studied were: perinatal death, intrapartum death, intrapartum asphyxiation, Apgar of <7 at 5 minutes, use of pediatric or specialized neonatal care. Maternal outcomes studied were: operative delivery, postpartum hemorrhage and perineal trauma including episiotomy (1998-2006 only). Results: Post-term pregnancy (≥42 weeks) was more common in the home birth group; 58% (n=25/43 post-term pregnancies) delivered at home. Perinatal mortality rates (including intrapartum fetal death and stillbirth) were similar between home and hospital groups (7.9 vs. 8.2 per 1000). There was no statistical difference in perinatal mortality between the home and hospital group (4.6 vs. 6.7 per 1000 respectively). Intrapartum fetal death was higher in the home birth group (1.8 vs .8 per 1000), though the absolute numbers were small. Cases of intrapartum death were not necessarily contingent upon place of birth. Of the 9 perinatal deaths total, 3 were antepartum (occurred after transfer to hospital and were unrelated to antenatal care), 2 were attributable to fetal congenital anomaly, and 4 occurred after the parents refused/delayed transfer or declined intervention after transfer. These deaths might indicate a lack of integration of South Australian midwives into the health care system or an underlying distrust of hospitals for parents. The home birth group had lower rates of caesarean delivery (aOR= .27), instrumental delivery (aOR= .33), and episiotomy (aOR= .14).

E) Chamberlain, G., Wraight, A., & Crowley, P. (Eds.). (1997). Home births: The report of the 1994 confidential enquiry by the National Birthday Trust Fund. Lancaster, UK: Parthenon Publishing Group Ltd.

Comprehensive investigation of the characteristics and outcomes of planned home births across the United Kingdom, endorsed by the Royal Colleges of Obstetricians, Midwives, and General Practitioners. A prospective trial of 6044 planned home births in Great Britain compared mortality and perinatal outcomes with a low risk hospital group and found no significant differences in mortality. The home birth group experienced significantly fewer medical interventions and perinatal complications. The study report is published as a book.

F) Ackermann-Liebrich, U., Voegeli, T., Gunter-Witt, K., Kunz, I., Zullig, M., Schindler, C., Maurer, M., & Zurich Study Team. (1996). Home versus hospital deliveries: follow up study of matched pairs for procedures and outcome. BMJ, 313(7068):1313-18.  (full text)

Prospective matched cohort study of 489 planned home and 385 planned hospital births. The study design carefully attended to issues of planning status, transfer criteria, and actual place of delivery. The groups were matched according to age, parity, gynecologic and obstetric history, medical history, partner situation, social class, and nationality. The main outcome measures were need for medication and/or intrapartum intervention, duration of labor, severity of lacerations, hemorrhage, neonatal condition and perinatal mortality. They found a lower incidence of interventions, medications, lacerations and higher Apgar scores in the home birth group and no differences in birth weight, clinical condition, or gestational age between groups. There were no differences in mortality between groups.

G) Wiegers, T. A., Keirse, M. J., van der Zee, J., & Berghs, G. A. (1996). Outcome of planned home and planned hospital births in low risk pregnancies: prospective study in midwifery practices in the Netherlands.BMJ 313(7068):1309-13.   (full text)

Prospective cohort study of 1836 women with low risk pregnancies (1140 planned home and 696 planned hospital births). The design controlled for provider type, parity, social, medical and obstetric background. The authors developed a tool that assigns an overall perinatal outcome index score based on “maximal result with minimal intervention”. This tool assigns scores for each of 22 intrapartum variables (indicating risk factors and intervention), 9 items on the condition of the newborn, and 5 postpartum outcomes/conditions to assign an overall perinatal outcome index. The authors assert that this tool allows researchers to evaluate factors that detract from optimal perinatal health as well as to weight each variables’ clinical significance and cumulative effect. The optimality index has subsequently been adapted and validated for North American and international contexts with evidence based rationale for the exclusion or inclusion of each variable. This study found no relationship between planned place of birth and perinatal outcomes in nulliparas when controlling for background variables (more or less favourable background); multiparas had significantly better perinatal outcomes in the home setting, regardless of background.

H) Northern Region Perinatal Mortality Survey Coordinating Group. (1996). Collaborative survey of perinatal loss in planned and unplanned home births. BMJ 313(7068):1306-09.  (full text)

The Coordinating Group collected and analyzed data for 558,691 births over 14 years in the UK (1981-1994), with 2888 booked for home delivery at term. They found perinatal mortality in the planned home birth group was less than half the average for all births even when the cases referred to hospital were included. Mortality for unplanned home births was four times as high as for all registered births. Perinatal mortality for women booked for home delivery was judged mostly unavoidable and not associated with place. Home birth critics often misquote this study as 134 losses in 3466 births, but 97% of those losses occurred in unplanned home births. The remaining losses were due to causes unaffected by birth site. Further analysis comparing data from the planned home birth group to low risk term hospital births concluded that there were no significant differences in rates of perinatal mortality.

