Articles

A critique of the leading study of American homebirth

Its authors boast that it is one of the ten most downloaded papers from the British Medical Journal (BMJ). That makes it even more unfortunate that the conclusions of the paper are directly at odds with the findings of the paper. Outcomes of planned home births with certified professional midwives: large prospective study in North America by Kenneth Johnson and Bettye Ann Davis is the premier paper on the safety of American homebirth. It claims to show that homebirth is as safe as hospital birth, but actually shows that homebirth has nearly triple the neonatal death rate of hospital birth for comparable risk women.

Johnson and Daviss, in collaboration with the Midwives Alliance of North America (MANA), the organization of American homebirth midwives, collected data on all homebirths attended by Certified Professional Midwives (CPMs, homebirth midwives, as distinct from CNMs, Certified Nurse Midwives) in the year 2000. Then the authors compared the outcomes for interventions and for neonatal deaths with a hospital group.

According to Johnson and Davis, when analyzing the different intervention rates of home and hospital:

We compared medical intervention rates for the planned home births with data from birth certificates for all 3 360 868 singleton, vertex births at 37 weeks or more gestation in the United States in 2000, as reported by the National Center for Health Statistics [Births: final data for 2000. National vital statistics reports. Martin JA, Hamilton BE, Ventura SJ, Menacker F, Park MM. Hyattsville, MD: National Center for Health Statistics, 2002;50(5)]

They used singleton, vertex births at 37+ weeks as a proxy for low risk women. They found, not surprisingly, that intervention rates are lower for homebirth. Then they turned to neonatal mortality rates. They should have compared the neonatal mortality rate of the homebirth group to the neonatal mortality rate of the hospital birth group, but they did not. Instead, they compared homebirth deaths to hospital births in a variety of out of date studies extending back more than 20 years.

The authors conclude:

Planned home birth for low risk women in North America using certified professional midwives was associated with lower rates of medical intervention but similar intrapartum and neonatal mortality to that of low risk hospital births in the United States.

But the authors never compared mortality rates to low risk hospital birth in 2000, because that would have led to a very different conclusion. Using the same dataset that Johnson and Daviss used, we find hospital neonatal death rate for white,  babies at 37+ weeks of 0.9/1000. This is not corrected for congenital anomalies, pre-existing medical conditions, pregnancy complications or multiple births. The neonatal mortality rate for white, singleton babies at 37+ weeks is 0,72/1000. The true rate is substantially lower. Nonetheless, we can make an important comparison. Johnson and Daviss reported a neonatal death rate at homebirth of 2.7/1000 (uncorrected for congenital anomalies, breech or twins). In other words, the neonatal death rate of CPM attended homebirths in 2000 was nearly triple the rate for low to moderate risk hospital births in 2000.

Simply put, the authors pulled a bait and switch. They claim to be comparing homebirth in 2000 with hospital birth in 2000. Indeed, they are comparing intervention rates for homebirth in 2000 with hospital birth in 2000, but when it comes to neonatal deaths, they used data extending back to 1969. It was the only way to make homebirth look safe by comparison.

Why might the authors deliberately intend to deceive readers? It turns out that Johnson and Daviss are not impartial researchers, though you would not know that from reading the paper. Johnson is the former Director of Research for the Midwives Alliance of North America (MANA) Statistics and Research Committee. Daviss, his wife, is a homebirth midwife. The paper does acknowledge that the study was funded by Foundation for the Advancement of Midwifery, a homebirth advocacy group.

Johnson and Daviss have created a website, Understanding Birth Better, to answer criticism. However, their explanation for the bait and switch is not merely disingenuous, it is an outright lie.

… Since our article was submitted for publication in 2004, the NIH has published analysis more closely comparable than was available at that time, and some have tried to use it as a comparison. While we still do not offer the comparison as a completely direct one, … it is the closest we have …

As they say in politics, it’s not the crime, but the cover up. Johnson and Daviss acknowledge that they used the wrong group for comparison with homebirth, but claiming that the correct data was not available at that time. That is flat out false.The relevant data was published in 2002, long before their paper was submitted (Infant Mortality Statistics from the 2000 Period Linked Birth/Infant Death Data Set, published August 29, 2002). Moreover, even before publication of the analysis, Johnson and Daviss had the raw data in their possession. They used that raw data from 2000 to calculate the rates of hospital interventions, so they were fully aware of the mortality data at all times.

It is difficult to imagine a legitimate reason why a professional statistician would deliberately use the wrong statistics for comparison when the right statistics were available and actually in his possession. It seems to me that the only possible explanation is that they knew all along that their study showed that homebirth has an increased risk of preventable neonatal death compared to hospital birth.

Regardless of reasoning or excuses, the bottom line is stark: rather than showing that homebirth with an American homebirth midwife is safe, the Johnson and Daviss study actually showed that homebirth with a CPM in 2000 had nearly triple the neonatal death rate of moderate to low risk hospital birth in 2000.

Posted in: Obstetrics & gynecology

Leave a Comment (120) ↓

120 thoughts on “A critique of the leading study of American homebirth

  1. mxh says:

    How did this pass the review process at BMJ? The editors of the journal are as much at fault for publishing these misleading conclusions.

  2. “How did this pass the review process at BMJ?”

    Although this is particularly surprising because it is so obvious, a lot of bad analysis gets by reviewers at leading medical journals. That’s why it is so important to read a scientific paper in full, not just the abstract. All too often, the data in the paper does not support the conclusion in the abstract.

    In addition, I wouldn’t be surprised if the reviewers were unaware of that Johnson was the former Director of Research for the Midwives Alliance of North America.He certainly does not disclose that information in the paper. They may have looked at it more carefully if they had known.

  3. Zoe237 says:

    So is the British Medical Journal itself unreliable? Why have they not corrected the supposed mistake? No offense, but as a layperson, I can believe you, a non-practicing online blogger, or I can believe the editors at the BMJ aren’t as dumb as you say.

    I have read the study as well as crunched the numbers at the wonder database, and read the authors’ response to these allegations, and while I can believe that they could have used a better control group, no expert, matched cohort analysis of the hospital control group (if you use the CDC data) has been done. Any amateur picking at filling in the blanks in the wonder database is unreliable. At best, you can say that we still don’t know.

  4. windriven says:

    Comparing apples with oranges when more apples are available runs perilously close to scientific fraud. Science is supposed to be the search for the truth and it is a perversion to twist the truth to suit political or personal agendas. At the bottom line it fuels distrust of science and plays right into the hands of the worst woo-meisters.

    Behold for instance the damage done to the cause of cleaning up the atmosphere and oceans by those clowns at East Anglia. Their Nixonesque e-mails smacking of data-fudging, secrecy and intimidation calls into question the scientific integrity of the corpus of AGW research. Already important climate legislation in Australia has collapsed at least in part because of ‘Climategate.’

    A good starting point for bringing more rigor to the game would be publishing the names of peer reviewers. Another would be a good deal more transparency in availability of source data. Truth always flourishes in sunlight.

  5. windriven says:

    @Zoe237

    “At best, you can say that we still don’t know.”

    Close, but no cigar. At best we can say that the authors have not proved their hypothesis: that at-home deliveries among low risk mothers offer similar neonatal mortality to hospital deliveries.

    The burden of proof is on those who contend that home birth is as safe as hospital birth.

  6. edgar says:

    Why are your comments not in BMJ’s rapid response?

  7. Basiorana says:

    winddriven– Is it? I mean, yes, extraordinary claims require extraordinary evidence, but home birth safety is not an extraordinary claim. It was the original method, and hospital births for low-risk women were introduced as the norm without any scientific evidence to support them– hence, puerperal fever problems.

    Of course, now we can look at the data and say it is safer in the hospital, but that’s because obstetrics have realized that they never actually proved it and then sought to determine safety rates. However, the burden of proof was always on obstetrical hospitals to prove that their new, bold, and different idea– giving birth in a hospital ward to prepare for potential problems, and thus, having the mother travel to the doctor instead of vice-versa– was safer.

    I would say they have, but that just means that homebirth advocates are disproven, not that they were ever responsible for proving anything. That’s like saying that doctors are responsible for proving that colloidal silver doesn’t work– they aren’t, the makers of colloidal silver are responsible for proving it DOES.

  8. moderation says:

    Basiorana – lets take your view to the extreme. Based the arguement you are putting forward “modern” medicine should have to do a double blind, plecebo controlled study to show that drilling a hole in the skull to drain bad ethers or bleeding a patient is not as good a a blood culture and antibiotics … because they are the “newer” therapies and should have to “prove” themselves.

    In the case that something untoward and unexpected happens during a home delivery … can the midwife intubate, place an umbilical line, control maternal bleeding, etc? One can argue that homebirthing is more comfortable care, but not that it is a higher level of care, and appearently with three times the neonatal mortality rate, not the safer care.

    BTW – is it just me or does there seem to be a lower standard of admission for papers to the BMJ than other journals. First the vaccines and autism paper and now this one. I seem to remember during my training a rating system for accuracy and relevency of medical journals … anyone know what the name of that system is and where you can look it up?

  9. Zoe237 says:

    “Close, but no cigar. At best we can say that the authors have not proved their hypothesis: that at-home deliveries among low risk mothers offer similar neonatal mortality to hospital deliveries. ”

    Fair enough. My issue it with the claim that hospital birth is safer than homebirth. If it’s so easy to prove, why hasn’t it been yet, with low risk planned homebirth vs. low risk planned hospital birth in the US? It already has been proven safe in other countries (who have more intensive homebirth midwive training and better continuity of care between home and hospital), so it’s not a completely out there hypothesis.

    de Jonge A, van der Goes B, Ravelli A, Amelink-Verberg M, Mol B, Nijhuis J, Bennebroek Gravenhorst J, Buitendijk S, Perinatal mortality and morbidity in a nationwide cohort of 529688 low-risk planned home and hospital births BJOG 2009

    http://news.bbc.co.uk/2/hi/health/7998417.stm (over 500,000 births in this one)

    CMAJ. 2009 September; 181(6-7): 359–360.
    doi: 10.1503/cmaj.091240.PMCID: PMC2742151

    Copyright © 1995-2009, Canadian Medical Association
    The safety of home birth: Is the evidence good enough?
    Helen McLachlan, PhD and Della Forster, PhD

    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2742151/

  10. Todd W. says:

    @Basiorana

    Is [the burden of proof on those who contend that home birth is as safe as hospital birth]? I mean, yes, extraordinary claims require extraordinary evidence, but home birth safety is not an extraordinary claim. It was the original method, and hospital births for low-risk women were introduced as the norm without any scientific evidence to support them– hence, puerperal fever problems.

