New study showing homebirth increases risk of nenonatal death. Thoughts?

[loraxc]

FTR, I had a homebirth and would have another.

I realize this will probably be a controversial thread, but the MDC community should be aware that this study is out there. (An acquaintance who knows I homebirthed emailed it to me this morning.) I'm interested in discussing whether this study is flawed, whether it is politically motivated (probably!), but also whether it has something important to tell us.

"Home birth 'trebles risk of baby's death' "

http://www.dailymail.co.uk/health/ar....html?ITO=1490

"Home births are good for mothers but riskier for babies, says study"

http://www.guardian.co.uk/lifeandsty...h-babies-study

[*MamaJen*]

We're talking about it over here: http://www.mothering.com/discussions....php?t=1239367

I got a copy of the original study, and wasn't terribly swayed by it. It's a meta-study, meaning it looks at a number of different studies from a number of different counties. Hence, the data is only as good as the original studies.

They looked at 342,056 planned homebirths, of which 321,307 came from a Dutch study which showed very high levels of homebirth safety. The other 20-odd thousand came from several other studies, some of which date back to the 1970s. They also didn't, as far as I can tell, include the North American Homebirth Study that showed good safety rates.

This is what I wrote in the other thread:
So basically, the bottom line in this study is they're showing a perinatal death rate of .07 for homebirths (229/331,666) vs. a .08 percent death rate for hospital births (140/175,443). Then they're showing a neonatal death rate of .2 percent (31/16,500) for homebirth, versus .09 for hospital (31/33,302).
That "three times as deadly" number comes from they're findings of nonanomalous noenatal death, where they're showing .15 percent (23/15,633) for homebirth, vs. .04 (14/31,999) for hospitals.

So basically, what that means to me is they were primarily using the Dutch study to find that perinatal death rate, which shows very similar mortality at home and in hospital. Then I guess that triple neonatal death rate was pulled from the other studies, many of which are pulled from the 1970s and 1980s.
***
So until I saw which studies had the high death rates, I wouldn't feel comfortable extrapolating too much from this analysis.

[AlexisT]

I posted my thoughts on another thread. I am skeptical of this study. I don't think you can aggregate results from 7 different countries, in studies done over a 30 year span when obstetric practice changed quite a bit. The quality and size of the studies varies, too. I simply don't think it was well done--and I am far from a "homebirth advocate".

[kittywitty]

http://www.theglobeandmail.com/life/...rticle1624918/

[MammaB21]

I have two counterpoints to that 'study' which to me are extremely important to consider. First of all, stating it the way the article does: 'infant death is "THREE TIMES AS LIKELY" in a homebirth', is misleading. While the sentence is true, they're talking about .03% and .3%. Both studies came back with an infant mortality rate of under 1% which is an extremely low risk either way. Certainly moms who chose homebirth are not endangering their babies.

Secondly, and the most important, is that the study doesn't provide any answers to where that extra percent is coming from. What is causing the increase in infant death at a homebirth? Which types of homebirths are being included in this study? Are we comparing strictly planned homebirths with an actively trained and practicing midwife in attendance? To me, that would be the only way I would ever take a study like this to heart. Otherwise, we're including all of the births that took place out of hospital. That would include planned and unplanned UC. It would include births that went too fast to get to the intended place of birth (ie: car births, parking lot births, births on the toilet). All of those have an increase risk of infection for baby due to birthing in a place likely to have higher exposure to bacteria. We would be including all the pregnancies that didn't have prenatal care. In those cases there could have been a problem with the baby that would have otherwise been detected and treated in a hospital setting. We would even be including the births in which there was a known problem with baby and mom decided to continue with her homebirth plans anyway. Think a heart problem that was not compatible with life. I know that in a case like that I would make the decision to birth my baby at home, possibly even if my prior plans had been a hospital birth.

Anyway, it saddens me that 'studies' like this are even out there when they are so utterly inconclusive. I'm also surprised that a doctor from the UK would take any study about homebirth done in the US seriously. There are so many other factors here that aren't being considered.

[*MamaJen*]

Another thing to clearly point out: the cohort of 300,000+ homebirths showed very good safety rates, while the cohort of 15,000 homebirths showed bad safety rates.
Also, another thing I pointed out in the other thread -- in the planned hospital births, they showed a C-section rate of 9.3 percent. Of course most of you know that in America today, it's more like 32.5 percent.
And like a PP said, they don't specify if the births are midwife attended, which IMO makes a huge impact.

[AlexisT]

Originally Posted by MammaB21

I'm also surprised that a doctor from the UK would take any study about homebirth done in the US seriously. There are so many other factors here that aren't being considered.

This part I can answer. The researchers were from the US. However, they didn't do a study per se: they did a meta-analysis of studies from the US, Canada, Australia, UK, Netherlands, Sweden, and Switzerland. It wasn't solely about home birth in the US. However, like I said, I don't see how you can achieve meaningful results by mushing together outcomes from completely different maternity systems.