V: Descriptive Studies & Registry Reports Observational Studies: International

A) MacDorman, M. F., Declercq, E., & Menacker, F. (2011). Trends and characteristics of home births in the United States by race and ethnicity, 1990-2006. Birth 38(1):17-23.

MacDorman et al. used data from the U.S National Center for Health Statistics to examine the trends and characteristics of home births in the United States from 1990 to 2006 with a focus on race, ethnic and geographic differences. Home birth was more common among non-Hispanic white women, over the age of 30, multigravid, married, delivering a singleton, term baby, and delivering with midwives. While home birth rates steadily increased for non-Hispanic whites, they declined for all other races and ethnic groups. Home births to non-Hispanic white women were mostly attended by midwives and were less likely to be preterm. Home births for all other ethnic groups were more likely to be preterm and delivered by either physicians or ‘other’ attendants, suggesting that these births were likely ‘unplanned’ emergency home births. Birth certificates in many states in the US currently do not distinguish between planned and unplanned home births.

B) Declercq, E., MacDorman, M. F., Menacker, F., & Stotland, N. (2010). Characteristics of planned and unplanned home births in 19 states. Obstet Gynecol 116(1):93-9.

Declercq et al. used data from the 2006 U.S. vital statistics in 19 states to compare the sociodemographic profiles of women choosing planned home births with women who had unplanned home births. Approximately 83.2% (n= 9,810) of the total home births occurring in the 19 states (N=11,787) were planned home births. Women in the unplanned home birth group were more likely to be non-white, younger, unmarried, foreign-born, smokers, have no prenatal care and no college education. Unplanned home births are more likely to be pre-term, and attended by someone who is listed as ‘other’ or unknown on the birth certificate. The majority of planned home births were attended by “other midwives”. Birth certificate data do not include information about planned or unplanned home birth transfer to hospital, nor can they guarantee the accuracy of the planning status variable.

C) Amelink-Verburg, M. P., Verloove-Vanhorick, S. P., Hakkenberg, R. M., Veldhuijzen, I. M., Bennebroek Gravenhorst, J., & Buitendijk, S. E. (2008). Evaluation of 280,000 cases in Dutch midwifery practices: A descriptive study. BJOG, 115(5):570-78.

This study discusses the importance of effective home birth risk selection in the Dutch obstetric system. The authors found that the current selection process results in a small number of urgent referrals and favourable perinatal outcomes for home births.

D) Murphy, P. A. & Fullerton, J. (1998). Outcomes of intended home births in nurse-midwifery practice: A prospective descriptive study. Obstet Gynecol, 92(3):461-70.

Prospective study describing various outcomes of home births attended by CNMs during 1994-1995 (n=1404). Of those beginning labour at home, 102 (8.3%) were transferred to the hospital in labour, 10 (0.8%) were postpartum transfers and 14 (1.1%) infants were transferred. For the whole sample of women beginning labour at home, fetal and neonatal mortality was 2.5/1000. For those actually birthing at home this mortality was 1.8/1000. Intrapartal problems were positively associated with transfer to hospital-based care, and overall outcomes were consistent with expected outcomes for low-risk birth.

E) Cawthon L. (1996). Planned home births: Outcomes among Medicaid women in Washington State. Olympia, WA: Washington Department of Social and Health Services.  (full text PDF)

This study described perinatal data for 2,054 Medicaid women who were cared for by licensed midwives between 1989 and 1994. Births were categorized by birth place, maternal characteristics, prenatal care; outcomes between planned home births and births in birth centers or in hospitals were compared. Researchers compared all women receiving some care from licensed midwives, women receiving care from certified nurse midwives, and all other Medicaid women and found no statistically significant differences in mortality rates. Congenital anomalies and SIDS caused the majority of deaths. The number of stillbirths or neonatal deaths among women who delivered at home was zero (0), and the rate of transfer to hospital delivery for the women who experienced fetal or neonatal death was 100%, suggesting appropriate screening and site selection by licensed midwives.

F) Anderson, R. E. & Murphy, P. A. (1995). Outcomes of 11,788 planned home births attended by certified nurse-midwives: A retrospective descriptive study. J Nurse Midwifery, 40(6):483-92.

A retrospective survey study of perinatal outcomes associated with 11,788 planned home births attended by certified nurse-midwives (CNMs) from 1987 to 1991. Over 60% of identified CNM home birth practices participated in this study. Perinatal mortality rates were very low: 0.9 per 1,000, excluding deaths due to congenital anomalies. Nurse-midwives who offer home birth utilized standard risk-assessment criteria, and were prepared for immediate resuscitation of the newborn and maternal complications. The authors conclude that planned home birth with qualified care providers is a safe alternative to hospital birth for low risk women.