    Yes, it is. If someone comes along, claiming that home birth is as safe as or safer than hospital birth, then the onus is on them to support their claim with proper evidence. In the same manner, if someone comes along claiming that hospital birth is safer than home birth, then the burden would be on the claimant to provide evidence to support their contention.

    It is not about “extraordinary claims requiring extraordinary evidence”. It is about supporting one’s claims with evidence, period.

  11. apteryx says:

    Zoe237 and Basiorana, I’m surprised neither of you yet caught the spot where Dr. Amy says:

    “…we find hospital neonatal death rate for white, singleton, vertex babies at 37+ weeks of 0.9/1000. This is not corrected for congenital anomalies, pre-existing medical conditions, pregnancy complications or multiple births. The true rate is substantially lower. Nonetheless, we can make an important comparison. Johnson and Daviss reported a neonatal death rate at homebirth of 2.7/1000 (uncorrected for congenital anomalies, breech or twins). In other words, the neonatal death rate of CPM attended homebirths in 2000 was nearly triple the rate for low to moderate risk hospital births in 2000..”

    So the “true rate” for hospital births is lower than the observed rate, because many of the dead babies had problems or risk factors, and they shouldn’t count – but we must not likewise presume that the “true rate” for home births is “substantially lower” for the same reason. Worse, Dr. Amy claims that the hospital death rate is elevated by the lack of “correction” for risk factors including multiple births – but even a nonpracticing emdee knows very well that “singleton, vertex babies” means no multiple-birth babies were included in the raw fatality count, and the hospital death rate could only be increased by adding them in. Worst, any deaths that may have involved multiple births or breech births apparently would have been counted in the homebirth death rate, but are excluded from the “low-risk” hospital death rate. Speaking of comparing apples to oranges!

    As an aside, regarding the recent discussion on the superiority of C-sections, one of the continuing education cases on the cultist woo-meister website NEJM.org right now is advertised as follows:

    “Meet the CPC patient at NEJM.org
    “Three days after an elective cesarean section, a 35-year-old woman was transferred to Massachusetts General Hospital because of abdominal pain, fever and hypotension. The patient underwent 34 operations, and she remained in the ICU for 4 weeks and on the surgical service for 5 weeks. Learn more about her presentation and diagnosis by reading the case report, and watch the video showing her current functional status after bilateral mid-forearm and below-knee amputations. Access to this article and video is free for a limited time to non–subscribers.”

    To translate, that’s both hands and both feet gone, on a woman who was healthy going in. Yikes. I hope I’m not terribly eeeevil to suggest that it would have been much a much better outcome if her baby had suffered a pinched nerve or cracked skull or whatever complication of vaginal birth one cares to advertise.

  12. Science Mom says:

    @moderation, you are probably referring to impact factor and BMJ’s is quite good: http://resources.bmj.com/bmj/about-bmj
    If you are referring to the infamous Wakefield et al. MMR-autism paper, that was published in Lancet.

  13. Scott says:

    So the “true rate” for hospital births is lower than the observed rate, because many of the dead babies had problems or risk factors, and they shouldn’t count – but we must not likewise presume that the “true rate” for home births is “substantially lower” for the same reason.

    Note that the conditions not accounted for in the home birth number, plus more, are not accounted for in the hospital birth number. Thus, this lack of correction will increase the mortality rate for hospital births relative to home births, assuming rates of such conditions are comparable (reasonable to first order). So it’s a perfectly valid argument, as far as it goes.

  14. Harriet Hall says:

    “…we find hospital neonatal death rate for white, singleton, vertex babies at 37+ weeks of 0.9/1000. This is not corrected for congenital anomalies, pre-existing medical conditions, pregnancy complications or multiple births. .. Johnson and Daviss reported a neonatal death rate at homebirth of 2.7/1000 (uncorrected for congenital anomalies, breech or twins).”

    Singleton rules out multiple births. Vertex rules out breech. If breech and twin deliveries were included in the home birth statistics, they might contribute to an increased death rate, but I wouldn’t think it would account for a 3-fold difference.

    How many home births involve breech or twins? By definition, those are high risk pregnancies. Wouldn’t those usually be seen as risk factors contraindicating a home birth?

  15. windriven says:

    @ Basiorana

    I cannot imagine that you wouldn’t agree that the risk to ALL (as opposed to low risk) mothers and babies is lower in hospital deliveries than home deliveries? So now we are cleaving off a subset of all deliveries – low risk pregnancies. To me that leaves the burden of proof on the at home advocates.

    @Zoe237 please see above.

    @apteryx

    I went to the NEJM site and did not find the article you mentioned so I have no idea what the woman’s medical condition was prior to admittance for the CS. But in the largest sense, who cares? I’m sure I could comb the literature and find horror stories about women giving birth at home. It was once done that way all the time. Of course maternal and neonatal mortalities were rather higher then.

    What is your actual point?

  16. apteryx says:

    Scott, you are confused, and are making the argument backwards – not your fault, as Dr. Amy chose to present the issue in a confusing way.

    The presence of risk factors, such as breech or multiple births, increases the risk of neonatal death. Dr. Amy is comparing the all-homebirth death rate (including those that turn out to be breech or twins) with the death rate for a select portion of hospital births (Caucasian, single, head-down only, which were assumed to be largely low-risk). Exclusion of nonwhite, multiple, and breech babies from the hospital death rate does not INCREASE the apparent rate; it LOWERS it. You got the impression from Dr. Amy that the hospital rate is for all deliveries and they have failed to correct it for the fact that complicated deliveries disproportionately take place in hospitals. That’s not the case.

    If any of the deaths in the homebirth cohort were of breech or twin babies, then the homebirth death rate comparable to the given hospital rate (only of “white, singleton, vertex” home-born babies) would very likely be lower than 2.7. That does not mean that the hospital death rate (in recent years) would not be lower, given genuinely comparable groups of low-risk women; I would bet that it is. But we have not been given data from comparable groups, so we have no basis for calculating the magnitude of the difference.

  17. JerryM says:

    You’d think data about birth rates and death rates would be easy to collect and research in this digital age, yet it seems there are big gaps in the available data?

  18. apteryx says:

    Dr. Hall, yes, usually people go to the hospital if they know they’re having twins or breech (unless perhaps they’re trying to avoid a compulsory C-section), but one gets the impression that people don’t always know they’re going to. My point is not that that accounts for all of the difference – I’m sure it does not – simply that Dr. Amy is twisting the stats as hard as anyone on the opposite side.

    Windriven, this site tends to lack long-term historical perspective. There was a time when homebirth with a midwife was safer than birth in a conventional hospital, purely in terms of death rates – not to mention in terms of avoiding babies brain-damaged by forceps. Things have improved most greatly in hospital care, but have improved in both venues; the home birth death rate now is lower than the hospital death rate of a few decades ago, which was considered pretty darn good at the time.

    Here’s a link to the NEJM article about the formerly healthy woman (also has a link to a video if you have the stomach for it):

    http://content.nejm.org/cgi/content/full/361/17/1689?ijkey=zD.g4JeyHjUH6&keytype=ref&siteid=nejm&emp=marcom

    My “actual point” there, Windriven, is a point that I have made before to Dr. Amy. Relative death rates for infants of people in various risk categories are a scientific issue; they are something that could in theory be carefully observed and statistically analyzed to draw definitive conclusions, even if it has not yet been adequately done. However, the question of whether infant death rates should be the only factor in decision-making is NOT a scientific question; it is a cultural one. Your philosophical values, or Dr. Amy’s, are not privileged over anyone else’s.

    If it turned out to be the case that one QALY-adjusted infant life was saved by hospital birth for low-risk women at the cost of extra suffering for a thousand women, major surgery for a couple hundred of them, long-term disability for an unlucky few who suffer severe complications, dozens of extra cases of asthma and allergies in the C-sectioned babies, and ten million dollars’ worth of health care costs that might have saved many more statistical years of life if spent in other ways – well, Dr. Amy would still say that was worth it to save that one baby. Personally, I would say it was not. Unless advocates of universal hospital birth (especially with MDs rather than nurse midwives) also address the non-fatal costs and risks of home and OBGYN-hospital birth, with an analysis that is robust to very different valuations of the various types of harm that can accure, you will not make much impression on those who are not already disposed to agree with you.

  19. Basiorana:

    “It was the original method, and hospital births for low-risk women were introduced as the norm without any scientific evidence to support them– hence, puerperal fever problems.”

    That’s what’s known as the naturalistic fallacy: it happened in nature, therefore it must be good. The fact is that childbirth is intrinsically dangerous. We know from epidemiological data that childbirth has a maternal death rate in nature of approximately 1% and a neonatal death rate as high as 7%. With hospital based modern obstetrics, the maternal death rate has dropped by 99% and the neonatal death rate has dropped by 90%. I think that issue has been settled definitely.

    No doubt homebirth with a CPM can be as safe as birth in nature (which is not safe at all). The issue is whether it is as safe as hospital birth, and it’s not.

