Make a Donation
     Connect with UsFacebookTwitterYouTube
    Google Custom Search

    Questions? Ask Henci!

    Find out what other moms-to-be are asking. Join in the discussion with Henci Goer, whose expertise is determining what the research tells us best promotes safe, healthy birth. If you would like to contact Henci outside of the Ask Henci forum, send an email to

    You must establish a username and password to participate in the Ask Henci forum, click here to submit your request.

    Archived User
    Dear Ms. Goer,

    What was the neonatal mortality rate in the hospital in the year 2000? That's what we need to know in order to evaluate the claims of the Johnson and Daviss study.

    I went back and looked at the neonatal mortality data for this group, the EXACT group that Johnson and Daviss felt was the perfect comparison for intervention rates. I did this by reviewing the exact same paper that Johnson and Daviss used. As you probably know, the paper is 105 pages long and has been divided into subsets for ease of research. The subset on neonatal mortality is Infant Mortality Statistics from the 2000 Period Linked Birth/Infant Death Data Set. Looking at the raw data we find:

    2,824,196 births to white women at term (37+ weeks), see Table 2
    2,602 deaths of white babies weighing more that 2500 gm see Table 6
    a death rate of 0.9/1000.

    That's before congenital anomalies, breech and twins are excluded from the hospital group.

    Johnson and Davis reported the following:

    "After we excluded ... three babies with fatal birth defects, five deaths were intrapartum and six occurred during the neonatal period. This was a rate of 2.0 deaths per 1000 intended home births. The intrapartum and neonatal mortality was 1.7 deaths per 1000 low risk intended home births after planned breeches and twins (not considered low risk) were excluded."

    If the congenital anomalies, breeches and twins are added back in, the death rate in 2.7/1000.

    So, the neonatal death rate for white women at term in the year 2000 was 0.9/1000 and the neonatal death rate at homebirth was 2.7/1000. This is almost 3 times higher than the hospital death rate. Chi square analysis shows this to be statistically significant (p is less than 0.005).

    Using the exact same raw data that Johnson and Daviss used, we see that homebirth with a CPM in the year 2000 had a significantly higher neonatal death rate than hospital birth. Therefore, Johnson and Daviss never showed that homebirth is as safe as hospital birth, regardless of what they claimed; they showed that it was considerably more dangerous for babies.
    By: ATuteur
    Archived User
    Amy, Amy, Amy, give it a rest. I'm going to paste in my response to the excess neonatal death rate argument you made in the Lamaze blog some weeks ago. As I pointed out, you were using the wrong statistic for making the comparison then, and you are still using the wrong statistic now.

    Pasted in from Lamaze blog thread:

    "Amy Tuteur has made the egregious error of comparing apples to oranges. She cites a neonatal death rate (babies born alive who die on or before the 28th day of life) of 1 per 1000 in low-risk white women and compares this rate with two home birth studies (Murphy and Fullerton 1998 and Johnson and Daviss 2005) in order to show that home birth results in an excess of deaths. However, contrary to her assertion, the studies she compares unfavorably with her statistic report neonatal deaths PLUS intrapartum deaths (babies alive at the start of labor who are born dead).

    Fortunately, both studies allow us to calculate neonatal mortality rates so that we can compare “apples to apples.” Johnson and Daviss report 11 deaths overall in 5418 women intending home birth at the onset of labor. Of these, five were intrapartum deaths and six were neonatal deaths. The neonatal death rate (6/5418) in the Johnson and Daviss study calculates to—wait for it—1 per 1000.

    "Turning to Murphy and Fullerton, Tuteur incorrectly states that the neonatal death rate was 4.1 per 1000. In actual fact, the intrapartum plus neonatal death rate was 2.5 per 1000 (3 intrapartum or neonatal deaths in 1221 women) in women who had a baby alive at the start of labor. The 4.1 number included two women whose babies were not alive at first labor examination and who were immediately transferred to hospital care. Only one of the three remaining deaths was a neonatal death, so again, the neonatal death rate (1/1221) calculates to 0.8 per 1000 or roughly 1 per 1000.

    "Of further note, two of the three deaths in Murphy and Fullerton could have had nothing to do with the choice to birth at home. In one case, the woman was transferred to the hospital during labor for slow progress and meconium passage. The fetal heart rate was normal at hospital admission. She labored several hours more in the hospital and delivered a stillborn infant. Meconium aspiration was listed as the cause of death. The second infant died a day after birth and postmortem studies could not establish a cause. The third case was the home birth of a stillborn infant in the presence of meconium."

    -- Henci

    By: Henci Goer
    Archived User

    Something else that I would point out in these comparisons...

    There is some evidence that women who seek care with homebirth midwives are significantly less likely to give birth preterm than those who seek care with obstetricians--possibly as much as 3 times higher in women seeking care with obstetricians.  Since preterm birth is a leading cause of neonatal death, I'd argue that the most accurate comparison of homebirth neonatal mortality compared to hospital neonatal mortality would require that a matched group of women intending hospital birth at the start of the pregnancy and women intending homebirth at the start of the pregnancy be followed to look at neonatal death rate in each group as a total, regardless of prematurity.

    Also, the "low risk" hospital birth group excludes twins in a defacto manner because the vast majority of all hospital birthed twins are born prior to 37 weeks because if they don't come before then on their own (labor starting or some indication arrising for induction/cesarean), then Dr.'s routinely induce twins at 37 weeks for no other reason than twin status.

    Henci Goer

    It's an interesting idea for a study, but there would be problems carrying it out. Neonatal death is sufficiently rare that you would probably need thousands of participants in order to have a reasonable probability of detecting a statistically significant (meaning not likely to be due to chance) difference between groups. I think you'd have some problems matching up the two populations as well.

    To add on to your point, though, obstetricians do way more cesarean surgeries on average than are necessary, and we have data that cesarean surgeries are associated with increased risk of perinatal death in the current and future pregnancies. It must logically follow, as it does in your "more preterm deliveries" argument, that those avoidable cesareans lead to some number of excess perinatal deaths, if not in the original pregnancy then in subsequent ones.

    -- Henci

    All Times America/New_York

    Forum Disclaimer

    Please note that this Forum is intended to help women make informed decisions about their care. The content is not a substitute for medical advice.

    Copyright 2015 Lamaze International. All rights reserved. Privacy Statement | Terms of Use