Transfers
One of the things that I have emphasized repeatedly is that most homebirths will proceed without any problems, and in that case, it doesn't matter who is there. It only matters when something goes wrong. Of most concern to me is situations in which a complication occurs that requires an obstetrician or neonatalogist to save the baby. Homebirth advocates counter that they can recognize such situations in time and transfer the mother or the baby to the hospital.
Let's take a look at what happens in these situations. Below is a table created from data from four major studies that claim to show homebirth as comparable in risk to hospital birth. This table shows the percentage of attempted homebirths where the mother was transferred in labor or where the baby was transferred after birth. It also shows the death rate of babies in the transfer group.
The transfer rate in labor or immediately after delivery is between 10-16%. So even in situations in which the mothers are considered perfect candidates for homebirth, more than 1 in 10 will require transfer to the hospital. This is not a trivial number. It means that anyone contemplating homebirth must anticipate that there will be a significant need for transfer to the hospital.
Of course, we expect the midwives to transfer patients to the hospital when they encounter a problem. The assumption is that the transfer will ensure that the baby will be healthy, because it will receive needed medical treatment in time. The truly startling result is the death rate in the transfer group. It is extraordinarily high, ranging from 21-74/1000. Put another way, the need for transfer in labor or immediately after delivery resulted in the deaths of 2.1%-7.4% of the babies transferred.
These numbers call into question some of the central claims of homebirth advocates. The first claim is that they can accurately predict which women are of such low risk that homebirth will be safe. The second claim is that their ability to recognize problems during labor and transfer patients ensures the safety of all babies. These extraordinarily high death rates suggest that if something does go wrong in labor, a disaster is may well ensue.
Labels: neonatal mortality
30 Old Comments:
Of course babies of low risk women who deliver in hospital, even with the assistance of obstetricians and neonatologists die too. Sometimes we forget that some babies die - very tragically, but all the medical assistance in the world is not going to save every baby. Can we see some stats on hospital born babies who have birth injuries and/or die in order to compare the homebirth rate accurately?
I'm working on that now. I need to find information about death rates for low risk women at term divided into specific complications.
Of course, we already know that death rates at homebirth in the aggregate are higher than death rates at hospital birth in the aggregate. The most likely reason for this is that death rates from specific complications are higher at homebirths than at hospital births, but we need more data to be sure.
How did you arrive at those numbers? In looking at The Farm I do not see they account for anything other than 17 deaths in the study linked. How do you get 74/1000 died and after transfer of care?
I cite in every consultation there is a 1 in 10 chance of transfer and most often that 1 is a primip.
" How do you get 74/1000 died and after transfer of care?"
In the Farm study 230 women and/or babies were transferred to the hospital after labor began. Of those transfers, 17 resulted in neonatal deaths. 17/230 = 74/1000.
Jamie:
"How does planned homebirth cause death before labor?"
It was not easy to tease out the data from these studies. I tried to figure out which IUFDs occurred before labor and which ones were discovered in early labor. I certainly could have made a mistake. If you find an errors, just let me know and I will correct them.
In the Johnson & Daviss paper, I did subtract 3 IUFDs that occured before labor began.
None of this changes the basic result which is that the chance of dying after a transfer is very, very high.
"The studies don't say which babies died at home and which died post-transfer, which is what you'd need to know to make this table meaningful."
No, that does not matter for this purpose. We are looking at what happens when a complication occurs at a homebirth. Some of the transfers happened before the deaths, some of the transfers happened after the baby was born, but the bottom line is that among the transfers, the death rate is very high.
This is an important statistic because homebirth advocates claim that by having hospital back up, they have virtually replicated the benefits of delivering in the hospital. Evidentally, that is not the case. Of course, we won't know for sure until I get good statistics for hospital births separated by complications.
However, I can already see some important trends. In some groups, 100% of the babies transferred for a particular complication (eg abruption) died. I doubt that the death rate for abruption in the hospital is 100%.
Whether the infants died at home or post transfer doesn't matter. The idea is that because they were transfered, rather than already being present at the hospital, medical interventions that could have saved them were delayed, and that such a delay resulted in more deaths.
and you keep saying low risk and I will keep saying that it is a range of risks-- no hospital considers twins and breech births as "low risk" and we both that the Farm specializes in such high risk births-- if a vaginal delivery of a breech or set of twins is done here- it is done by a perinatologist--
also I don't think that using a vacuum in an urban homebirth setting is low risk either- several of these midwives do not stay with in the low-risk category as far as what midwives do and should not be measured against low risk births.
amka saidWhether the infants died at home or post transfer doesn't matter. The idea is that because they were transfered, rather than already being present at the hospital, medical interventions that could have saved them were delayed, and that such a delay resulted in more deaths."
nope not true-- look carefully at least one cord prolapse was caused by an in hospital procedure- happened in the hospital and baby died in the hospital.
"look carefully at least one cord prolapse was caused by an in hospital procedure- happened in the hospital and baby died in the hospital."
Look again. The mother was transferred to the hospital for fetal distress. At the time of ruptured membranes, the cord prolapse was discovered and the fact that the cord had likely been compressed by the baby's head was noted at that time.
I have tried to find a proxy for the neonatal death rate after an in hospital emergency so we can compare it to the neonatal death rates for homebirth emergencies transferred to the hospital.
I think I found a good proxy in data from the following study:
National cross sectional survey to determine whether the decision to delivery interval is critical in emergency caesarean section
J Thomas, S Paranjothy, D James, BMJ 2004;328:665 (20 March).
This study looked at 17,780 emergency C-sections in England and Wales in 2000. Of these 4,622 were performed for immediate threat to the life of baby or mother. There were 43 deaths in this group for a neonatal death rate of 9/1000.
That is considerably less than the neonatal death rates for emergency transfers from home births.
I don't have time to do your mathematical gymnastics, but just wanted to chime in and say that there is nothing wrong with a high transfer rate. It shows that the midwives are being cautious. I know I decided ahead of time with my homebirth that if I even had a twinge of doubt I would transfer and not feel bad about it.
Not all women in labour transfer because of health complications. Some of that 10-16% do it so that they can get an epidural. Do you have any statistics that show how many women transfer for complications vs pain relief?
Mama Liberty:
"there is nothing wrong with a high transfer rate."
Of course, there is nothing wrong with a high transfer rate. The question we are looking at is whether transfer from home is equivalent to emergency treatment in the hospital. It looks as though transfer from home has a much, much higher neonatal death rate than emergency treatment in the hospital.
As a physician, this is certainly what I would expect.
Jamie:
"I would like to reiterate my objections to these numbers, Amy. You just don't have the information to draw these conclusions."
I got the idea for doing these calculations from Angela Horn's website. She pointed out that the neonatal mortality rate was worst for transfers. At first I thought that this was so obvious as to not be worth mentioning. Then I decided to look into it a bit more.
As I have said repeatedly, most homebirths will go quite well, and require no assistance of any kind. The midwife is there for support and to keep the baby from falling on the floor. The real issue is what happens if there is an emergency, since every homebirth study has shown that there will be emergencies. What appears to happen is just what I would expect: babies die of things that they could have been saved from at the hospital.
We could quibble about what should and should not be included in these calculations. For example, we could argue about whether these IUFDs in early labor are IUFDs or intrapartum deaths. Obviously, subtracting deaths is going to make the results look better.
On the other hand, we could argue about whether non-emergency transfers should be included in these statistics. I divided the number of deaths by the number of transfers, regardless of the reason for transfer. As someone pointed out, most transfers are for non-emergencies like failure to progress or desire for pain relief. If I had divided the number of deaths by the number of transfers that were truly emergencies, the neonatal deaths rates would have been dramatically worse, and they are already very bad.
It is very difficult to get a handle on comparable numbers in the hospital. What we are looking for is emergency complications in a very low risk population at term. Then we want to know the neonatal death rate from these complications.
I searched high and low for such data, but most studies do not break down the participants in this way. The study that I quoted (which was done as an assessment of hospital response to emergencies) revealed a neonatal death rate of 9/1000 for the most serious emergencies. The death rate for less serious emergencies was, of course, less. I chose the number that made hospital birth look worst. Keep in mind that this was not a low risk population, either. The neonatal death rate for emergencies in a low risk population might have been even lower.
"It's interesting to me that the paper you reference questions the 30-minute decision-to-incision timeline. They compared outcomes of intervals from 15 minutes to 75 minutes, and here's what they found: "We repeated this analysis with cases delivered within 30 minutes as the reference group. We found no significant difference in the odds of a poor outcome for babies delivered in less than 30 minutes compared with those delivered between 31 and 75 minutes (1.1, 0.9 to 1.4 for five minute Apgar score of < 7). Babies delivered after 75 minutes, however, had an 80% increased odds of a five minute Apgar score of < 7 (1.8, 1.3 to 2.4)."
Counterintuitive, huh? You'd think the faster, the better."
I don't know whether is is counter-intuitive or not since waiting more than 75 minutes produced dramatically worse outcomes. If those results are true, they tell us that babies have an innate reserve. They can withstand oxygen deprivation for a relatively long period of time, perhaps up to 75 minutes long. After they exhaust their reserve, they begin to die.
I have some doubts about this study because of this. This study can be used (and probably is used) to protect doctors and hospitals against claims that they did not perform emergency C-sections in a timely fashion. The doctor could claim that, yes, it is true that he was SUPPOSED to perform a C-section within 30 mintues, but the fact that he saw a few more patients in the office before he sauntered over to the hospital shouldn't be held against him. This study shows that it didn't make a difference anyway.
Even though we don't have exact numbers, a very significant trend emerges. The risk of dying from a complication at a homebirth is very substantially higher than the risk of dying of the same complication in the hospital.
Jamie:
"In a three-month period, then, there would have been roughly 151,110 babies born. That means in 11.9% of pregnancies, women were delivering via emergency C-section."
I did not agree with that assessment either. That's why I used only the 4622 true emergencies (approximately 3% of the total) to calculate the neonatal death rate of 9/1000. We can refine that even further by subtracting women who were sectioned for failure to progress, breech, and malposition (which are not true emergencies) and find that there are only 4162 true emergencies. In that group, the death rate is 10/1000.
As I said before, I think there are some serious problems with this study. One of them is that many of the C-sections characterized as "emergencies" are not emergencies at all. Of the 17,780 C-sections in the study, only 23% of them are truly emergencies.
In order to answer the question about whether the interval to C-section matters, the authors should have only looked at those 4162 C-sections. 43 of the 52 deaths in the entire study occurred in that group. I would like to know how those deaths compare with the timing of the C-section. Unfortunately (conveniently for them), the authors neglect to provide that information.
Another thing that makes the data questionable is that there is no explanation as to why it took them much longer to start C-sections for some women than for others. Although they imply that every C-section was of equal urgency, that may not be the way that the staff at the hospital interpreted it.
It could be that the women who waited longer for C-sections were actually lower risk than the women who had immediate C-section. I say that because fully half of the deaths occurred in babies who were delivered within 30 minutes. Then the death rate drops precipitously and does not begin to rise again until an hour.
ACOG believes that an emergency C-section must be started within 30 minutes of noting the emergency. That suggests that there are other papers which show a definite difference after 30 minutes.
In any event, I only used that paper because I was looking for some estimate of the percentage of C-sections that are performed for emergencies and the death rates in that group.
This is interesting though not especially surprising.
After all, this is exactly the subset of home births in which you would expect to see the largest number of problems. To be honest I am surprised that some people here are, well, so surprised by this.
Doesn't it make sense?
1) SOME major issues--about 10%, it seems--CANNOT be detected until they happen.
2) Of those, most will require transfer to the hospital (note the use of "major" in the first sentence.) How do we know this? Well, we can look at when midwives transfer their clients. Clearly if those issues had been known ahead of time, they never would have delivered at home: nobody wants to do an in-labor or emergent transfer.
3) Of those that DO require emergent transfer, the infant is more likely to have problems or die.
How to we know this? Again, this is to be expected: A bedroom and/or an ambulance with a midwife in attendance, are not equivalent to a surgical suite and/or a NICU with the appropriate doctors, drugs, and equipment.
What can we conclude from this?
Well, from this alone, not much--which is why I am always amused if this gets protested. After all, this is merely ONE issue under comparison, and it is not conclusive on the OVERALL issue of safety.
Can we compare these rates with hospital births with the same complications?? Otherwise, this is meaningless.
What about the National Birthday Trust study reported on the homebirth.org.uk website you link to? That describes 769 transfers (a rate of 16%) of which 4 babies died, or a rate of 5.2/1000. Unless there's something wrong with the study (and I defer to your expertise at that discernment), it seems that this is anomalous with regard to your other citations, not to mention much more favourable to homebirth since the national death rate for babies during birth in the UK is 8/1000. Any explanations for this?
- Blaise
Blaise:
"What about the National Birthday Trust study reported on the homebirth.org.uk website you link to? That describes 769 transfers (a rate of 16%) of which 4 babies died, or a rate of 5.2/1000."
That's true. I can't find information that breaks the transfer rate down by reason, but the summary says that at least 62% were transferred for failure to progress or prolonged rupture of membranes. The remaining 38% (292) probably included additional non-emergent transfers like undiscovered breech etc. Therefore, I can't get find out the true number of emergency transfers. If all 292 transfers were for emergencies, the neonatal death rate in that group would be greater than 14/1000.
Amy wrote:
Look again. The mother was transferred to the hospital for fetal distress. At the time of ruptured membranes, the cord prolapse was discovered and the fact that the cord had likely been compressed by the baby's head was noted at that time.
MY reply is that this woman in first stage may have just arrived at the hospital for a normal birth or may have had some undetected fetal distress-
Dr Amy,
I think it would be wise to do further research on transfers and death rate. Despite what you think you can concur as a physician, you are not an epidemiologist, and making your own judgements and conclusions about what numbers may or may not mean is out of your scope.You can't just decide to take a low number like "200" and multiply it by 5 to make a "per thousand" statistic. That is almost laughable.
The fact is, there are no credible studies that prove that homebirth is less safe than hospital birth for low risk women.
Furthermore, there is not a midwife I know, including myself that has "anywhere from a 10% to 16% transfer rate." Try more like 3-6%. That would be more accurate.
~Midwife of many years!
"The fact is, there are no credible studies that prove that homebirth is less safe than hospital birth for low risk women."
No, the fact is that while there are studies that CLAIM to show that homebirth is as safe as hospital birth, none of them actually show it. You don't need to have a degree in epidemiology to know that you need to compare like with like. None of these studies compare homebirth with groups of women of equally low risk.
we don't know that , again not all this is low risk and cannot be measured that way-- we will keep going round and round- what are the risks for vaginal breech? what are the risks for vaginal twins? --- how do these groups match?
what about congenital anomalies incompatible with life? do you have a hospital population who does not test or abort of anomalies? - the match for them could be Amish- what are the regional hospital stats on that? how about in Montana, Indiana, Pennsylvania.
I am so sick of this crap
here goes
as for the Farm-
lets really look at it- have you read their book? it tells you much of what happened and you need to remember the times as well.
So risk factors I would put on the Farm - early days- NO prenatal care- NO trained attendants - poly drug exposure including in labor, college students and drop outs ,many jobless and homeless ( bus caravan folks with no homes) unwed-- what is the fitting demographic for this group in the early days?
Then training from an old doc- from the first losses there is an understanding of need to learn more and change. Also belief system stated- anti-abortion in the early books (and maybe in the later) offer to have pregnant moms come work on farm and have baby there and leave baby there if need be instead of having an abortion-- so unwed moms with few resources and morally against abortion going to the Farm to give birth -- this is once again not a low risk group-- more training and refinement along with changing times- due to experiences now specializing in other high risk groups- women giving birth to twins, or breech babies- known internationally for doing births in this fashion- women come from all over to give birth in this setting. How an OB would assess risk the majority of births that have occurred on the Farm would not be considered low risk.
-------------------------------------------
"as for the Farm- lets really look at it- have you read their book?"
No one with any scientific training thinks that the Farm study is valid. The only people still quoting from it are homebirth advocates.
Furthermore, that study is old. Where are the statistics from the Farm for the past 25 years? Why don't they list their neonatal death rates on their website? Are they hiding something?
since that one collection on the farm included stats thru 2000???? where do you get the last 25 years?
"since that one collection on the farm included stats thru 2000"
Amazingly enough, the stats from 1970-2000 show fewer deaths than 17 reported in Durand's study from 1970-1989. Obviously the stats in the book are not accurate.
one point - births at the Farm just from demographics and known prenatal conditions such as breech and twins make it a high not low risk group... and if you are going to compare at all that would be how it is done.
Now I did try to find stats in the same time period that would cover vaginal twin and breech births- but guess what- they controlled out all the congenital abnormalities... something you say isn't suppose to be done.
Dr Amy wrote:
"No one with any scientific training thinks that the Farm study is valid. The only people still quoting from it are homebirth advocates."
Yes.. the homebirth advocates and you Dr. Amy. !
My family all the time say that I am wasting my time here at web, but I know I am getting familiarity every day by reading
thes fastidious articles or reviews.
Feel free to visit my blog post ... sacramento cash advance
The transfer rate in labor or immediately after delivery is between 10-16%. So even in situations in which the mothers are considered perfect candidates for homebirth, more than 1 in 10 will require transfer to the hospital. This is not a trivial number. It means that anyone contemplating homebirth must anticipate that there will be a significant need for transfer to the hospital.
<< Home