  20. Harriet Hall says:

    The NEJM case report was of a woman who had had cephalopelvic disproportion in a previous pregnancy and gestational diabetes in this otherwise uncomplicated pregnancy. The C-section was characterized as “elective” with no further explanation. All went well until she developed symptoms of a Group A Strep postoperative infection which progressed to necrotizing fasciitis, toxic shock syndrome and complications of shock including bowel necrosis and gangrene of the hands and feet. At one point she had a cardiac arrest and was resuscitated.

    Her infection was probably not a hospital-related infection: most commonly strep infects the uterus as a result of vaginal colonization with strep. She received prophylactic antibiotics for the C-section and was promptly treated with vancomycin when the postoperative complications appeared. It wasn’t enough.

    This is a rare and devastating case, but it points out the advisability of only doing C-sections when there is a good indication.

  21. apteryx:

    “…we find hospital neonatal death rate for white, singleton, vertex babies at 37+ weeks of 0.9/1000. This is not corrected for congenital anomalies, pre-existing medical conditions, pregnancy complications or multiple births. The true rate is substantially lower. Nonetheless, we can make an important comparison. Johnson and Daviss reported a neonatal death rate at homebirth of 2.7/1000 (uncorrected for congenital anomalies, breech or twins).”

    Good pick up. I find it very hard to proof read my own writing. It was supposed to say:

    ……we find hospital neonatal death rate for white, babies at 37+ weeks of 0.9/1000. This is not corrected for congenital anomalies, pre-existing medical conditions, pregnancy complications or multiple births. The true rate is substantially lower. Nonetheless, we can make an important comparison. Johnson and Daviss reported a neonatal death rate at homebirth of 2.7/1000 (uncorrected for congenital anomalies, breech or twins).

    The neonatal death rate for white, singleton babies at 37+ weeks is 0.72/1000.

    I will edit the post.

  22. Johnson and Daviss have already publicly acknowledged on their website Understanding Birth Better:

    “Thus a crude comparison of the comparable rates for non-Hispanic white >37 week babies in hospital in the year 2000 would be about 0.91 neonatal deaths/1000 live births …”

    They offer no disagreement that they compared homebirth in 2000 with the wrong group. They offer no disagreement that the correct comparison for hospital neonatal mortality is 0.9/1000.

  23. apteryx:

    “Here’s a link to the NEJM article about the formerly healthy woman”

    So what? You have to demonstrate that both she and her baby would have been healthy had she given birth at home. Given her history of CPD and gestational diabetes, it is entirely possible that both she and her baby would have died without a C-section.

    Moreover, no one has claimed that bad things don’t happen in hospitals. The claim is that even when you INCLUDE the bad things that happen in hospitals, homebirth has a higher mortality rate.

  24. windriven says:

    @apteryx

    My interest is purely scientific. As a matter of fact I agree with you to the point that adults should be able to make their own decisions about health care and most everything else regardless of scientific evidence. I am less committed to that libertarian ideal when children are involved and, if the scientific data suggests better neonatal outcomes either at home or in the hospital, I would be inclined to argue in favor of that venue.

  25. windriven says:

    @apteryx

    Having carefully reviewed the NEJM case study I wonder why you chose this case. The woman had an earlier pregnancy that required vacuum-assisted delivery. The choice of CS seems entirely reasonable.

    I don’t know if the subsequent infection was likely iatrogenic or not. That isn’t clear in the case study.

    I would love to see The Pus Whisperer weigh in on this case (though it is really far afield from the post we’re discussing).

  26. apteryx says:

    Dr. Amy —

    Tee hee! No, I certainly don’t “have to demonstrate that both she and her baby would have been healthy had she given birth at home.” As any competent observer would agree, it would be utterly impossible to prove any such thing in any specific instance. I cannot demand that you PROVE that any particular infant who died at a homebirth would have not only lived, but been healthy, in a hospital. Of course you could prove no such thing – it might not be probable that it would have died in the hospital, but it’s “entirely possible.”

    But the utter ridiculousness of that challenge – which IMHO betrays a very unscientific view indeed – isn’t the only ridiculous aspect of your message above. Three more ridiculous points:

    1. You present the alternatives as being only home birth or C-section, as if I had argued that woman should have given birth at home. Wrong. You can reread the message if you need to. This was in response to your older thread in which I saw that you have been touting the miracle of surgical birth. I just noted that if she had given birth, in the hospital, through her vagina, she’d still have hands and feet and a lot more of her gut.

    2. You imply that one cannot wish a course had been followed that would have spared this woman severe disability unless that course could have provided the infant with a guarantee of survival and perfect health. Again, we don’t “have to” think anything of the sort. Minor harm to an infant, for example, may not be seen by everyone as more tragic than major harm to a woman. Indeed, some would argue that the cost when a functioning adult woman is crippled outweighs even the death of an infant. Feel emotional about that suggestion? You’re free to, but morality and science are two different things.

    3. In all probability, neither the baby nor the mother would have died from lack of a C-section. That’s what “elective” means. You can also read the case report. Having previously had a large-headed baby does not mean that your current baby will magically get stuck; and if that had been the case, it would have been listed in the case report as a clinical indication for C-section. You don’t think that after what happened to this woman, the MDs would have loved to be able to give some medical purpose for a C-section so they could claim that “at least they saved the baby” or “her life”?

    With this as with everything else, the invasive procedures are considered to be superior precisely because they are invasive, and you’re telling the rest of the world that at the very best they have the burden of proving that those things are not superior. Whether it’s forceps, enemas, episiotomies, or elective C-sections, all these things were imposed on women without benefit of randomized controlled trials, and yet if a woman suggests that she might do without them until they are proven, she’s a heretic.

    Well, I’ve done my share for this thread. Good luck to whoever will have the task of pointing out future errors.

  27. apteryx:

    “In all probability, neither the baby nor the mother would have died from lack of a C-section. That’s what “elective” means.”

    No, that is not at all what “elective” means in this context. Elective merely means “non-emergent.” I pointed this out several days ago in regard to the Villar study about the effects of C-sections. A C-section done to save the life of both mother and baby because of cephalopelvic disproportion would be classified as “elective” even though it was life saving.

    “As any competent observer would agree, it would be utterly impossible to prove any such thing in any specific instance.”

    Right. That’s why it is not an example of the “dangers” of the hospital. Unless you can demonstrate the the woman would have been better off outside the hospital, it is pointless to even mention the case in this context.

  28. edgar says:

    Amy is guilty of what she accuses the authors day, misrepresentation of data. Were she REALLY interested in finding the ‘truth’ she would know that choosing white women only biases her presupposition in her favor. At the very least she should offer a range based on race/ethinicty

    If I were the authors’, I would have adjusted based on race, and I would have pulled out amish people (it is unclear if amish were included or not.)

    As for the wrong dataset, I have no opinion on this as of yet.

  29. windriven says:

    @ edgar

    What is the racial bias in this issue? Are there racial differences in maternal choice of birth venue, neonatal mortality? It wouldn’t have occurred to me to account for this. I just wonder what the rationale is. I’m also interested in why you would exclude Amish. Cultural choices would seem even less important that racial groups.

  30. The reason that the comparison group is restricted to white women is because race is a risk factor for neonatal mortality. Babies born to Black women have substantially higher rates of neonatal mortality than those born to other women. Almost all women who choose homebirth are white, so it makes sense to use white women as the comparison group.

  31. edgar says:

    ” They should have compared the neonatal mortality rate of the homebirth group to the neonatal mortality rate of the hospital birth group, but they did not. Instead, they compared homebirth deaths to hospital births in a variety of out of date studies extending back more than 20 years.”

    and

    “Indeed, they are comparing intervention rates for homebirth in 2000 with hospital birth in 2000, but when it comes to neonatal deaths, they used data extending back to 1969.”

    Amy, I think you have a responsibilty to point to exactly which reference you take issue with and why

  32. edgar says:

    If it were my study, I would exlude the Amish from the HB population they population is probably not generalizable to the greater US population. There is a higher rate of congenital malformations. Aside from that, their lifestyle is vastly different from the mainstream.
    That is just me though,I would try to make my study as generalizable as possible, another epidemiologist might have another opinion.
    I would exclude any such distinct, subpopulation that is as culturally isolated as this.

    Asw for the white women, Amy, it doesn’t makes sense, anymore that what the authors did. Again, if it were me, I would have adjusted for race in the paper. You should have at least listed the rates for each race, and stated that the mortality rate fell somewhere it that range, probably within the lower end of the range, since not all of the HB cohort were white.
    That would have been a scientific way of presenting your data.

    “What is the racial bias in this issue? Are there racial differences in maternal choice of birth venue, neonatal mortality”

    I do not know the answer to this in terms of HB’ers. Amy would say that are mostly white, and they are in the Daviss and Johnston paper, but as for the HB community as a whole, I don’t know.
    Neonatal mortality rates are higher for Aferican Americans (and most everyone who ain’t white). Whether this is a marker for poverty/access to care or has biological basis (or a mixture of the two) is debated and unknown. There is new research about chronic stress and health problems, so this may be a new area of research in the future.

  33. According to the authors of the paper:

    On interventions,

    “Individual rates of medical intervention for home births were consistently less than half those in hospital, whether compared with a relatively low risk group (singleton, vertex, 37 weeks or more gestation) that will have a small percentage of higher risk births or the general population having hospital births (table 3).”

    Table 3 includes notes on the control group:

    “Based on data from birth certificates for all 3 360 868 such births in United States in 2000. Data reported by National Center for Health Statistics.10 This subset of birthing women would generally be low risk, but would include a small percentage of higher risk women who would likely require more medical intervention.”

    On neonatal deaths:

    “The results for intrapartum and neonatal mortality are consistent with most North American studies of intended births out of hospital11-24 and low risk hospital births (table 4)”

    Here is Table 4. As you can see, none of the papers are contemporaneous and extend back as far as 1969.

