Home Birth & Apgar Scores in the AJOG October 2013, some clarity.
In the October 2013 issue of the American Journal of Obstetrics & Gynecology (AJOG, also known as the Gray Journal) there are two “research” articles discussing outcomes related to birth setting. In today’s blog I take a critical look at both of them so if asked about them by a prospective client you can give a salient response. Since science has been corrupted by money and ideology it is difficult to know what to believe these days making reliance on common sense and clarity of intent that much more important.
The first article begins on page 323 and is titled, “Apgar Score of 0 at 5 minutes and neonatal seizures or serious neurologic dysfunction in relation to birth setting”. This ominously titled article is authored by Amos Grunebaum, MD out of Cornell University with multiple co-authors including Frank Chervenak, MD who we all know to be fervently biased against home birthing. In a previous blog I have been critical of the authors of this article for their deviation from scientific norms by promoting their paper through an uncontested press release more than a month prior to its publication. Thus, headlines received prior to any critical review. Let’s take a look at the data and methodology of this study and then analyze their conclusions.
The stated objective is to examine the occurrence of 5-minute Apgar scores of 0 and seizures or serious neurologic dysfunction for 4 groups by birth setting and birth attendant (hospital physician, hospital midwife, birth center midwife, and home midwife) in the United States from 2007-2010. They used birth certificate data files from the U.S. Centers for Disease Control’s National Center for Health Statistics for singleton newborns >37 weeks and >2500 grams.
In order to trust the results of any study we must first trust the information upon which the study is based. A retrospective study of birth certificate records, especially as used in this study, is ripe with unreliability. A simple example is the inability to determine precisely who attended a birth as the signature on the record does not mean that practitioner was actually there. I have signed birth and death certificates in my career for patients I did not care for simply because I was the physician on call and was asked to do so by the hospital staff. In 2003 the standard Certificate of Live Birth was upgraded to include the location of birth as hospital, birth center, or home. And it was further specified as accidental, intended, or unknown if intended. In 2008, only 27 states were using this upgraded form yet the Grunebaum study admits to using all U.S. births from 2007-2010 that met their inclusion criteria of >37 weeks and >2500 grams (of the 16,693,978 births they included 13,891,274). The second study I will discuss calls this methodology into question as their use of the more accurate revised form had them exclude nearly 50% of American births in the year 2008 alone. To summarize, birth certificate information is notoriously unreliable and often does not differentiate whether the birth was planned or unplanned, attended or unattended. Yes, someone has to sign it but it does not mean the doctor or midwife signing actually attended.
The authors then rely on the 97.5% of states (?) that had collected data on the presence or absence of neonatal seizures or serious neurologic dysfunction in their birth certificates. I would question the reliability and collection methods of this data. The definition of a neonatal seizure is seizures occurring in the first month of life. Birth certificate data is usually filled out in the first couple days of life. I am not sure how the authors then used this data responsibly. Also, it would be important to know long term outcomes of the neonates who experience seizures or serious neurologic dysfunction and how that correlates to lifelong disability. This study does not clarify this endpoint.
The term relative risk (RR) is used in this study to emphasize their conclusion that homebirth attended by a midwife has an overall 10.55 RR of a 5 minute 0 Apgar score. The RR is slightly higher in primips than in multips for all categories but RR can be quite misleading especially in the headline of a press release which was their intention, I believe. The overall risk in this study of a 5 minute Apgar score of 0 in the hospital attended by a physician was 1.6/10,000 births. For home births with a midwife it was 1.63/1,000. While this is a 10-fold increase it is still rare when you consider all the other risks of birth. It is purposeful deception to use statistics selectively in this way. Especially when they conclude that all obstetric practitioners must disclose this information to all pregnant women who express an interest in out-of-hospital birthing. “Surprisingly”, this article concludes with a recommendation that concurs with an article written last year by Dr. Chervenak on the professional responsibility of physicians to recommend against planned out-of-hospital births to women who express an interest in it. Amazing that based on a risk of .163% of a 5-minute Apgar of 0 doctors are told to advise against home birth but a 33% risk of surgical birth is no problem. A higher risk of infection, breast feeding issues, psychological trauma, mother-baby separation, induction/augmentation are not reasons to warn mothers against hospital birthing? The skewing of data to funnel women down the medical path is without shame. What more can I say?