My understanding is that all the studies were meant to look at planned home birth; however, the quality of the studies varied a lot. I know some of the US studies relied on birth certificate data, which isn't always reliable.

[loraxc]

I thought the study was looking only at planned homebirths (excluded UCs, car births, etc.)

So basically, what that means to me is they were primarily using the Dutch study to find that perinatal death rate, which shows very similar mortality at home and in hospital. Then I guess that triple neonatal death rate was pulled from the other studies, many of which are pulled from the 1970s and 1980s.

I don't quite follow, but are you saying that neonatal rate is drawing only from the US numbers, which are potentially very old? Can we then say that perhaps there was at one time a higher neonatal death rate for midwives (who did this include--trained people, anyone using the title?) in the US, but this is probably no longer the case?

[*MamaJen*]

Originally Posted by loraxc

I thought the study was looking only at planned homebirths (excluded UCs, car births, etc.)

The study only looks at planned homebirths, but does not note if they were with a trained attendant. Basically, they could be treating a UC in a cabin in the wilderness the same way that they would treat a homebirth attended by a CNM one city block from a hospital.

Originally Posted by loraxc

I don't quite follow, but are you saying that neonatal rate is drawing only from the US numbers, which are potentially very old?

Basically, yes. It's hard to tell exactly where those high neonatal numbers are pulled from. That's the frustrating thing. They give a list of 12 studies they analyzed -- three from Canada (1998 - 1999; 2003 - 2006; and 2000 - 2004); two from the Netherlands (1990 - 1993 and 2000 - 2006); two from the U.S. (1976 - 1982 and 1989 - 1996, and note that the big North American homebirth study is not in there); two from the UK (1978 - 1983 and 1994); one from Western Australia (1981 - 1987), one from Switzerland (1989 - 1992) and one from Sweden (1992 - 2004).

By a huge amount, the biggest one is the recent Dutch study -- and we know that that study showed no higher mortality rates for planned midwife attended homebirths. That amounts to 321,307 of the 342,056 homebirths. The remaining 11 studies showed another 20,749 births.

To find the perinatal death number, which was similar to the hospital mortality rate, they combined six studies, for a total of 331,666 births. So obviously, the Dutch study and five others.
To find the neonatal death, which was higher, they combined the remaining seven studies, for a total of 16,500 births. The most significantly higher number was the nonanomolous neonatal death rate, which was pulled from 6 studies amounting to 15,662 births. There were 23 deaths in that group.

So basically, if the majority of those 23 deaths came from UCs in the Australian outback in 1980, that's not convincing data. If they came from Canada in 2004, that would be more alarming. Note that there is no US data less than 15 years old.

[*MamaJen*]

Okay, so just for example: of the 20,749 births that were from the 11 studies other than the Dutch study, 30 percent (6,133) were from a 2002 study that looked at Birth certificates in Washington state from 1989 - 1996, and showed a high mortality rate for home births. Here's a strong critique of that study: http://www.lamaze.org/Research/WhenR...2/Default.aspx

[loraxc]

To find the neonatal death, which was higher, they combined the remaining seven studies, for a total of 16,500 births. The most significantly higher number was the nonanomolous neonatal death rate, which was pulled from 6 studies amounting to 15,662 births. There were 23 deaths in that group.

What was their justification for doing it this way?

[*MamaJen*]

Originally Posted by loraxc

What was their justification for doing it this way?

I have no idea. Basically, it looks like they took five percent of their data, pulled a mortality rate from that, and sent out press releases saying that home birth it three times more dangerous -- based on that five percent break out group, despite the fact that 95 percent of the data showed similar mortality rates for home and hospital. I'm really baffled.

[liz-hippymom]

Originally Posted by *MamaJen*

99 percent of the data showed similar mortality rates for home and hospital. I'm really baffled.

but that in its self is a problem, as the woman with "planned homebirths" are always low risk, whereas "planned hospital births" include both low risk and high risk woman. if homebirth were safer than the number should be lower for deaths, not higher.

and two times as high is twice as many dead babies. when your baby is one of then it matters.

[Banana731]

Originally Posted by liz-hippymom

but that in its self is a problem, as the woman with "planned homebirths" are always low risk, whereas "planned hospital births" include both low risk and high risk woman.

I have not read the study. Is that the case? I would find that surprising.

[*MamaJen*]

Originally Posted by liz-hippymom

but that in its self is a problem, as the woman with "planned homebirths" are always low risk, whereas "planned hospital births" include both low risk and high risk woman. if homebirth were safer than the number should be lower for deaths, not higher.

and two times as high is twice as many dead babies. when your baby is one of then it matters.

I think you misunderstood what I was saying. They had 12 different studies. They grouped half of them into one cohort, amounting to 99 percent of the data (300,000+ births) and found similar mortality rates. Then they took five percent of the data (15,000 births) and found higher mortality rates. Three times higher in five percent of your total data is not twice as much.