    Look down at the bottom of Table 4. Note the author’s disclaimer:

    “Table is presented for general comparison only. Direct comparison of relative mortality between individual studies is ill advised. as many rates are unstable because of small numbers of deaths, study designs may differ (retrospective versus prospective, assessment and definition of low risk, etc.), the ability to capture and extract late neonatal mortality differs between studies, and significant differences may exist in populations studied with respect to factors such as socioeconomic status, distribution of parity, and risk screening criteria used.”

    The authors acknowledged at the time that direct comparison was “ill advised” and then staked the conclusion of their paper on a direct comparison.

    Finally, as I have pointed out repeatedly, Johnson and Daviss have since acknowledged (on their website Understanding Birth Better, in December 2007):

    “Thus a crude comparison of the comparable rates for non-Hispanic white >37 week babies in hospital in the year 2000 would be about 0.91 neonatal deaths/1000 live births”

    They attempt to excuse the use of the out of date papers by claiming:

    “Since our article was submitted for publication in 2004, the NIH has published analysis more closely comparable than was available at that time”

    However, the relevant data was published in 2002, long before their paper was submitted (Infant Mortality Statistics from the 2000 Period Linked Birth/Infant Death Data Set, published August 29, 2002). Moreover, even before publication of the analysis, Johnson and Daviss had the raw data in their possession. They used that raw data from 2000 to calculate the rates of hospital interventions, so they were fully aware of the mortality data at all times.

    I can’t imagine what more proof you need. The authors pulled a bait and switch, have acknowledged the existence of more accurate figures (the same ones that I calculated) and gave an excuse for not using them which appears to be a complete lie.

  34. edgar says:

    “Almost all women who choose homebirth are white, so it makes sense to use white women as the comparison group.”

    That. Right. there. Amy,
    is what I am speaking about. Your manner of writing and sweeping generalzations make me wonder what your motive is, because I question if it is truly in the interest of science.
    No, 89% were white. To me that is NOT al most all.
    A scientist would have stated that precisely. It also does not make sense to use the white group without a qualifier, as you did not do.

  35. edgar says:

    Raw data? what raw data? Are you speaking of Births: Priliminary Data for 2002?

    That is what I am asking, instead of saying ‘raw data’, state which source they should use.

    WHy don’t you post the link to their website where you got those quotes so people can read for themselves.

  36. “No, 89% were white. To me that is NOT al most all.”

    10% of the women were Hispanic or Asian. They have neonatal mortality rates that are the same or lower than white women. If you’d like, we can adjust the hospital neonatal deathrate downward by factoring correcting for those women.

  37. edgar says:

    We can do that, I am just stated what I would have done, if it were my study. It should have been done, IMO. ALong with the exclusion of the Amish.

    I disagree with your assertation about the direct cxomparsion, I totally agree with the authors, compairsons BETWEEN studies should be dome very cautiously.
    A comparison among their results and NCHS is valid.

  38. They’ve changed the format of their response since they originally published it on their website. It now exists as a downloadable pdf formatted to look like a published scientific paper.

    http://understandingbirthbetter.com/section.php?ID=31&Lang=En&Nav=Section

    I have critiqued it in my own pdf. Thanks to the wonders of Adobe Acrobat, I added comments on the original.

    http://www.homebirthdebate.com/BMJCPM2000_corrected.pdf

  39. edgar says:

    Wrong
    4% were hispanic.
    1.3% were African-American
    2.6% were other

    No mention of Asian at all

  40. edgar says:

    And I ask again why you haven’t put your response on the BMJ website?

  41. edgar says:


    … Since our article was submitted for publication in 2004, the NIH has published analysis more closely comparable than was available at that time, and some have tried to use it as a comparison. While we still do not offer the comparison as a completely direct one, … it is the closest we have …”

    I cannot find this quotation anywhere on their website, Amy, shame on you

  42. “why you haven’t put your response on the BMJ website?”

    They only publish responses submitted in the first few weeks after the paper is released, hence the designation “rapid responses.”

  43. Didn’t I just write that they have changed their website since then?

    I went into the Internet Archive to find the cached version of the page from January 2008.

    http://web.archive.org/web/20080115164850/http://understandingbirthbetter.com/section.php?ID=31&Lang=En&Nav=Section

    You can find the relevant reference in the 4th full paragraph from the top.

  44. edgar says:

    Amy, I really don’t care if you don’t like it, the fact of the matter is you can’t have quotations like that without citations from where you got them, so that other can go back and look at them. If they are no longer available, then you can’t use them, as you have no proof.

    That IS the scientific way.

  45. Zoe237 says:

    Amy Tuteur, MD:

    “The reason that the comparison group is restricted to white women is because race is a risk factor for neonatal mortality. Babies born to Black women have substantially higher rates of neonatal mortality than those born to other women. Almost all women who choose homebirth are white, so it makes sense to use white women as the comparison group.”

    As a total aside, any theories on why this might be (that race is a risk factor)? I’ve read that it is true even when you adjust for socio-economic status. The only two possible ideas I can surmise are a biological basis (but what?) and institutional racism in the maternity wards of hospitals (but what?). But those are complete guesses. Why is the neonatal mortality rate for African American babies many times higher than for white babies born at home?

  46. windriven says:

    “Were she REALLY interested in finding the ‘truth’ she would know that choosing white women only biases her presupposition in her favor.”

    Maybe I’m stupid but I don’t get this at all. If only white women are compared, and if white women are over represented in home births, and if non-white women have higher infant mortality rates and are more likely to give birth in a hospital, how is using white women biasing this in Dr. Tuteur’s favor? It seems quite the opposite.

  47. edgar says:

    No, she chose the white group which has the lowest neonatal mortality of all groups.
    Thus the rate for the hospital group is the lowest it could possible be.

    Those in the HB group were not all white, but were disproportionatley so. Therefore we can resonable surmise that those who were not-whote may have had a higher than average rate of neonatal death, and comparing that group as a whole to the white hospital group.

    Now allowing for the factg that we do not have the raw data on which to perform calulcations, an accurate way to do this would be to quite the rates of whites and all others, and allow that the rate probably falls closer to the whote group.

  48. edgar says:

    Zoe,
    As I mentioned, there is promising research out there that indicates that non-white people in the US suffer chronic stress from racism poverty, etc, and that could potentially be the reason for many poor health outcomes.
    It is interesting, and the jury is still out, but it makes sense. It is funny, I was at APHA, and i mentioned this to an African American Doc, and he smiled and said “we knew this all along, the old folks called it being ‘worried to death’”

  49. edgar:

    “she chose the white group which has the lowest neonatal mortality of all groups.”

    No, the white group does not have the lowest neonatal mortality. There are several minority groups that actually have lower neonatal mortality.

    I don’t really understand why you are disagreeing. The authors themselves have already capitulated. Why are you still arguing to defend what they themselves have stopped trying to defend?

  50. windriven says:

    “Those in the HB group were not all white, but were disproportionatley so. Therefore we can resonable surmise that those who were not-whote may have had a higher than average rate of neonatal death, and comparing that group as a whole to the white hospital group.”

    Edgar, with all due respect, this doesn’t hold up. By your own admission the HB group was disproportionately white. So even if the non-white HB mothers had a higher rate of infant mortality, that rate would have to be astronomically higher to generate a total HB mortality rate 3 times that of the all white hospital group.

  51. edgar says:

    You are right, it would. But we don’t know what it is, do we? Maybe it is atronomically hgher. Probably not.

    My argument is with Amy’s lack of precision in order to further her own belief. If this is really Science based medicine, lets be as precsise as possible. We should be trying to figure out as accurately as possible what the ‘true rate’ is.

    Amy if you read the author remakrs in their entirety you will see that they stand by their analysis. So apparently does BMJ.

  52. edgar says:

    “she chose the white group which has the lowest neonatal mortality of all groups.”

    No, the white group does not have the lowest neonatal mortality. There are several minority groups that actually have lower neonatal mortality

    Yes, you are correct, there are racial groups with lower mortality rates. I would question racial misclassification, though (not that it is relevant to this topic), American Indians/Alaska Natives are notoriously misclassified.

  53. edgar says:

    “she chose the white group which has the lowest neonatal mortality of all groups.”

    No, the white group does not have the lowest neonatal mortality. There are several minority groups that actually have lower neonatal mortality

    Yes, you are correct, there are racial groups with lower mortality rates. I would question racial misclassification, though (not that it is relevant to this topic), American Indians/Alaska Natives are notoriously misclassified.
    test

  54. edgar says:

    The rate of 3x is Amy’s assertation. One that I take with caution due to her lack of precision and general disrespect for scientific process.

  55. “if you read the author remakrs in their entirety you will see that they stand by their analysis.”

    No, they don’t stand by their analysis. In light of their admission that the hospital neonatal mortality rate was 0.9/1000 in 2000 they have decided to “re do” their analysis and (surprise!) they now claim that their original analysis was wrong and that the neonatal mortality rate in the homebirth group was actually 0.9/1000. Quite a coincidence, don’t you think?

    They presented a poster at the American Public Health Association meeting fall 2008 (not published, not peer reviewed, no data or analysis). They claimed:

    “We compared the neonatal mortality rate among 5,418 planned homebirths attended by Certified Professional Midwives in the year 2000 (CPM2000 study) to the U.S. National Institutes of Health (NIH) neonatal mortality rate for births in hospital to U.S. non-Hispanic white women of 37 weeks plus gestation. Prematurity rates were also examined for the two populations.

    Adjustments were made to ensure that the comparisons were as close as possible to comparing like with like. This included removal from the CPM2000 study death rate of intrapartum mortality, 3 deaths involving lethal birth defects unlikely to have been carried to term in the hospital population, and 1 death and 286 births among African-American and Hispanic women. After making the necessary adjustments that were possible, the neonatal death rate in both datasets was just under 1 death per 1000. …

    Our conclusions remain unchanged from those in the original article — the neonatal mortality rate for low risk women in North America using Certified Professional Midwives is similar to that for low risk women in hospital in the U.S., and the intervention rates are much lower.”

    Once again they removed lethal congenital anomalies from the homebirth group, but not the hospital group. we can apply the same adjustments that Johnson and Daviss applied. According to the 2000 dataset on CDC Wonder, in the group of white women, 37+ weeks, 2500+ gm, with singletons who delivered in the hospital in 2000, we find that there were 1863 deaths, of which 1001 were due to lethal congenital anomalies. That means that the neonatal death rate for hospital birth in 2000 was 0.34/1000 after we performed the EXACT SAME adjustment that Johnson and Daviss performed on the homebirth data. Now that the groups are once again comparable, the neonatal mortality rate for homebirth in 2000 is STILL almost TRIPLE the neonatal death rate for hospital birth in 2000.

    What’s their excuse for removing lethal congenital anomalies from the homebirth group, but not the hospital group? They claim that lethal congenital anomalies would not be carried to term in the hospital. That’s an unsubstantiated claim and a ridiculous one. As I just showed, more than 50% of the deaths in the hospital group were from congenital anomalies.

    The bottom line is that Johnson and Daviss can “adjust” the data to their hearts’ content, but those “adjustments” must also be applied to the hospital data. When both data sets are treated the same way, the conclusions remain the same. Homebirth has an increased rate of neonatal death almost triple that of hospital birth for low risk women.

  56. windriven says:

    While readily admitting that I am not an expert in obstetrics, Dr. Tuteur’s arguments seem perfectly defensible to me as, frankly, do some of yours. But I am troubled by the ad hominem nature of some of your … arguments. Dr. Tuteur has raised what appear to the eyes of a physicist to be legitimate concerns about the BMJ data. You have raised seemingly valid arguments of your own. It seems to me utterly righteous to question data and methodology closely. But ascribing ‘general disrespect for scientific process’ to Dr. Tuteur strikes me as out of line.

    This seems to be an argument that revolves around the fact that there is no pristine and solid data set available. I don’t imagine that to be your fault or Dr. Tuteur’s either.

    It seems to me that Dr. Tuteur has the high ground with physician-attended hospital deliveries being the gold standard. It is incumbent on the HB advocates to demonstrate that HB is a superior – or at least equivalent – venue for childbirth. Moreover, it seems to me that subtracting all of the risk factors out in an effort to balance HB with physician-attended hospital deliveries is in and of itself a form a cherry picking. ‘If you’re a healthy, blond-haired, blue-eyed, fair-skinned lass with wide-body hips and a compliant head-down fetus, home birth is just as good as a hospital delivery.’ Umm … so what? It seems to me that SBM obstetrics is about minimizing maternal and infant mortality and morbidity regardless of ethnicity, risk factors, or other variables. If the HB crowd wants to slice off a carefully selected subset of the gravid population and proclaim their venue superior, don’t you think it is incumbent on them to provide a bullet-proof data set to substantiate their claim?

  57. Zoe237 says:

    Here are reports of two more hospital deaths (Yes, these cases are anecdotal, no, they don’t prove anything, but yes, they raise questions):

    “Valerie Scythes, 35, and Melissa Farah, 28, were friends who taught at the same grade school in Avon, N.J., and died two weeks apart last spring after delivering their firstborns, both healthy girls, via planned C-sections at Underwood Memorial Hospital in Woodbury, N.J. Whether Scythes and Farah died solely as a result of their C-sections isn’t clear. John Baldante, the Scytheses’ attorney, and Todd Miller, the Farahs’ attorney, did not return phone calls. The Philadelphia Inquirer reported that Farah’s death certificate said she died of “shock due to bleeding and anemia.”"

    http://www.usatoday.com/news/health/2008-01-07-csections_N.htm

  58. edgar says:

    Oh yes, I ABSOLUTELY do, and they absolutely should. And I hope I see more of it in the future. And let the data fall where it may. Amy has been less tha n precise in her assertations, and too sweeping in her anaylses for me to take seriously. SHe also seems to cast doubt on things, which to the lay person, may seem ominous. The poster session at APHA comment for instance. It is a venue to present work, nothing more, nothing less. yuet her (not peer reviewed, not published) sounds very wooish to the unitiated, when in fact is is not.
    Just professionals presenting work to other professionals.

    And the assertation they they used the dataset that they said they didn’t have is misleading, too.
    “The relevant data was published in 2002, long before their paper was submitted (Infant Mortality Statistics from the 2000 Period Linked Birth/Infant Death Data Set, published August 29, 2002). Moreover, even before publication of the analysis, Johnson and Daviss had the raw data in their possession. They used that raw data from 2000 to calculate the rates of hospital interventions, so they were fully aware of the mortality data at all times.”

    I do not know if they had Infant Mortality Statistics from the 2000 Period Linked Birth/Infant Death Data Set, published August 29, 2002 or not as Amy says.

    Their citation was for Births: Final data from 2000.
    The Linked Birth Infant Death was not in their citations, so Amy’s allegation that it was ‘in their possession’ seems to be a bit strong. How does she know what they possess and what they didn’t?

    But back to the paper, as I mentioned if it were me, I would have adjusted my rates to mirror societal makeup. I.E. adjusted my numbers so that blacks would make a proportional respresentation of the cohort. Also, the folks in the study tended to be I see what you are saying about slicing and dicing the data to make a narrow group. But that’s a dicy proposition once you move past the data piece of it…Categorizing black as ‘high risk’ simply because they are black. They also show that the HB group were of Low SES and older. I would have adjusted for those things too. And all of the differences in the maternal charecteristics.

    As for a good data set, I am leaning more and more toward federal birth and death certificates run by medicaid billing. One error and they are sent back for correction!
    I analyze death certs, and the amount of people that die from ‘respiratory failure’ or ‘heart failure’ astounds me. I hope its only saving grace is that it has been wrong consistantly.

  59. Windriven
    “eyes of a physicist ”

    Did your spell checker mis-correct that from a typoed “physician” or was that indented to be “physicist”?

  60. windriven says:

    @ Karl

    Physicist. I am not a physician. That sometimes leads me into deep water here.

  61. IndianaFran says:

    You know, for a post which criticizes other researchers for alleged intellectual dishonesty and misrepresentation, you would think that there would be a genuine effort to avoid precisely this behavior.

    Yet Amy says:
    ++++++++++++++++++++++
    Johnson and Daviss reported a neonatal death rate at homebirth of 2.7/1000
    ++++++++++++++++++++++
    which is an absolute flat-out falsity.
    There is no such figure anywhere in the the original paper or the author’s later analysis. The only way this “statistic” can be presented is by adding back in the 5 intrapartum deaths reported, and incorrectly calling the combined intrapartum plus neonatal death rate the “neonatal” death rate.
    It is simply, plainly, statistically wrong to compare this with a hospital rate based on linked birth/death certificates, which includes only those deaths that occurred after live births.

    This is not an oversight or a typographical error. Based on previous writings, Amy apparently feels entitled to do this because *one* of the intrapartum deaths *might* have been accidentally miscoded with an Apgar score of 1/0, or *might* have been mistakenly coded as an intrapartum death rather than a neonatal death, so she decided that *all* of the reported intrapartum deaths are questionable, and that she can lump them all together with the neonatal deaths, and still call this the “neonatal” death rate associated with homebirth.

    Depending what hospital data you are using for a comparison, you can argue whether the correct number of deaths to include in the homebirth study should be 10 (all the neonatal deaths plus the one questionable intrapartum death), or 9 (all the neonatal deaths reported) or 8 (all the neonatal deaths among white women if you are comparing to an all-white hospital group) or 6 (all the neonatal deaths excluding congenital anomalies) or 5 (all white neonatal deaths excluding congenital anomalies), yielding a neonatal death rate somewhere between .9 and 1.8 per thousand. But you simply cannot use 14 as the numerator in this calculation, because you are not providing a comparable hospital death rate that includes intrapartum deaths. Claiming that the study “reported a neonatal death rate of 2.7 per thousand” is an intentional misrepresentation.

    It’s really difficult to argue on the one hand that Johnson and Daviss are stretching the truth in order to make home birth look as good as possible, while on the other hand you are stretching the truth in the opposite direction in order to make home birth look as bad as possible.

  62. IndianaFran says:

    Amy says:
    ++++++++++++
    According to the 2000 dataset on CDC Wonder, in the group of white women, 37+ weeks, 2500+ gm, with singletons who delivered in the hospital in 2000, we find that there were 1863 deaths, of which 1001 were due to lethal congenital anomalies.
    ++++++++++++

    I cannot duplicate these results from the CDC wonder dataset. Can you confirm exactly what inclusion criteria you used?

  63. “The only way this “statistic” can be presented is by adding back in the 5 intrapartum deaths reported”

    Yes, the intrapartum deaths. Let’s address that issue. But before we do let’s revisit what we have established so far:

    1. Johnson and Daviss knowingly and deliberately compared the neonatal death rate for homebirth in 2000 with a bunch of out of date papers extending back to 1969 instead of with hospital birth of low risk women in 2000.

    2. When this was publicly pointed out, J&D claimed that the data for 2000 was not available when they submitted their paper in 2004. That is false since it was published in 2002 and, in any event, was part of the raw data that was in their hands all along.

    3. J&D removed congenital anomalies from the homebirth group, but not from the hospital group.

    Now let’s turn to the so called intrapartum deaths. Intrapartum death has a very specific meaning. It can only be called an intrapartum death if the baby shows absolutely no sign of life at the time of birth. Even one pulsation of the umbilical cord makes it a neonatal death, not an intrapartum death. In other words, the baby must drop into the attendant’s hands absolutely dead.

    J&D make it clear that they used some other definition of intrapartum death when they chose to put a baby with low Apgars into the intrapartum death group. Since they were clearly not following the accepted definition of intrapartum death, I chose not to classify those deaths as intrapartum.

    Let’s assume for a moment that those 5 (or 4) deaths really were intrapartum deaths. J&D would like you to think that babies who die DURING a homebirth should just be discarded, not counted at all, as if it had nothing to do with homebirth. That, of course, is simply absurd. Those deaths MUST be counted.

    So if they want to call those deaths intrapartum deaths, they MUST compare the intrapartum death rate at homebirth with the intrapartum death rate at hospital birth.

    Intrapartum death in a hospital setting is quite rare. During my entire career I never saw a single baby drop dead into the attendant’s hands. That’s because of intrapartum fetal monitoring. We know what the baby’s heart rate is all along.

    In published studies, the intrapartum death rate for all hospital births, of all gestations, and all manner of complications is approximately 0.03/1000 [EDITED: this is an error. The intrapartum death rate is 0.03% which is 0.3/1000]. The intrapartum death rate for term babies without risk factors is undoubtedly much lower Now let’s calculate the intrapartum death rate for homebirth.

    According to J&D, there were 5 intrapartum deaths out of 5418 births for a truly astounding intrapartum death rate of 0.9/1000, fully 30 times [EDITED: This is an error. The correct number is 3 times] higher than the intrapartum death rate in the hospital!

    So, classify the intrapartum deaths separately if you want, but I’m sure you’ll agree that you can’t simply ignore them. They must be accounted for somewhere. If J&D prefer to classify them as intrapartum deaths, they MUST compare the death rate to the hospital equivalent.

    Therefore, we could say that homebirth has a startlingly high rate of dead babies dropping into the hands of CPMs who were completely unaware that the baby had already died.

  64. IndianaFran says:

    “In published studies, the intrapartum death rate for all hospital births, of all gestations, and all manner of complications is approximately 0.03/1000.

    I’m sure that you can provide the references that support this?

  65. IndianaFran says:

    Amy says:
    “J&D would like you to think that babies who die DURING a homebirth should just be discarded, not counted at all, as if it had nothing to do with homebirth. ”

    Again, you are misrepresenting the author’s intent. J&D DID indeed count these births, they reported a rate for “intrapartum plus neonatal death” and they attempted to find, in the published literature, other studies that reported a combined rate. “Table 4 Combined intrapartum and neonatal mortality in studies of planned out of hospital births or low risk hospital births in North America (at least 500 births)”.

    The CDC data, which is based on linked birth/death certificate, cannot be used to calculate a comparable combined rate. I can’t read minds (at least not as well as you), so I can’t say for sure, but maybe that’s why the authors chose not to use the CDC data in the original paper – because using that comparison would force them to “ignore” the intrapartum deaths.

    When the later analysis turned to CDC based datasets, then in order to do an apples to apples comparison, you must include only the neonatal deaths.

    NO ONE is asking anyone to ignore intrapartum deaths. But if you are going to use a combined rate, then you need to find an appropriate source for a hospital comparison with combined rates. If you want to evaluate neonatal and intrapartum deaths separately, then separate comparison groups are needed. But you can’t just claim than you have a better knowledge than the original authors, of how to properly classify the intrapartum deaths. And you really shouldn’t state as fact that “Johnson and Daviss reported a neonatal death rate at homebirth of 2.7/1000″ when that is simply not the case.

  66. Zoe237 says:

    “So, classify the intrapartum deaths separately if you want, but I’m sure you’ll agree that you can’t simply ignore them. They must be accounted for somewhere. If J&D prefer to classify them as intrapartum deaths, they MUST compare the death rate to the hospital equivalent.”

    That’s not what you were doing in your “3x as many deaths” assertion. You were comparing your own doctored numbers in the homebirth group, intrapartum+neonatal, with the CDC dataset, which you know darn well is only *neonatal.* Sigh. This is why you need to publish and be peer-reviewed to be taken seriously by anybody.

    Perinatal mortality would be a much better measure, but this is difficult to study in this case.

  67. “I’m sure that you can provide the references that support this?”

    Here’s the reference, but I must apologize for my math mistake, the overall rate of intrapartum stillbirth in the study is 0.03% which is the same as 0.3/1000, not 0.03 per thousand.

    That means that the increased rate of intrapartum death at homebirth was “only” triple that of term hospital birth of all risk levels.

  68. IndianaFran:

    “NO ONE is asking anyone to ignore intrapartum deaths.”

    They had 14 deaths in 5418 homebirths for a rate of 2.6/1000.

    First they excluded congenital anomalies, but they’re not allowed to do that. Then they failed to appropriately define intrapartum deaths; so we don’t really know whether the intrapartum deaths are truly intrapartum deaths. I chose to classify them all as neonatal deaths.

    But suppose you classify them as intrapartum deaths. Then you have an obligation to compare the intrapartum deaths at home and in the hospital. J&D elided that comparison by claiming that no such statistics exist, but that’s not true They then tried to compare the homebirth deaths to a bunch of out of date studies.

    Five (or four) intrapartum deaths among 5418 births is a very high number. Most clinicians will never see an intrapartum death in their entire career.

  69. edgar says:

    Amy, I asked you to clarify these points above, and you have not done so:

    1. Johnson and Daviss knowingly and deliberately compared the neonatal death rate for homebirth in 2000 with a bunch of out of date papers extending back to 1969 instead of with hospital birth of low risk women in 2000.

    2. When this was publicly pointed out, J&D claimed that the data for 2000 was not available when they submitted their paper in 2004. That is false since it was published in 2002 and, in any event, was part of the raw data that was in their hands all along.

    I reat:

    I do not know if they had Infant Mortality Statistics from the 2000 Period Linked Birth/Infant Death Data Set, published August 29, 2002 or not as Amy says.

    Their citation was for Births: Final data from 2000.
    The Linked Birth Infant Death was not in their citations, so Amy’s allegation that it was ‘in their possession’ seems to be a bit strong. How does she know what they possess and what they didn’t?

  70. IndianaFran says:

    “Here’s the reference”

    ??

  71. “Their citation was for Births: Final data from 2000.
    The Linked Birth Infant Death was not in their citations, so Amy’s allegation that it was ‘in their possession’ seems to be a bit strong. How does she know what they possess and what they didn’t?”

    Technically you are correct, but I’m not sure why that matters. Even though both the birth data and the linked birth-infant death data were both published in 2002, it is theoretically possible that they only read the former and never saw a copy of the latter. I doubt it, but, as you point out, it is possible. Even though the linked birth infant death data is regularly published approximately 18 months after the year in question, they could have been entirely ignorant of it.

  72. IndianaFran says:

    Again and again, you are asserting facts about the original study and its authors which are not true:

    “J&D elided that comparison by claiming that no such statistics exist,”

    Where do you find any such claim?

  73. IndianaFran says:

    “Intrapartum mortality data is here:

    Life-table analysis of the risk of perinatal death at term and post term in singleton pregnancies”

    for those of us mortals who do not have full-text access to this source, could you kindly paste the pertinent part of the results?

  74. “Where do you find any such claim?”

    Isn’t it implied by not bothering to mention or use them?

    Let me ask you: Do you wish to claim that Johnson and Davis were honest in their presentation of the data?

    Do you wish to claim that Johnson and Davis did not know the hospital neonatal mortality rate for 2000 when they submitted their paper for publication in 2004?

    Do you wish to claim their assertion that it was unavailable to them is true?

    Do you wish to claim that it was appropriate to remove congenital anomalies from the homebirth group and not the hospital group?

    Do you wish to claim that their comparison to out of date papers extending back to 1969 was appropriate?

  75. edgar says:

    Amy,
    Thank you. You have asserted over and over that they ‘had the raw data into their possession.’ And that they ‘used that same raw data to calculate intervention rates.”

    They ARE TWO DIFFERENT PUBLICATIONS! Yet you do not acknowledge this, you make it sound as if these publications are one in the same. At no time did they ever have raw data. They had a paper that summarized NCHS birth stats. Published 2-12-02.

    The Linked Infant Birth Death for 2000 was published 8-28-08.

    Now, I do not know the timeline the authors were on for submission, peer review, editing, publication. You your assertion that they deliberately used favorable data and did not use more recent data is entirely possible, I do not know. And neither do you. Your allegations are stunning. I would have no problem with you raising the question as to why this is,but I strongly object to you acting as judge and jury, when you yourself do not have the facts, are are guilty of your own particular subterfuge.

  76. Harriet Hall says:

    This is getting very boring. Does anyone claim that the study Amy is criticizing is a good study that should be believed?

  77. edgar says:

    Amy you have lost all credibility:

    “Let me ask you: Do you wish to claim that Johnson and Davis were honest in their presentation of the data?”
    Yes, so far as I can see.

    Do you wish to claim that Johnson and Davis did not know the hospital neonatal mortality rate for 2000 when they submitted their paper for publication in 2004? Do you wish to claim their assertion that it was unavailable to them is true?”

    You state that they had the data in their possession and could have calculated the true rates if they wanted to, but here are the facts:
    -You calculated your rates using CDC wonder, a tool that was not available at the time of the paper. The inference is of course, that D&J could have made the same calculations if they so chose to. This is untrue.

    -You castigate D&J for not using Infant Mortality Statistics from the 2000 Period Linked Birth/Infant Death Data Set. NVSR Volume 50, No. 12. 27 pp. (PHS) 2002-1120. As I said, I do not know their timeline. But after pursuing that paper, I would not have used it. Why? BECAUSE IT DOES NOT BREAK OUT PLACE OF BIRTH which is the hypothesis of the paper.

    Do you wish to claim that their comparison to out of date papers extending back to 1969 was appropriate?

    Absolutely. I would have done the same thing. Lacking other viable data I think it is entirely valid to put up ALL relevant previous studies as a general comparison and address it in the limitations.

  78. edgar says:

    Harriet,
    if you are bored, why are you reading? Is is this a standard tactic, to claim that this is tedious and boring, when in fact we are trying to get to the truth, which is supposed to be the purpose of this blog? If it doesn’t interest you, then stop reading.

