3.1. Characteristics of the included studies
The comprehensive search produced a total of 2,042 articles. A total of 1619 duplicated studies were removed, and 423 studies were screened for title and abstract. After title and abstract screening, 346 studies were excluded, 77 were sought for retrieval, and 61 were assessed for eligibility by full-text screening. The search was completed on August 26, 2023; 36 full-text articles were obtained for review, and 21 studies that met our predetermined inclusion criteria for a systematic review and 20 for meta-analysis of planned birthplace were found.
Of 21 studies that meet the criteria for a systematic review of obstetric interventions and foetal-maternal outcomes, one study did not provide any published or author-provided data on perinatal, maternal, or foetal outcomes that could be incorporated into our meta-analysis (22). The meta-analyses comprised 20 original studies published and reported obstetric interventions and foetal-maternal outcomes for ~ 750,000 intended home births. The exact number of births can be different by analysis, which includes more of the extensive study sample from the same country (The Netherlands), where the data more substantially overlapped (15, 23–29). Only two randomised controlled trials containing relevant outcomes were included in our analysis (30, 31).
With caution, when the multiple comparison groups were selected and compared, strictly consider the used planned home birth group (led by midwives) and obstetric-led or hospital comparator or conventional centre (led by midwives) as the institutional group. Out of 21 studies that meet the criteria for a systematic review (of which 20 were included in the meta-analyses), they were in 10 settings. Twelve studies were from a country where home birth was considered to be “well integrated” into the healthcare system (The Netherlands, UK, and Denmark) (15, 21–28, 30–33); eight of them were in The Netherlands (15, 22, 32, 34, 23–29, 31), and nine studies from seven settings where home births considered “less well-integrated” settings (7, 30, 35–40). Although some countries in “less well-integrated” settings were well-integrated, the degree of integration varied by region (Belgium, Spain, Norway, Hungary, Italy, Iceland, and Lithuania), and the I2 score ranged between 0 and 100% (Table 1). Table 1 explains the characteristics of the included studies in detail (see Additional file 1)
Table 2
Characteristics of the included studies
Author(s), year, country | Study design | Data source and period | Data analysis by parity | Place of birth | Outcome(s) reported | NOS score (4-2-3) |
Planned homebirth n/N (%) | Planned Institutional birth n/N (%) |
Isaline et al., 2019, Belgium (34) | Retrospective cohort | MLOZ (Mutualit ØLibres Onafhankelijke Ziekenfondsen) was the third Belgian statutory healthcare insurer from 1 March 2014 until 31 October 2015 | not stratified | 59/89 (66.3) | 30/89 (33.7) | 2–5, 20 | 3-2-3 |
A de Jonge, 2013, Netherlands (22) | Retrospective cohort | "LEMMoN” data from 1 August 2004 to 1 August 2006, combined with data from the Netherlands perinatal registry | Adjusted and stratified | 92,333/146,752 (62.9) | 54,419/146,752 (37.1) | 8 | 3-1-2 |
A deJonge et al, 2014, Netherlands (23) | Retrospective cohort | National perinatal and neonatal registration data over seven years from 2000 to 2006 | Adjusted and stratified | 321,307/529,688 (60.7) | 63,261/529,688 (30.8) | 12, 15, 16 | 4-2-2 |
Wiegerinck et al, 2018, Netherlands (24) | Retrospective cohort | National Perinatal Register (PRN) and extra medical information obtained from hospitals and midwifery units in Amsterdam between January 1, 2005, and December 31, 2007. | not stratified | 46,764/57,396 (81.5) | 10,632/57,396 (18.5) | 2,4, 5, 8, 12, 19 | 3-2-3 |