[sept04mama]

I just want to say that we shouldn't automatically discount EVERYTHING that is found to be (possibly) negative about homebirth. People need to be informed. I had a hard time researching the "cons" of homebirth when making my decision. I am glad we have an opposing view in this thread, even though most of us will still ultimately still choose homebirth.

I think possibly the reason it may be higher is that a lot of homebirths are postdate, which does include a higher risk of death for the baby. Maybe they took that into account, I don't know.

And Liz, yes, it *does* make a difference when yours is one of the babies who die. I would know.

[mwherbs]

the stats would include all "planned" homebirths- no matter the risk category- remember that OB's consider breeches and twins as high risk and I really don't see any way that they are keeping UCs out of the stats because they are planned homebirths- I will have to go to the med library and read their source studies- in the past the Pang study supposedly had stats on all planned homebirths in Washington but we know how that was off because the researchers use health department data on all births that did not occur in the hospital- so although the mws in Wa are licensed and it would be easy enough to collect that data directly - that is not the info the study used-

[GuildJenn]

Originally Posted by Banana731

I have not read the study. Is that the case? I would find that surprising.

In the B.C. (Canada) study they only included women who met the criteria for home birth whether they delivered at home or not. (And yes in my province, Ontario, and in B.C. there are guidelines.)

However they also only went on the planned location - if a woman was having a home birth and transferred, that outcome counted as a home birth. Which seems fair because in that case the system works.

Anyways what I wanted to add to the discussion is that I think one of the big issues in neonatal death research in first-world countries today is that so many complications and issues now so rarely result in death that the stats can be confounding.

Lumping all complications together -- and putting the cut-off at 7 days for perinatal mortality -- means that it's hard to get a picture of which intervention "would have/might have/should have" helped.

So I don't find it surprising that perhaps particular complications would have better (or at least, not death) outcomes in hospital, and if the complications for the period of study tilted in that direction the study would show that.

It might be that homebirths have higher rates of neonatal deaths because hospitals can intubate newborns more quickly, for example. So you might have more babies surviving to 8 days in the hospital, taking them out of perinatal death stats.

Or conversely, if a study doesn't look at homebirths with a transfer, you could throw the stats in the other direction because more people who transfer will be in the middle of complications.

What I personally would try to get (and have tried to get) from studies like that is a personal sense of where I stand.

If the fractional difference in risk in the meta-study were true, for example, and that fraction were changing my mind about a homebirth, I think what it would be telling me personally is that I wouldn't be ready to be in a situation where there was a possibility that my choice to labour at home was the cause (however rare) of a loss.

In terms of policy, I don't think this study would be a good one to base public policy on because it is cross-country, and each country has a very different climate. In the US for example I am betting that there could be a small but significant number of women who hesitate longer before transferring because of insurance issues, whereas in Canada it's not as much of a concern so women who are homebirthing and find themselves in one of those rare situations might have a smoother transfer (and so better outcomes). Etc.

[*MamaJen*]

So here's something really interesting. If you add up ALL mortality in the study, the hospital mortality rate is actually higher than the homebirth mortality rate (and yes, the studies did match low-risk births for the hospital cohort).

Perinatal mortality, home: 229/331,666
Neonatal mortality, home: 32/16,500
Perinatal mortality, hospital: 140/175,443
Neonatal mortality, hospital: 32/16500

Total mortality, home: 261/348,160 = .0007496, or .075 per thousand
Total mortality, hospital: 172/208,745 = .00082397, or .0824 per thousand

So there you have it.

(I'm actually very excited, because we just had a problem in my statistics class exactly like this, and I'm a big geek and it's cool seeing a real world version of it.)

And one other thing -- that study of homebirths in Western Australia from 1985 - 1990 looks like it was used for a large portion of the data in the neonatal death category. That study showed abysmal mortality rates, along the lines of 7 per thousand. Which makes a lot of sense. Western Australia is enormously rural and has a high aboriginal population, you might be 1,000 miles from a hospital, it included high risk pregnancies attended by unqualified birth attendants -- pretty much everything you don't want to see in a homebirth, and not in line with how homebirths occur today in Canada or the UK or most of the US. I would be really interested in seeing the neonatal mortality rate with that study eliminated.

[MidwifeErika]

Originally Posted by liz-hippymom

but that in its self is a problem, as the woman with "planned homebirths" are always low risk, whereas "planned hospital births" include both low risk and high risk woman. if homebirth were safer than the number should be lower for deaths, not higher.

While that seems like that should be the case, it really is not. I know of midwives who are willing to provide care to women with varieties of health issues so long as the women can get co-care or the midwife is able to consult on these issues. Also, many homebirth midwives provide care to women who are VBACs, advanced maternal age, multiples, breech, grand multips, postdates, etc. Each of these groups I would not consider "high risk," however there is a slight increase in risks for each of these groups.

Also, I am curious about when the planned birth location was taken into account. Does anyone know? Was it the beginning of care or was it at the time of labor?