    Yes, I think it is good study. I think it has its limitations and some things I would have done differently (I have addressed those points above). Actually I find Amy’s unfounded allegations, half truths and misstatements much more egregious that the study flaws.

  79. Zoe237 says:

    “J&D would like you to think that babies who die DURING a homebirth should just be discarded, not counted at all, as if it had nothing to do with homebirth. That, of course, is simply absurd. Those deaths MUST be counted.”

    Well, no. In the original BMJ study, they WERE included. In the subsequent analysis with the newly available NIH data on their website, they excluded it because linked birth/death certificates don’t include intrapatum deaths. Two different contol groups. Nice strawman though.

    “According to J&D, there were 5 intrapartum deaths out of 5418 births for a truly astounding intrapartum death rate of 0.9/1000, fully 30 times [EDITED: This is an error. The correct number is 3 times] higher than the intrapartum death rate in the hospital!”

    I can’t access the paper that claimed .3/1000 either right now, but I did search intapatum deaths United States and found Amy’s own website claiming a death rate of .62 in hospital in England in 1999. So add that to the neonatal. (.62+.72=1.34/1000 in hospital, 2003). (Very unreliable, but so is the othe analysis). The .3/1000 that Dr. Tuteur cites now is from a Scotland study for 1985-1996. This study also excludes congenital anomalies.

    From website: “…that place intrapartum death rates in the hospital in the same range. The Confidential Enquiry into Stillbirths and Deaths in Infancy in the UK. This was a study of 648,409 births during the year 1999 in England, Wales and Northern Ireland. From the report:

    ‘… Many initiatives at national and local level have occurred in response to these and it is pleasing to see a downward trend in deaths of this type. The number and rate of deaths in this category, weighing 1 kg and over, have fallen from 529 (0.77/1000 total births) in 1993 to 398 (0.62/1000 total births) in 1999.’”

    Anyway, I’ll stop. It’s purely academic to me anyway. Forgive some of my missing rrrs, I appaently need a new keyboad.

  80. edgar:

    “Yes, I think it is good study.”

    Fine. Ultimately each person makes her own assessment. What I’ve tried to do is point out the flaws in the paper. In attempting to address my arguments, people have read the paper very carefully, parsing each sentence, checking references, demanding references for my claims. That’s the way every scientific paper should be read.

    We can argue about various aspects, but some of my claims are beyond dispute:

    1. Johnson and Daviss compared homebirth death rates in 2000 with a bunch of out of date papers when they should have compared them to the hospital death rates for comparable risk women in 2000.

    2. The relevant data were available in 2002, long before J&D submitted their paper for publication.

    3. J&D removed congenital anomalies from the homebirth group, but not the hospital group.

    4. J&D have since acknowledged that the correct neonatal mortality rate for moderate to low risk hospital birth is 0.9/1000. (The actual mortality rate is somewhat lower, but for purposes of this discussion, it doesn’t matter).

    5. J&D failed to disclose that Johnson is the former Director of Research for the Midwives Alliance of North America, the professional organization of homebirth midwives.

    6. The conclusions of the paper are not justified by the data since the authors used the wrong group for comparison purposes.

  81. IndianaFran says:

    “Does anyone claim that the study Amy is criticizing is a good study that should be believed?”

    Of course the study has limitations, the authors include a specific statement related to this:
    “As with the prospective US national birth centre study19 and the prospective US home birth study,23 the main study limitation was the inability to develop a workable design from which to collect a national prospective low risk group of hospital births to compare morbidity and mortality directly. Forms for vital statistics do not reliably collect the information on medical risk factors required to create a retrospective hospital birth group of precisely comparable low risk,38-40 and hospital discharge summary records for all births are not nationally accessible for sampling and have some limitations, being primarily administrative records.”

    In fact, the authors were careful not to overstate their results:
    “Planned home birth for low risk women in North America using certified professional midwives was associated with lower rates of medical intervention but similar intrapartum and neonatal mortality to that of low risk hospital births in the United States.”
    They use the language “similar to”, not “statistically equivalent” or any more precise wording, because they recognize that a direct statistically valid comparative cohort is not available.
    Different readers may question how wide their interpretation of the word “similar” might be stretched, but this statement cannot be said to be categorically false.

    Amy can argue that some home birth advocates do over-interpret the actual study results if they claim that this study “proves” that home birth is “just as safe” as homebirth. It’s really not a definitive answer one way or the other, and it doesn’t claim to be. But it is an important addition to the available literature on the subject because of its scope and its prospective nature (intent to treat).

    Amy continues to assert here and in other places that it is a trivial statistical task to assemble a well-matched cohort of hospital births to compare to home births. Serious scholars who have looked at this question would disagree. Johnson and Daviss add additional discussion of this subject in their followup analysis. (http://understandingbirthbetter.com/files/uploads/BMJCPM2000QuestionsAndAnswersMay2008v2.pdf
    , see pages 3 and 4).
    Perhaps if Harriet is still bored, she might find it an interesting issue to research.

    There is plenty of room to discuss and argue specific data points in the original analysis and followup. The question of how to account for congenital anomalies is also a complex one – especially as an increasing number of these are diagnosed prenatally and may not show up in the full-term pregnancy statistics.

    But arguing the scientific merits of the paper is not the same thing as imputing nefarious intent on the part of the researchers. I (and other posters) have noted several specific places where Amy’s reading goes far beyond an unbiased review.
    “J&D would like you to think that babies who die DURING a homebirth should just be discarded, not counted at all, as if it had nothing to do with homebirth. ” No fair-minded reading of the original paper or followup can support this statement of personal attack.
    “Simply put, the authors pulled a bait and switch. ”
    “Why might the authors deliberately intend to deceive readers?”
    It is far beyond a fair scientific criticism to assert some knowledge of the authors’ intent.
    Edgar has also pointed out Amy’s claim to know exactly what raw data the authors possessed at what time has no basis in fact.
    And of course Amy’s frequently repeated claim that this study “actually shows that homebirth has nearly triple the neonatal death rate of hospital birth for comparable risk women” – when she is knowingly comparing one calculation of neonatal deaths only to another calculation including neonatal plus intrapartum deaths.

    I strongly agree with the previous posters here – whatever the shortcomings of the Johnson and Daviss study, Amy’s critique here has far less credibility.

  82. edgar says:

    Is their discolsure in violation of BMJ disclosure policy?

  83. Basiorana says:

    @moderation– the difference is that those are all active means of treating a problem. Homebirth is basically doing nothing, maybe one step above (hence, safety concerns). And yeah, they did have to prove that antibiotics were better than doing nothing. Homebirth isn’t a treatment, it’s a lack thereof.

    I don’t think the evidence isn’t there, I just disagreed that homebirth was a woo treatment that needed to prove it’s efficacy. It’s more like prayer healing. Of course, at the current time it’s obvious that hospital birth is safer. That doesn’t mean they have to prove homebirth safety. It means they already FAILED, and the answer is not to tell them to prove their point, it’s to show them that their point has already been disproven.

    @ Amy Tuteur: I never said homebirth was good. I said hospital birth was originally introduced without trials and they originally had higher death rates because of the lack of such trials (since they introduced hospital birth before infection controls). They then introduced scientific rigor and quickly became safer than doing nothing, aka, homebirth. But the burden of proof was still on them to prove they were better than doing nothing, not on the people who were doing nothing. They did, obviously.

    But the burden of proof is never on the people doing nothing, it’s on the people doing something. Homebirth isn’t a treatment, it’s a lack of it, so the burden of proof was not on them. What they’re doing now is trying to appeal the fact they were disproven.

    (Keep in mind, this is general about homebirth, not about the specific homebirth philosophies of individuals like Grantly Dick-Read and Ina May Gaskin, where the burden of proof is on them).

  84. edgar says:

    Homebirth isn’t a treatment, it’s a lack of it, so the burden of proof was not on them.

    I actually disagree with that. Hospital birth is the standard treatment (regarding place, not any specific treatment). I think the burden IS on the modalities that do not follow the standard treatment, no matter what they are.

    Furthermore, evaluating alternative treatments can help to shed light in the limitations of current treatment.

  85. Plonit says:

    I think the burden IS on the modalities that do not follow the standard treatment, no matter what they are.

    ++++++++++

    Using that approach it would have been very difficult to counter the practice of routine episiotomy. Routine episiotomy was introduced without evidence of benefit, but based on the notion that surgically expediting delivery would be protective of women’s (sometimes just primips) pelvic floors and also babies (since they would be out sooner). The first RCT studies of episiotomy (Sleep et al) found no evidence of benefit, but they didn’t show serious harms either. Was the burden of proof really on “episiotomy limited to specific indication” (since liberal use of episiotomy was at that time the standard of care)?

    I think you can argue that both hospital and home birth are ‘modalities’, since a home birth with an attendant is not “do nothing.” Therefore the burden of proof is on both modalities, and any ‘advocate’ in the arena should acknowledge that there is an absence of strong evidence regarding the safety of place of birth (note, not just birth at home, but also in the hospital).

    Appealing to the “standard of care” in a particular place and time is not a scientific approach, but a historical and political one. The current standard of care in the UK is that women are offered a choice of place of birth (home, midwife-led unit aka birth centre, obstetric hospital). Presumably there is also no “burden of proof” on this approach, since this is also the “standard modality”?

  86. noelle says:

    I was reading the newspaper this morning and there was an article from AP by Jeff Barnard. It’s about unattended homebirths, the movement is called “freebirth”. Not even with a midwife, it’s the woman and her husband who deliberately chose this way to deliver. Read the article here:
    http://abcnews.go.com/Entertainment/wireStory?id=9130439

  87. pmoran says:

    Even if the studies are held to be conflicting because of low signal to noise —

    Surely the issue is what women are being told about the home birth option and there is plenty of relevant material. They should know that even in the Canadian study supposedly showing the safety of home birth 7.2% of women embarking on it ended up requiring Caesarean section because of unanticipated problems.

    That in another study “shwoing that it is safe” 30% of the women needed transfer to hospital mainly because of fetal heart problems and the need for better pain control. 12%required transfer in another large study.

    These figures do not fit in with the notion of a “safe, natural” home birth” in the year 2009. They show that home birth is highly dependent for its safety upon the back -up of fully equipped birthing resources and the hope of timely transfer to hospital in an emergency. It cannot be an inherently safe option unlesss all possible outcomes of childbirth can be predicted, and they are not.

    Now it is reasonable to argue that hospitals are not entirely safe places either. A lot depends upon the quality control being applied to hospital practices including those factors favouring potentially hazardous interventions.

    But if safety is the issue, then home birth is still not clearly the answer. It introduces its own risks into the birth process. It is a different matter if some will choose to run such risks, but they should be told about them.

  88. fitzerald says:

    “But if safety is the issue, then home birth is still not clearly the answer.”

    Except that the “safety” issue of a home birth is not about staying at home no matter what. With most home births you can have adequate access to emergency treatment, which is why transportation is an important issue. A C-section rate of 7.2% does not make home birth unsafe, it’s just like many women who labor in hospital rooms are taken to the OR for C-sections. Also, the Dutch study that found a 30% transfer rate could be for many reasons. It might be because the women or midwifes did not want to take any chances and were extra cautious. The transfer rate in the US would probably be lower since many women opting for home birth are doing so because they don’t like the environment of the L&D ward.

    Of course, the studies of home births would include women who do not want to be transferred to a hospital, or may not have access to a hospital. And even with them included, the risks of neonatal death are similar to those of hospital births.

    “It introduces its own risks into the birth process.”

    What are the specific risks of home births? Based on the current data available, what is wrong with planning a home birth, especially if you have a trained attendant and adequate services to a hospital?

    The OP’s harping about the risks of neonatal deaths in home births seems to totally ignore the fact that the researchers discouraged a direct comparison of the death rates between hospital and home births. That’s because the difference is so small that it could just be random. So while a critique of the study’s methods is fine, a definite conclusion like the one the OP made using similar data sets is anything but “science based medicine.”

    I was reading the latest post on SBM about the anti-vacciners, and there’s a similarity between the attitudes of anti-vacciners and that of Dr. Tuteur and her supporting commentors on the blog. It’s really just a personal dislike for unconventional birthing rather than the studies themselves. Of course many home birthers also have a stubborn dislike for hospitals, but why would a similar attitude be on the SBM blog?

  89. Harriet Hall says:

    Where fitzgerald sees a “personal dislike for unconventional birthing” I see a concern for giving patients accurate risk information so they can make informed decisions. I would not personally want to be involved in a home birth because of feelings derived from my personal experiences with childbirth emergencies (put simply, it scares me). But I fully support the right of a patient to choose a home birth as long as she is fully and accurately informed and precautions are taken to make it as safe as possible for the baby’s welfare.

  90. apteryx says:

    Dr. Hall, we independent observers would find that easier to believe if “Dr. Amy” had not snarled about women who prefer to avoid any unnecessary intervention in birth almost as loudly as she snarls about women who want to give birth at home in the hot tub. The venom she directed at “natural childbirth” encompassed lots of women who want to avoid anesthetics, choose comfortable positions for labor or delivery, etc. even if they’re in a hospital. Any evidence whatsoever that not getting spinal anesthesia during a hospital delivery risks killing your bay-bee? If not, this isn’t about “accurate risk information”; it’s about “my cultural practices are right, everyone else’s are wrong.” (I do her the favor there of assuming that she and others in the profession never let the thought cross their minds that more intervention means more income.) This is why I tell you that you’ve got a real trust problem with her. If there are no clear numbers on the homebirth issue (and it appears from comments by edgar and others that they were even more unclear than I realized) why should I believe she’s being dispassionately objective, when in other contexts she so obviously is not?

  91. Calli Arcale says:

    Just curious: I didn’t perceive her as snarling at women who wish to avoid unnecessary interventions. I didn’t even see her snarling at women who wish to avoid spinal anesthesia in a hospital setting as you allege. Maybe I missed it. Could you point me to where she said that it’s bad to give birth in the hospital without lots of interventions?

  92. Harriet Hall says:

    apteryx,

    I find it difficult to believe that an “independent observer” would see anything in Amy’s comments that could be considered “venom” or even criticism directed at those who reject unnecessary interventions or want to choose their position or avoid any method of pain relief. Certainly no recommendation for spinal anesthesia.

  93. Plonit says:

    Dr Tuteur neatly encapsulates her approach in the statement:

    “Most people are interested in childbirth that is healthiest, and that, of course, is childbirth with all the tools and interventions of modern obstetrics at hand and used liberally”*

    This viewpoint is at odds with the premise of evidence-based medicine that the benefits of interventions need to be demonstrated rather than assumed. Moreover, it paints those who prefer judicious, rather than liberal, use of interventions as uninterested in the health of childbearing women and their babies.

    *The context can be found at http://www.theunnecesarean.com/blog/2009/12/4/stuff-white-people-like-talking-about-birth.html#comments

  94. apteryx says:

    Dr. Hall:

    She herself, in comments following the rather vile piece portraying women who refused anesthetics as the dupes of eugenic racists, defined the “natural” childbirth she was demonizing as “the childbirth with pain and without episiotomy movement,” indicating specifically that her intention was to promote the use of anesthetics and episiotomies. She claimed that “natural” birth, so defined by her, was “not better, healthier, safer or superior in any way.” This is in part a values claim she has no right to make for anyone but herself, and in part a scientific claim that depends upon either denial or abysmal ignorance of the scientific evidence debunking routine episiotomies and of the potential side effects of anesthetics.

    Several times, she openly maligned and disparaged women who gave birth without drugs (“racist,” “woo,” “pseudoscientific,” “affectation,” “But what made you choose that as a goal?”). Why this deep-seated hostility? Suppose I have a bad headache and, after reading the warning labels on the Advil, decide to just use an icebag instead. Would she fling similar insults at me? What’s it to her if someone chooses to accept “unnecessary” pain?

    She therefore tried to turn it into a safety issue, ranting repeatedly about “wastage” in nature and the dangers of birth: “in the absence of modern obstetrics, approximately 40,000 American women would die each year in childbirth.” If true, so what? Since the subject was natural childbirth per her definition, are we supposed to conclude that women having hospital births are still risking their lives if they refuse drugs or episiotomy? Is she kidding??? To her this seems to be an all-purpose threat: Do whatever the OB says or you’ll DIE, or your baby will DIE. You don’t want your bay-bee to die … DO you?

    Again, you can listen to me or not on this, but believe me, as long as you let any authors on your site rely entirely on insults, threats of doom, and false representations of science that anyone these days can look up online, you’ll be preaching to the choir and the rest of the potential congregation will be down the block in the bar. Good luck to you. You, Dr. Hall, are in my opinion (despite your tendency to conflate philosophical and scientific issues) the most logically minded and civil of the SBM contributors. Here they seem to have lined up the least logical and civil contributor to date, and I really don’t know why you are so willing to defend her. Not for her gender, I hope.

  95. Plonit:

    “This viewpoint is at odds with the premise of evidence-based medicine that the benefits of interventions need to be demonstrated rather than assumed.”

    No one is assuming anything.

    As mentioned up thread, the standard of care is hospital birth with all the “bells and whistles.” The standard of care is determined by what produces the best and safest outcomes; the standard of care is not what most closely recapitulates nature. And there is not one scintilla of doubt that the best and safest outcomes are produced by modern obstetrics.

    Not only is there an extensive scientific literature to support everything that is done (new papers are produced at the rate of hundreds, if not thousands, each year), but there is extensive national and international data to support the fact that modern obstetrics has not only been spectacularly successful; it is the ONLY system that produces low levels of neonatal and maternal mortality.

    In the past 100 years alone, modern obstetrics in the US has lowered the neonatal mortality rate 90% and the maternal mortality rate 99%. To my knowledge, midwifery has not made even a single contribution to this effort.

    Critics of modern obstetrics, particularly midwifery critics, have claimed that obstetrics is not evidence-based. However, they’re wrong about that. Indeed, it has been quite shocking for midwives, particularly those in the UK and Australia, to learn that evidence-based practice supports modern obstetrics and not their criticism. Hence there is a movement afloat within the midwifery community to redefine “evidence” as including the ideological beliefs of the practitioner. I’ll be writing about that in my next post.

  96. Harriet Hall says:

    I can’t believe how Amy’s words are being misrepresented and twisted. For instance, she does not recommend routine episiotomy, which is not evidence-based and is no longer standard practice in any hospital. She was referring to episiotomy as one of the reasons some women chose homebirth back when it was standard practice. She did not disparage women who refuse drugs; she disparaged the faulty reasoning that led some of them to make that choice. When she recommended liberal use of interventions, I don’t believe for a minute that she meant interventions that are not evidence-based. The idea that she meant “Do whatever the OB says or you’ll DIE, or your baby will DIE. You don’t want your bay-bee to die” is a gross distortion, to the point of being offensive.

    This subject is very emotion-laden. Commenters seem to be missing the very valid overall points of Amy’s posts and getting bogged down in minor errors and minutiae.

    When I tried to discuss the scientific evidence for and against circumcision, I was reviled as an evil mutilator of baby boys. I see the same kind of thing happening here. Attacking the messenger instead of the message. That usually means the attackers can’t produce any valid evidence to support their counter-message.

  97. Plonit says:

    The standard of care is determined by what produces the best and safest outcomes

    +++++++++++++++

    This is certainly an arguable point.

    In what sense does routine, universal use of cEFM better and safer than intermittent auscultation with cEFM restricted to specific indications?

    Has restriction of food and oral fluids been shown to be better and safer than permitting (even encouraging) eating and drinking in labour?

    Is the standard of care with regard to indications for augmentation better and safer than more conservative management of labour progress?

Comments are closed.