Palau-Costafreda et al, 2023, Spain (35) | Retrospective cohort | Medical history, obstetric characteristics, and maternal and neonatal outcomes were recorded by obstetric health- care professionals in patient charts and the hospital's information system in MLU and OU between January 2018 and December 2020 | NP (not performed) | 255/878(29.0) | 623/878(71.0) | 2–5, 8–11, 16, 19 | 3-1-3 |
Blix et al, 2012, Norway (36) | Retrospective cohort | Midwives' registers, telephone interviews, and birth protocols were used for homebirth data, and the Medical Birth Registry of Norway (MBRN) was used for institutional data from 1990 to 2007. | Adjusted and stratified | 1631/17,941(9.1) | 16,310/17,941(90.9) | 2–5, 8, 9, 12, 15, 16, 17, 19 | 2-1-3 |
Bolten et al, 2016, Netherlands (25) | Prospective cohort | DELIVER Study participants from 20 midwifery practices between January 2018 and December 2020. | Adjusted and stratified | 2050/3495 (58.7) | 1445/3495 (41.3) | 1, 3–5, 8, 9, 12–18 | 4-2-2 |
Dobbie G et al, 1993, UK (30) | Randomised controlled trial | Leicester Royal Infirmary Maternity Hospital and home-like-home delivery scheme from 1 March 1989 until 6 July 1990. | not stratified | 2304/3510 (65.6) | 1206/3510 (34.4) | 1–5, 8, 10, 12, 13 | Rob 2 (see Additional file 2) |
Wami et al, 2022, Hungary (7) | Retrospective cohort | Home: Hungarian Tauffer database (perinatal registry) Institutional: University Hospital Gyn/Oby Clinic from 2012–2020 | not stratified | 1792/2997 (59.8) | 1205/2997 (40.2) | 8–14, 17, 19 | 3-2-2 |
Campiotti et al, 2020, Italy (37) | Case-control study | All out-of-hospital births reported in birth registry data from 2014 to 2018 were extracted. | not stratified | 848/1099 (77.0) | 251/1099 (23.0) | 1, 4, 9, 13, 15 | 3-2-2 |
Overgaard et al, 2012, Denmark (31) | Prospective cohort | Data were collected in a sparsely populated region of North Denmark from two FMUs and two OUs over 3.5 years (2004–2008). | NP (not performed) | 839/1678 (50.0) | 839/1678 (50.0) | 1,2, 4,5, 9, 19, 20 | 3-1-2 |
Kooy et al, 2017, Netherlands (15) | Retrospective cohort | The Perinatal Registry (PRN) in the Netherlands between 2000 and 2007 | not stratified | 402,912/679,952 (59.0) | 219,105/679,952 (32.0) | 2, 4, 5, 6, 12, 13, 15, 16, 18 | 4-1-3 |
Offerhaus et al, 2020, Netherlands (26) | Retrospective cohort | Data from all women who gave birth in 2015 and were registered in the Dutch Perinatal Registry (Perined). | stratified | 657/2611(25.2) | 1954/2611(74.8) | 3–5, 8, 9, 12–17, 19 | 3-2-2 |
VanHaaren-tenHaken et al, 2015, Netherlands (27) | Prospective cohort | A Dutch prospective cohort study involved 150 randomly selected practices to enrol women in midwife care, using a questionnaire and medical records from 2007–2011. | stratified | 226/576 (39.2) | 182/576 (31.6) | 1–5, 8–16, 18, 19, 21 | 4-2-2 |
Wiegerinck et.al, 2015, Netherlands (28) | Retrospective cohort | Data from hospitals and midwife practices in the catchment area of both academic hospitals linked with the perinatal register (PRN) in Amsterdam between 2005 and 2008. | not stratified | 26,128/56,294 (46.4) | 30,166/56,294 (53.6) | 1–5, 8, 9, 11, 12, 19, 20 | 3-2-2 |
Eide et al, 2009, Norway (38) | Prospective cohort | Data from the hospital and pregnancy records were gathered between November 3, 2001, and May 31, 2002 | NP (not performed) | 252/451 (55.9) | 201/451 (44.6) | 2–5, 9 | 3-1-2 |
Bernitz et al., 2011, Norway (29) | Randomised controlled trial | Women who delivered at the Østfold Hospital Trust, recorded in the electronic journal system of the department, partus (Clinsoft) from 2006–2009 | not stratified | 282/1,111 (25.4) | 412/1,111(37.1) | 1, 2, 4, 5, 8, 9, 17, 19, 20 | Rob 2 (see Additional file 2) |
