A flawed study published in “a second-tier journal”

That was the conclusion of Dr. Nicholas Fogelson about the much-publicized recent meta-analysis on home-birth, in a comment on Science & Sensibility’s post about it. [Honestly, I didn’t know that the AJOG was so poorly respected. Makes me rethink what studies I’d look at, if it was published there.] In the S&S post, Amy Romano questions whether the meta-analysis was even appropriate, in light of the huge de Jonge study of over 300,000 planned home births in the Netherlands. She writes,

Lack of statistical power could not possibly be the rationale for conducting a meta-analysis on the safety of home birth. [The de Jonge study] contributed 94% of the data on planned home birth in the meta-analysis (321,307 of 342,056 planned home births). That study found virtually identical rates of neonatal death in both the planned home and planned hospital births, with relatively narrow confidence intervals.

The blog “Birth Sense” has two posts on this study: a prologue of sorts and then a discussion of the 12 studies included in the meta-analysis. And despite Dr. Fogelson’s comment that AJOG is a “low-impact” journal, many news sources apparently didn’t get that memo, and had no problem trying to make a huge impact with the sensational headlines that “home birth is three times as deadly as hospital birth.”

I have read numerous blog posts and news articles about this study and will include links to them below, but I want to comment on some things. First, since the study isn’t going to be released for a few more months, I haven’t read it (I know it’s available by request, and I presume for a price or if you have a subscription), so I’m not sure what I can add to what those who have read it have said. In some ways, what I’ve read has raised more questions than answers. The main question is, what were the inclusion criteria? From some sources, I’ve discovered that the original study had to have been published in English (a common enough restriction — gotta make sure the doctors don’t get bogged down in any bad translations from reading foreign language studies!), and also to studies that reported outcomes of mothers and babies in both home and hospital births (apparently the reason why the Johnson & Daviss 2005 article on CPM births, published in the BMJ was not included — they didn’t have a hospital “control group”).  But I wonder, did they not have any exclusion criteria for the size? They included one study that had 5 women in each group — a total of 10 women studied! If I may say so, “Good grief, Charlie Brown!”

Additionally, they included the Pang study, which itself was flawed, in that it reported not only planned home-births (with or without a qualified attendant), but it also included unplanned home-births — this despite a title that declares that it looked at planned home births. Nor was the study limited to the “term” period, but included births as premature as 34 weeks. Since I don’t have the meta-analysis in front of me, I can’t say how heavily it influenced the conclusion, but “garbage in, garbage out” — if the data going into the study is wrong, you can’t reach the right conclusion.

In all meta-analyses I’ve read, there has always been a discussion of inclusion and exclusion factors; if there are others besides the ones I’ve mentioned above, I’d like to know — out of curiosity, if nothing else.

Some of the articles I’ve read have touched on problems researchers may have in reconciling different studies, which may all have different designs, or look at different populations, etc. Again, I don’t have the study, so I don’t know if Wax and associates looked at all the different factors and did an appropriate exclusion or analysis of differing studies. Some differences among studies that I can think of that might need some statistical calculation or exclusion of some births (which may or may not have been included in any of the cited studies), include prospective vs. retrospective study design; inclusion of term, preterm or post-term births; cross-over (women planning a home birth who transferred to the hospital, or women planning a home birth who gave birth precipitously outside of the hospital); whether babies with congenital anomalies were excluded; whether the mothers were properly matched in the groups — same general age, income, health, parity, etc.; whether twins or breech or post-term babies were included or excluded (many studies are limited by design to “singleton vertex babies from 37-42 weeks gestation”); cause of death (for example, excluding accidental deaths by smothering or car wreck would be a good idea, I’d think, if possible); whether there was a trained birth attendant; quality of the original study, and undoubtedly many other things I can’t think of.

One of the confounding factors that some of the articles I’ve read, is that some of the studies examined only the early neonatal period (up to 7 days), while others examined the full neonatal period (up to 28 days). Which reminds me of another factor I should have included in the paragraph above — perinatal deaths including stillbirths, and also possibly intrapartum deaths. Several of the writers have noted that the largest study, the Dutch study, was not included in the actual analysis of the most loudly trumpeted conclusion, namely, that home birth carries a 3x death risk for babies when compared to hospital birth. The largest study which had almost 95% of all of the births in all of the studies, and showed no difference for neonatal death, was excluded for the purposes of  neonatal death. It seems that it is because that study only went up to 7 days. Was there not some sort of statistical analysis that could be done to extrapolate deaths that may have occurred from days 8-28, so that this study could have been included? Or, an analysis that could have excluded deaths from 8-28 days in the other studies? As I said, I have more questions than answers. I’d love to read the study for myself (and the Birth Sense link has links to all 12 of the included studies — some of them are abstracts but some of them are the full studies, so I may start by reading those), but I wonder if any of the questions I have would be answered if I did read it.

Now for the links that I read, in addition to those I cited above (some of them include other links to other articles, as well) — obviously, most of them are going to be favorable to the “this meta-analysis is junk” viewpoint, but not all of them are:

Now for some links to other home-birth related stuff that is not about this flawed study:

  • The Unnecesarean’s blog post on a new study in this month’s issue of Obstetrics & Gynecology that says it’s important to distinguish between planned and unplanned home births [and I’m just in the mood right now, due to lack of sleep, to say cheekily, “and next, they’ll tell us that water is wet, and the sun is hot!” :-)]; but seriously, it’s because many times “unplanned” home births are to women with known risk factors that may cause problems with either themselves or their babies
  • the NHS write-up of the de Jonge (Dutch) home-birth study
  • High Tech vs. Nature’s Way,” an article from Minneapolis-St. Paul
  • The College of Physicians and Surgeons of British Columbia “Resource Manual” on planned home births

Ahhh, now I can close out some tabs on my browser.

Oh, and Dr. Amy, I’m pretty sure you’ve been waiting for me to write this up since I know you read me (not to mention you probably have “home birth” on your Google Alerts, and probably regularly add other keywords to make sure you don’t miss any internet discussions), so if/when you comment, just remember that you have to stay on topic. No cut-and-paste jobs like you usually do, m’kay? Oh, and before you bluster that the studies which showed that home birth is safe only in the countries of the study (like the Netherlands and Canada), because of how their midwives are integrated into the health system, I want you to answer a question: if that is true, why don’t you try to integrate midwives into the health system in America, rather than trying to elbow them out? That would solve that problem.🙂

3 Responses

  1. Did you see the NYT article? Wax sort of backs down.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: