California College of Midwives 

Press Release ~  January 2003  

Obstetricians Use Dubious Method to Discredit Homebirth --

Motives Questioned by Midwives and Public Health Researchers

For the last six months, the American College of Obstetricians and Gynecologists (ACOG) has widely publicized a new study from Washington State that makes homebirth appear unsafe. Unfortunately, ACOG promoted the study entitled “Outcomes of Planned Home Birth In Washington State” (by Pang et al, 2002) in newspapers across the country before it could be reviewed by the wider scientific community. A closer look reveals major flaws in the study skewed that the results. Parents and midwives’ associations across North America are concerned that a prestigious medical organization is actively promoting this flawed homebirth study.

The great majority of studies of healthy women who choose home-based birth services under the care of an experienced birth attendant document safe outcomes for both mothers and their babies
. This newest study from Washington State was intended to compare neonatal outcomes for “planned” home births with those of planned hospital births. However, it used birth certificate data from a state which does not collect data on the intended place of birth. The criteria used to determine planned place of birth was designed by the research group itself and did not include input or conformation from either the midwife or the parents. The authors admitted that an unknown number of unplanned and unattended home births may have been included. This makes the study’s conclusions unreliable.  

Out of the approximately 4 million births each year in the US, the total number of births in the Washington state study was only 16,726. The total number of neonatal deaths for both home and hospital groups was only 38. This small number is further reduced when congenital anomalies, which were responsible for one half the losses (19 babies) are subtracted, leaving only 19 deaths out of nearly 17,000 births. This included 7 in the hospital cohort (0.7 per 1,000) and 12 (1.95 per 1,000) in the home birth group. However, no chart reviews were done, even though this extremely small number means that one or two mistaken or misclassified cases would tip the scale one way or the other. In many categories the research paper substituted "educated guesses" for factual data, used "soft" data to arrive at "hard" conclusions and came to global conclusions based on very narrow criteria that included (by their own admission) missing and misclassified data.

The study design ignored the increased cost and maternal complications caused by a very high rate of medical interventions associated with "planned hospital birth", (“Listening to Mothers” Survey ) such as induction and Pitocin-accelerated labors, narcotics, episiotomies, forceps and 20 percent-plus Cesarean section rate. The study also overlooked all the long-term complications that go along with the high operative delivery rate in planned hospital births. This includes pelvic organ problems and incontinence related to forceps use (20%) and potentially fatal complications in post-cesarean pregnancies. (3 ObGynNews) Previous Cesarean deliveries are strongly associated with placental abnormalities which cause maternal hemorrhages and often require an emergency hysterectomy. (3 separate ObGynNews)

In the January 2003 issue of Obstetrics and Gynecology, a letter to the editor (see attached) compares the Washington study to ones done with more sophisticated methods. Those studies show that about 10% of women planning home births are transported to the hospital for medical care. However, in the Washington study the percentage of home birth transfers  was much lower -- only 4%.  This strongly suggests that the so-called “planned home births” in Washington included a large number of unplanned and unattended home births. Since the 1980s, researchers in the field have known that it is critical to distinguish the homebirths that were “planned” with a skilled attendant, from those that occurred at home by accident, without a knowledgeable person or basic medical equipment available. 

Unplanned or unattended home births have always been more likely to have a bad outcome. Since there  is no professional present during the labor or birth, there is no one to prevent an emergency or recognize a complication early on and help the mother to get timely medical care. The abnormally low rate of transfer in this study suggests that bad sampling methods were used. However, the authors of this study failed to respond to  criticism about this problem.  Because of the many flaws in scientific methods used by the authors, the Washington study’s outcomes cannot be generalized to all women planning homebirths attended by midwives.

Long opposed to homebirth, ACOG has been unable to identify solid scientific evidence that supports their opposition. As the number of studies with good methods and large enough sample size increase, it is evident that with proper professional care, birth outcomes of healthy women are similar whether they choose to give birth in home or hospital.

Licensed midwives in Washington state brought these flawed methods to the attention of the research team long before the study was completed or published.  They pointed out that a physician often signs the birth certificate when a baby who was born at home without an attendant is subsequently transferred to the hospital. This type of birth certificate data gives the false impression that the physician attended a home birth.  The appeal by midwives’ to accurately represent all this vital information was ignored and the press was given the story before its publication in a medical journal. This violates the normal protocol followed in scientific circles. As a result, before the study’s methods were evaluated, misleading headlines reached the public.

In contrast to the ACOG stand, the American Public Health Association (APHA) passed a resolution in 2001 to increase access to out-of-hospital birth attended by direct entry (that is, non-nurse) midwives. The APHA based their decision on studies carried out with more sophisticated methodology now required of homebirth research, which the Washington study researchers did not use.  Most notably, a study using data from women intending to deliver with a Certified Professional Midwife in the year 2000 was scrutinized and the methodology critiqued by epidemiologists in the APHA even before it was carried out.

In Canada some physicians’ colleges have removed the ban from physicians doing homebirths since midwifery legislation in the 1990s.


In spite of the controversy over the numbers used by this study, the bottom line is that the neonatal mortality rate for home-based care is very low and generally on a par with hospital-based obstetric care. Most studies put neonatal mortality at about 2 per 1,000, excluding congenital anomalies (cite P. Schlenzka) for home, hospital or birth centers. Using that criteria, even this flawed study only recorded a rate of 1.95 per 1,000, with a total Cesarean / operative rate of less than 4%. That is a difference of slightly more than one per 1,000 or 1/10 of a percent.  

However this 1/10 of one percent “advantage” in the hospital cohort came with the consider cost (economic and otherwise) of an approximately 30% operative rate, which was seven times higher for planned hospital births and can result in maternal-infant mortality in subsequent pregnancies. Were immediate and down-stream complications associated with the normally high operative rate for planned hospital births, especially those in post-cesarean women, factored into the our definition of “relative safety”, then home-based birth services would have to be considered a protective form of care for the babies of healthy women.