Halfdansdottir et al, 2015, Iceland (32) | Retrospective cohort | An Icelandic electronic birth registry and direct contact with midwives and tertiary/rural secondary hospitals from 2005–2009 | Adjusted and stratified | 307/1,228 (25.0) | 921/1,228 (75.0) | 1–5, 8, 9, 13, 17–21 | 3-1-2 |
Van der Kooy et al, 2016, UK (33) | Retrospective cohort | Anonymised data of the UK National Perinatal Epidemiology Unit Report and the midwifery-led birth centre Sophia between January 2007 and June 2012 | Stratified | 443/5,953 (7.4) | 1391/5,953 (23.4) | 4, 5, 8, 9, 12–16, 18–20 | 4-2-2 |
Bartuseviciene et al, 2018, Lithuania (39) | Retrospective cohort | Hospital registry at the tertiary-care women’s hospital in Kaunas, Lithuania. | stratified | 477/1,283 (37.2) | 806/1,283 (62.8) | 3–5, 9, 10, 19 | 3-2-3 |
Christensen et al, 2017, Denmark (21) | Retrospective cohort | Data from two freestanding midwifery units and two obstetric units in the North Denmark Region from March 2004 to October 2008. | Adjusted and stratified | 839/1,678 (50.0) | 839/1,678 (50.0) | 1, 2, 4, 5, 9, 19, 20 | 3-1-2 |
Outcome variables: 1) Oxytocin administration/Augmentation, 2) Epidural analgesia, 3) Episiotomy, 4) Assisted vaginal delivery (forceps/vacuum), 5) Caesarean section(C/s), 6) Blood transfusion, 7) Maternal death, 8) Post-partum haemorrhage (PPH), 9) 3rd or 4th-degree perineal tear, 10) Vaginal or perineal laceration, 11) Retained placenta, 12) Perinatal or neonatal or foetal mortality, 13) Low birth weight < 2500g or < 10%, 14) Large birth weight ≥ 4000g or > 90%, 15) Preterm/premature born, 16) Post-term-born, 17) Non-vertex presentation, 18) Neonatal mal-formations (congenital anomalies), 19) Low Apgar score < 7 at one or 5mins, 20) NICU admission, 21) Neonatal ventilation/resuscitation |
For every stratum of analysis, an inverted funnel plot was made to evaluate the publication bias amongst studies on birthplaces. Eight plots were created, and the largest of our plots contained just fourteen studies. Nevertheless, plots with fewer studies (< 5) are more challenging to interpret and were not reported (Fig. 2). Figure 2 illustrates the funnel plot for publication bias evaluation by level of integration and outcomes.
3.2. Maternal outcomes and obstetric interventions
There were no maternal deaths reported in either group among the nearly 550,000 women included in these studies that specifically reported this outcome. Regarding maternal outcomes and obstetric interventions, homebirths were less likely to report an obstetric intervention. Without stratification by the level of integration and parity, women who are planning a home birth had a 60% lower chance of having a Caesarean section (11 studies, OR 0.40, 95% CI [0.24, 0.68]); 68% fewer women planned assisted vaginal births (vacuum or forceps; 14 studies, OR 0.68, 95% CI [0.55, 0.83]); 74% fewer women planned to use epidural analgesia (9 studies, OR 0.26, 95% CI [0.18, 0.37]); 14% higher likelihood of having an episiotomy (9 studies, OR 1.14, 95% CI [0.66, 0.73]) and 61% less likely had an oxytocin administration for labour induction (9 studies, OR 0.29, 95% CI [0.04, 2.39]). In addition, relatively fewer women who planned to give birth at home experienced adverse maternal outcomes than those who gave birth in an institution, with 59% lesser experience of having retained placenta (5 studies, OR 0.41, 95% CI [0.37, 0.45]) and 31% lesser reported postpartum haemorrhage (10 studies, OR 0.69, 95% CI [0.46, 1.03]). In contrast, vaginal/perineal laceration (5 studies, OR 1.22, 95% CI [0.58, 2.57]) and 3rd /4th -degree perineal tear (12 studies, OR 1.04, 95% CI [0.91, 1.19]) have shown no odds of statistical significance.