Well, on this article there is one more glaring point to make. I could not find anywhere in this article a distinction between a live born newborn with a 5-minute Apgar of 0 and a stillborn. If a woman at term was diagnosed with an intrauterine fetal demise at term and chose to have her baby at home would that count as a 0 Apgar at 5 minutes? Finally, I have spent more than 27 years working in hospitals with NICU teams. When a baby is born with a very low or zero Apgar at 1 minute it is pretty much standard of care for the NICU team to rush in and perform their initial magic of resuscitation. Even if the baby were to be allowed to expire at 30 minutes or was to be revived only to die hours or days later the Apgar score at 5 minutes would not be a 0!! I would submit that the risk of 0.16/1000 for a 0 Apgar at 5 minutes in the hospital setting as quoted in the article is artificially low. And if this reasoning sounds logical then the entire premise of using the 5-minute Apgar score as a marker of quality of care by practitioner or location is undermined and serves to fit the Chervenak model of selective ethics! Maybe it would be professionally responsible for physicians to tell their patients that!
(Please see my comments following this blog)
The second article on page 325 in the October 2013 issue of the AJOG is titled, “Selected perinatal outcomes associated with planned home births in the United States” by Y.W.Cheng, MD et al, from UCSF. This was a retrospective study, also, of term singleton live births in the U.S. in 2008. Of the over 4,000,000 births that year, 2,081,753 met the study criteria of using the 2003 birth certificate revision discussed above. Of these births 0.58% were planned home births. They found that planned home births had a higher rate of 5-minute Apgar score less than or equal to 4 with a RR of 1.87 (3.7/1000 vs. 2.4/1000) and a RR of 3.08 for neonatal seizure (6/10,000 vs. 2/10,000). They also concluded that women with planned home birth had fewer interventions such as operative vaginal delivery and induction/augmentation. They also broke down planned home births by CNMs or “other” midwives.
Their findings and the presentation of their data in style seemed much more neutral and unbiased. For instance, they do state that infants born to women with a planned home birth are more likely to have a neonatal seizure (6 vs. 2 per 10,000, still a small number) but also say that NICU admission was lower among infants of planned home birth (RR 0.23). They did emphasize that outcomes for homebirths with CNMs did not differ significantly from hospital births but that with “other” midwives the risks were greater. However, I could not find a definition for “other” midwives in the article.
I found the self-reflection and critiquing of their own research to be honest and refreshing in stark contrast to the first article’s air of certainty which feels like smugness to me (my bias!). Cheng, et al admit their study has limitations. “As a retrospective study, it may have included confounding or missing data that could potentially bias our findings.” They go on to say, “….administrative data, such as birth certificate data, may contain inaccurate information.” Now isn’t that refreshing to hear from a researcher? They also honestly admit they could not identify or differentiate women who planned a home birth but who were transferred to hospitals which they admit occurs in 10-15% of planned home births (Notice the use of 10-15% rather than the skewed numbers of up to 47% Dr. Chervenak used in a previous opinion piece last year). This could elevate the risks at home but they admit they cannot be sure. Finally, Dr. Cheng concludes, “Because of the complex tradeoff between maternal benefits and neonatal risk, women who contemplate location of birth should be fully informed about both sites”.
Here are 2 articles, side by side, both presenting a selected endpoint on the outcomes of term singleton births by location and practitioner. However, the difference in style and honesty is striking. The first article has a co-author with a deserved reputation for skewing his data to fit his ideology. While I cannot quantify his influence on his fellow co-authors I can use experience and common sense to recognize bias, selective data mining, intentional omissions and flaws in methodology and conclusions. It is important for all of us to look at research with a critical eye. Understand that relative risk is rendered far less meaningful when we are comparing a small number to a multiple of a small number. I must agree with Dr. Cheng. Once the sperm and egg unite there are risks assumed. True, not skewed, informed consent must be given about all options. And respect for the individual is paramount for it is the essence of ethics to respect the autonomy of patient decision making and it cannot be expected that given the same information two people with differing life experiences will always reach the same conclusion. Finally, good science does not need to be preempted by a press release. I have been concerned of late that science is more and more being corrupted by money and ideology. While we may not prevent this we must remain aware of it and read critically and maintain a healthy skepticism.
Dr. Stu