Three large studies have compared safety outcomes of midwives and doctors. The first is a systematic review of research on midwifery care that meets the “gold standard” for medical evidence. That review looked at studies from Canada, Australia, Ireland, and the UK. The other two studies do not meet the highest standards for research design, but they offer evidence specific to the United States.

Best evidence/International Review:

A 2016 Cochrane review analyzed data from 15 studies that examined outcomes for well over 17,000 babies and mothers delivering in Canada, Australia, Ireland, and the UK. Sandall J, Soltani H, Gates S, Shennan A, Devane D. Midwife-led continuity models versus other models of care for childbearing women. Cochrane Database of Systematic Reviews 2016, Issue 4. Art. No.: CD004667.  The review compared outcomes for women who received midwife-led care versus outcomes for women who received either doctor-led care or joint care from midwives and doctors. The women in the studies had some risk factors in both the midwife and doctor/joint care groups, but most were low-risk. None of the studies included out-of-hospital births. All 15 studies included in the review met the “gold standard” for research because they were randomized, prospective, controlled studies.

The researchers found that most outcomes were similar for midwife-led care and doctor-led or joint care:

There were no differences between groups for fetal loss equal to/after 24 weeks and neonatal death, . . . perineal laceration requiring suturing, postpartum haemorrhage, . . .  low birthweight infant, five minute Apgar score less than or equal to seven, neonatal convulsions, admission of infant to special care or neonatal intensive care unit(s) or in mean length of neonatal hospital stay (days).

However the Cochrane review did find a few safety outcomes for low-risk women that were better for midwife-led care, notably:

[W]omen who received midwife-led continuity of care were . . . less likely to experience preterm birth, or lose their baby before 24 weeks’ gestation, and to lose their baby overall, although there were no differences in losing the baby after 24 weeks. [Plain Language Summary at p. 2]

The authors conclude:

[T]he probability is that midwife-led continuity models of care are associated with a reduction in fetal loss and neonatal death by approximately 16%. [at p. 23]

It is worth noting that these differences were seen in a total population that included women with mixed risk. Of the 17,674 women represented in the review, “11,195 women . . . were defined to be at low risk by trial authors . . .  [and] 6578 women [were] defined to be at mixed risk of complications by trial authors.” In other words, more than a third of the women in the review were part of studies that included medium- or higher-risk women. So the excellent outcomes associated with midwifery-led care were observed with a group of women who were not uniformly low-risk.

The Cochrane review looked only at the highest quality of evidence, so these findings are very reliable. However, none of the studies included in the review used data from the United States.

Good evidence/US Studies:

The first of the two US-focused studies is a 2012 systematic review of 21 studies comparing midwifery care and doctor care in the hospital. Johantgen,M., Fountain, L., Zangaro, G., Newhouse, R., Stanik-Hutt, J., White, K., Comparison of Labor and Delivery Care Provided by Certified Nurse-Midwives and Physicians: A Systematic Review, 1990 to 2008, Women’s Health Issues 22-1 (2012) e73–e81.Johantgen,M., Fountain, L., Zangaro, G., Newhouse, R., Stanik-Hutt, J., White, K., Comparison of Labor and Delivery Care Provided by Certified Nurse-Midwives and Physicians: A Systematic Review, 1990 to 2008, Women’s Health Issues 22-1 (2012) e73–e81.

Unlike the Cochrane review, this US review was not limited to randomized, controlled trials (although 2 of the 21 studies met that standard). Because the included studies were not all randomized, and because the study designs varied in how they accounted for differences in underlying risk, the Johantgen et al. study is of lower quality than the Cochrane review. However, it still provides useful information about care in the US.

Looking at the pooled data in these 21 studies, the researchers concluded that doctors and midwives delivering in the hospital had similar outcomes on most measures relating to infant health. On a few, however—Apgar score, birthweight, and rates of Neonatal Intensive Care Unit (NICU) admissions—outcomes were better overall for babies delivered by midwives. (Table 3, “Infant outcomes”.) In addition, the one outcome relating to maternal health—3rd or 4th degree perineal lacerations—was better for mothers attended by midwives.

So, like the Cochrane review, this systematic review of data specifically from the US found that outcomes for hospital birth were generally similar for doctors and midwives, but that a few outcomes were better for midwives than for doctors.

The second US study is a very large retrospective cohort study of home birth outcomes published in 2017.  Grünebaum A, McCullough LB, Sapra KJ, et al. Planned home births: the need for additional contraindications. Am J Obstet Gynecol 2017;216:401.e1-8.

The goal of this study was to look at home birth outcomes, not to investigate differences between hospital midwives and doctors. In order to have a basis for comparing home birth outcomes, however, the researchers also looked separately at outcomes for hospital births attended by Certified Nurse Midwives (CNMs) and hospital births attended by doctors.

After excluding multiples, preterm infants, and babies with low birthweight or congenital anomalies, the researchers were left with a data pool that included outcomes for nearly 13 million births, of which over 1 million were midwife-attended births in the hospital, nearly 12 million were doctor-attended births in the hospital, and not quite 100,000 were planned home births.

In this very large sample, researchers found a somewhat lower overall infant mortality rate for midwife-attended hospital births compared to doctor-attended hospital births (3.08 per 10,000 births versus 5.09 per 10,000 births) (Grünebaum Table 2). When researchers separated the data for various risk factors, including number of previous births, gestational age over 41 weeks, maternal age over 35 years, breech presentation, and previous C-section, babies delivered by midwives had somewhat lower rates of neonatal mortality in every group except breech presentation (Grünebaum Table 3).

This study did not randomize women into midwife or doctor care groups. So, while the researchers controlled for many risk factors, it is possible that the doctor-attended group was overall higher-risk.

Nonetheless, the Grünebaum study offers some evidence to suggest that while outcomes for low-risk women who deliver in the hospital are generally similar for doctors and midwives, midwives may have have better outcomes overall.

Two of the three major studies comparing midwife and doctor care in the hospital looked at intervention rates as well as outcomes. These studies found that intervention rates for doctors are the same as or higher than intervention rates for midwives in every category.

The Cochrane review found similar intervention rates when comparing doctor-led or joint care to midwife-led care on most measures. However, several intervention rates were lower for midwife-led care:

The main benefits were that women who received midwife-led continuity of care were less likely to have an epidural. In addition, fewer women had episiotomies or instrumental births. Women’s chances of a spontaneous vaginal birth were also increased and there was no difference in the number of caesarean births. . . . The review identified no adverse effects compared with other models.

The 2012 systematic review of US studies by Johantgen et al. showed similar results. While many intervention rates were similar for midwives and doctors, some were clearly higher for doctors:

The four measures related to the process of birth have similar findings that favor the CNMs in the use of fewer cesarean deliveries, operative vaginal deliveries (forceps or vacuum), and episiotomy; and more vaginal births after cesarean delivery. 

Overall, then, the evidence suggests doctors have intervention rates that are similar to those of midwives for some measures and higher for others. [Note that while hospital-based midwives have excellent outcomes and intervention rates, intervention rates tend to be lower still for midwives who deliver in midwife-led birth centers.]