Compared to women planning an institutional delivery, those planning a home birth in well-integrated settings had a 76% lower chance of having a Caesarean section (6 studies, OR 0.24, 95% CI [0.12, 0.49]); 29% fewer women planned assisted vaginal births (vacuum or forceps) (8 studies, OR 0.71, 95% CI [0.56, 0.90]); 66% fewer women planned to use epidural analgesia (5 studies, OR 0.34, 95% CI [0.24, 0.46]); 59% had experienced retained placenta (3 studies, OR 0.41, 95% CI [0.37, 0.45]) and 10% had 3rd or 4th-degree perineal tear (6 studies, OR 1.10, 95% CI [1.01, 1.20]) and twice more likely had vaginal /perineal laceration (2 studies, OR 2.27, 95% CI [1.13, 4.56]), while episiotomy (4 studies, OR 1.69, 95% CI [0.71, 4.04]); postpartum haemorrhage (2 studies, OR 0.63, 95% CI [0.40, 1.00]) and oxytocin administration for labour induction (5 studies, OR 0.25, 95% CI [0.02, 3.80]) were statistically insignificant.
In “less well-integrated” settings, women who were planning a home birth had a 69% more likelihood of having a Caesarean section (5 studies, OR 0.69, 95% CI [0.46, 1.01]); 59% more women had assisted vaginal births (vacuum or forceps)(6 studies, OR 0.59, 95% CI [0.41, 0.85]); 19% fewer women planned to use epidural analgesia (4 studies, OR 0.19, 95% CI [0.05, 0.73]); 74% less likelihood of having an episiotomy (5 studies, OR 0.74, 95% CI [0.48, 1.13]); 51% had less likely experienced retained placenta (2 studies, OR 0.49, 95% CI [0.18, 1.36]) and 63% more had 3rd or 4th-degree perineal tear (6 studies, OR 0.63, 95% CI [0.41, 0.97]) and 43% less likely had an oxytocin administration for labour induction (2 studies, OR 0.43, 95% CI [0.23, 0.81]); 92% lesser reported postpartum haemorrhage (3 studies, OR 0.92, 95% CI [0.52, 1.64]) and 84% more vaginal/perineal laceration (3 studies, OR 0.84, 95% CI [0.50, 1.41]) (Table 3). Table 2 summarises the obstetric interventions and maternal outcome meta-analysis findings for all women by level of midwives' healthcare integration (see Additional file 1).
Table 4
Summary of the obstetric interventions and maternal outcome meta-analysis findings for all women by level of integration
Outcome (strata) | Number of studies | Place of birth | OR [95%CI] M-H, Random | I2(%) |
Planned homebirth n/N (%) | Planned Institutional birth n/N (%) |
Obstetric interventions | |
Oxytocin administration | 9 | 740/31908 | 11298/35432 | 0.29 [0.04, 2.39] | 100 |
Well-Integrated | 5 (19, 22, 23, 25, 27) | 556/30648 | 11158/35014 | 0.25 [0.02, 3.80] | 80 |
Less-well integrated | 2 (24, 29) | 184/1260 | 140/418 | 0.43 [0.23, 0.81] | 100 |
Epidural Analgesics | 9 | 2343/74753 | 4909/42733 | 0.26 [0.18, 0.37] | 100 |
Well-Integrated | 5 (18, 22, 23, 25, 27) | 2184/73775 | 4214/41597 | 0.34 [0.24, 0.46] | 93 |
Less-well integrated | 4 (24, 28–30) | 159/978 | 695/1136 | 0.19 [0.05, 0.73] | 97 |
Episiotomy | 9 | 3695/31164 | 5536/37474 | 1.14 [0.66, 1.96] | 98 |
Well-Integrated | 4 (21–23, 25) | 3448/30121 | 5120/35812 | 1.69 [0.71, 4.04] | 99 |
Less-well integrated | 5 (34,35,38,39) | 247/1043 | 416/1662 | 0.74 [0.48, 1.13] | 68 |
Assisted vaginal delivery (Vacuum/forceps) | 14 | 34862/480037 | 24827/266933 | 0.68 [0.55, 0.83] | 94 |
Well-Integrated | 8 (18, 20–23, 26, 27, 32) | 34734/478036 | 24670/264960 | 0.71 [0.56, 0.90] | 96 |
Less-well integrated | 6 (24, 28–31, 33) | 128/2001 | 157/1973 | 0.59 [0.41, 0.85] | 27 |
Caesarean section (C-section) | 11 | 12443/478751 | 15899/264590 | 0.40 [0.24, 0.68] | 99 |
Well-Integrated | 6 (18, 20–23, 27) | 12359/477296 | 15733/262648 | 0.24 [0.12, 0.49] | 99 |
Less-well integrated | 5 (24, 28–31) | 84/1455 | 166/1942 | 0.69 [0.46, 1.01] | 38 |
Maternal outcome | | | | | |
Post-partum haemorrhage (PPH) | 10 | 3905/172115 | 5107/104358 | 0.69 [0.46, 1.03] | 97 |
Well-Integrated | 7 (16, 18, 21–23, 25, 26) | 3867/169656 | 5068/102248 | 0.63 [0.40, 1.00] | 98 |
Less-well integrated | 3 (7, 24, 30) | 38/2459 | 39/2110 | 0.92 [0.52, 1.64] | 26 |
3rd or 4th-degree perineal tear | 12 | 2002/78532 | 1099/48133 | 1.04 [0.91, 1.19] | 20 |
Well-Integrated | 6 (18, 21–23, 26, 27) | 1960/74798 | 1045/44891 | 1.10 [1.01, 1.20] | 0 |
Less-well integrated | 6 (7, 24, 28, 30, 31, 33) | 42/3734 | 54/3242 | 0.63 [0.41, 0.97] | 0 |
Perineal or vaginal laceration | 5 | 1275/5831 | 991/6192 | 1.22 [0.58, 2.57] | 97 |
Well-Integrated | 2 (22, 25) | 961/3307 | 441/3649 | 2.27 [1.13, 4.56] | 84 |
Less-well integrated | 3 (7, 30, 31) | 314/2524 | 550/2543 | 0.84 [0.50, 1.41] | 88 |
Retained placenta | 5 | 939/75136 | 1221/42765 | 0.41 [0.37, 0.45] | 0 |
Well-Integrated | 3 (18, 22, 23) | 933/73089 | 1205/40937 | 0.41 [0.37, 0.45] | 0 |
Less-well integrated | 2 (7, 30) | 6/2047 | 16/1828 | 0.49 [0.18, 1.36] | 14 |
I2- Heterogeneity |
3.3. Foeto-neonatal outcomes
Although there was no significant variation in the perinatal mortality outcome by birthplace when the data was stratified by the level of midwives' health care integration and parity, we were limited in our evidence to conclude because additional studies were required for establishing an inference (pooled result stratified by parity: OR 1.87, 95%CI [0.74, 4.72], and by the level of integration: OR 1.05, 95%CI [0.62, 1.79]) (Figs. 3 and 4).
The foetal-neonatal outcomes of women planning a home birth demonstrated a 16% lower likelihood of low birth weight babies in well-integrated settings when compared to their comparable peers (5 studies, OR 0.84, 95% CI [0.75, 0.93]); 33% less likely had a non-vertex foetal presentation (3 studies, OR 0.67, 95% CI [0.48, 0.93]); 34% less likely to have an Apgar score of < 7 at 1/5minutes (7 studies, OR 0.66, 95% CI [0.58, 0.76]); 16% more likely to be a post-term born (5 studies, OR 1.16, 95% CI [1.04, 1.28]); 11% more likely had a preterm birth (6 studies, OR 0.89, 95% CI [0.81, 0.98]); 53% of the patients experienced NICU admission (4 studies, OR 0.47, 95% CI [0.34, 0.64]). However, there was no discernible variation in congenital malformations/neonatal abnormalities, which was large for gestational-age newborns by birthplace.
In “less well-integrated” settings, the foetal-neonatal outcomes of women planning home births, with limited evidence, revealed that there was a twofold increase in the likelihood of having a newborn with large for gestational age (1 study, OR 1.96, 95% CI [1.678, 2.29]) and a 54 per cent reduction in the likelihood of non-vertex foetal presentation (2 studies, OR 0.46, 95% CI [0.16, 1.37]). A lower Apgar score (< 7 at 1/5mins) (4 studies, OR 1.47, 95% CI [0.21, 10.26]); low birth weight babies (2 studies, OR 0.73, 95% CI [0.31, 1.70]); post-term born (1 studies, OR 4.95, 95% CI [0.90, 27.19]); and NICU admission (2 studies, OR 1.11, 95% CI [0.64, 1.93]) did not, however, significantly differ foetal-neonatal outcomes by birthplace and none of the studies from “less well-integrated” settings has reported congenital malformations/neonatal abnormalities, and preterm/premature births (Table 5). Table 3 summarises foetal-neonatal outcome meta-analysis findings by the midwifery health care integration level in detail (see Additional file 1).
Table 6
Summary of foetal-neonatal outcome meta-analysis findings by level of midwifery healthcare integration
Foeto-neonatal outcome (strata) | Number of studies | Place of birth | OR [95%CI] M-H, Random | I2 |
Planned homebirth n/N (%) | Planned Institutional birth n/N (%) |
Perinatal or neonatal or foetal mortality | 10 | 731/806794 | 552/448979 | 1.05 [0.62, 1.79] | 82 |
Well-Integrated | 8 (17–21, 23, 25, 26) | 716/803371 | 529/431464 | 1.00 [0.54, 1.85] | 0 |
Less-well integrated | 2 (7, 34) | 15/3423 | 23/17515 | 1.37 [0.61, 3.07] | 85 |
Low birth weight < 2500g or SGA < 10% | 7 | 28242/407185 | 18570/225009 | 0.84 [0.69, 1.03] | 46 |
Well-Integrated | 5 (20, 21, 26, 32) | 28105/404545 | 18503/223553 | 0.84 [0.75, 0.93] | 9 |
Less-well integrated | 2 (7, 33) | 137/2640 | 67/1456 | 0.73 [0.31, 1.70] | 85 |
Large birth weight ≥ 4000g or LGA > 90% | 3 | 844/2675 | 444/3341 | 1.41 [0.83, 2.40] | 79 |
Well-Integrated | 2 (21, 22) | 46/883 | 94/2136 | 1.04 [0.71, 1.53] | 0 |
Less-well integrated | 1 (7) | 798/1792 | 350/1205 | 1.96 [1.68, 2.29] | Na |
Preterm/premature born | | | | | |
Well-Integrated | 6 (17, 20–22, 25, 26) | 20236/728655 | 12586/389403 | 0.89 [0.81, 0.98] | 80 |
Less-well integrated | 0 | - | - | - | - |
Post-term born | 6 | 77296/725800 | 36760/386516 | 1.17 [1.05, 1.30] | 81 |
Well-Integrated | 5 (17, 20–22, 26) | 77292/725545 | 36758/385893 | 1.16 [1.04, 1.28] | 83 |
Less-well integrated | 1 (35) | 4/255 | 2/623 | 4.95 [0.90, 27.19] | Na |
Non-vertex presentation | 5 | 67/3394 | 204/4544 | 0.59 [0.41, 0.85] | 33 |
Well-Integrated | 3 (21, 22, 32) | 46/1190 | 173/3057 | 0.67 [0.48, 0.93] | 0 |
Less-well integrated | 2 (7, 24) | 21/2204 | 31/1487 | 0.46 [0.16, 1.37] | 71 |
Neonatal mal-formations (congenital anomalies) | | | | | |
Well-Integrated | 4 (20, 22, 26, 32) | 4727/403888 | 2936/221599 | 0.61 [0.34, 1.08] | 62 |
Less-well integrated | - | - | - | - | - |
Apgar score < 7 at one or 5mins | 11 | 559/79806 | 549/49293 | 0.73 [0.55, 0.96] | 44 |
Well-Integrated | 7 (18, 19, 21–23, 26, 27) | 523/76870 | 519/46377 | 0.66 [0.58, 0.76] | 0 |
Less-well integrated | 4 (7, 24, 30, 31) | 36/2936 | 30/2916 | 1.47 [0.21, 10.26] | 76 |
NICU admission | 6 | 186/74415 | 263/43117 | 0.60 [0.40, 0.90] | 65 |
Well-Integrated | 4 (18, 23, 26, 27) | 149/73944 | 241/42805 | 0.47 [0.34, 0.64] | 37 |
Less-well integrated | 2 (24, 29) | 37/471 | 22/312 | 1.11 [0.64, 1.93] | 0 |
I2- Heterogeneity, Na- not applicable, NICU-Neonatal intensive care unit