I think it is time to sum up and come to closure on safety, training and certification, and regulatory considerations.
With regard to safety:
First of all, yes, the chi-square test is a tool that is often used in looking for statistically significant differences. If the objective is to look for differences in outcomes arising from the setting, then it is important to address confounding factors in the data. There are many independent variables that affect the probabilities including presentation (i.e. head down vs. breech), lifestyle affecting health, economics, parity (how many babies the mom has had) etc
I know john has studied the data in the CPM2000 study, but for completeness, lets review the outcomes. There were 4 fetal deaths that delivered at home, 5 intrapartum deaths (I know there is a difference of opinion here relative to the bins), and 9 neonatal deaths. First of all, the 4 fetal deaths would not be counted as intrapartum or neonatal deaths in hospital statistics. They would be counted as fetal deaths or perinatal deaths. Fetal deaths are counted separately.
As for the intrapartum deaths, one was a precautionary transport in first stage that was allowed to proceed vaginally in hospital. The planned home setting was irrelevant and this illustrates the limited abilities to prevent bad outcomes in hospital. There were 2 bad outcomes associated with breech presentation (out of 80 total). The safest form of delivery for breech presentation may very well be planned C/S and these are not considered low risk. The thing to keep in mind here is that this discussion is generally to assess safety for low risk women. There are women who will not birth in hospital even with breech presentation and it is far better that they have a midwife than to go unassisted. In any event, in the discussion of safety for low risk women, it is appropriate to remove the breeches (2 bad outcomes and 80 total pregnancies). The fourth was a severe bleeding complication that would not likely have been resolvable in hospital. The fifth was a severe cord accident. Nevertheless, the appropriate number of intrapartum bad outcomes to assess for low risk women is 3.
Now for the neonatal period. Weve discussed that there is an excess of congenital anomalies with the Amish and Mennonite families. John is right that if we take out the outcomes then we must take out the entire population. There were 467 Amish and 194 Mennonite families that should be removed. As for the other 6 bad outcomes, 2 were SIDS, 1 likely pulmonary defect or possible SIDS, 1 GBS disease undetected in hospital. So, the setting was likely irrelevant, and it is appropriate to count 6 neonatal deaths for the population we describe as low risk.
The appropriate analysis basis values for low risk women are then 3 intrapartum and 6 neonatal deaths in (5418 467 Amish 194 Mennonite 80 breech 13 twins) = 4664 births. Even if you place all 3 intrapartum deaths in the neonatal period, the chi-square test will conclude they are indifferent. Alternatively you can compare 5/5000 to 6/4664 neonatal deaths and conclude they are equivalent by inspection.
Now this is just one aspect of safety. We know the C/S rate is way too high (>30% nationwide) and we know there is an excess of maternal mortality with C/S; although, the probability is lower for a maternal mortality as compared to a neonatal death. We also know that the likelihood of a hysterectomy is greatly increased with C/S (lost opportunity babies). Furthermore, the frequency of premature low birth weight babies is much lower for us due to improved maternal health associated with midwifery care which will resulted in fewer bad outcomes when we look at the entire perinatal period. With a circumspect view, it is easy to conclude planned home birth is safer. Now, if a woman feels safest birthing in hospital, then the hospital is the perfect setting for her. For women who choose to birth at home, it is outrageous to declare them out of process and deny them care and normalcy.
While Amy is quite vocal on this, the obstetrical community of practice is not all on the same page here. I urge everyone to read (it is not long) the joint RCOG/RCM statement on home birth from the UK.
With regard to training and certification:
We had a very thorough evaluation performed comparing the clinical requirements to obtain the CPM vs. CNM credential, and it was presented to the House Select Committee on Licensing Midwives in March. The conclusion was clear that the clinical requirements are comparable related to providing care during the childbearing year. Keep in mind that the CPM credential is accredited by the same organization that accredits the CNM credential. Yes, the training is largely in the clinical setting, as opposed to a classroom setting.
With regard to regulatory considerations:
This is very easy. Some women will always choose to birth at home and they deserve access to care. The national standard for midwives attending women at home is the CPM. More and more states are licensing CPMs and no state is going backwards. There are women in North Carolina birthing unassisted at home because they do not have access to credentialed, licensed midwives. Aside from assuring access to care, licensure also assures that the standards of care are maintained. Clearly the correct action is to license CPMs.
From every perspective, increasing access to credentialed, regulated midwives is good for North Carolina just like it is for the states with the best perinatal mortality statistics. I will provide some coaching to my adversaries, however, on how to minimize the planned home birth population. Simply address the quality of care in hospital, and never again will there be a woman who says I understand the logic related to being in close proximity to obstetrical problem solving when birthing, but I have birthed in hospital and I was traumatized, and I am not going to do that again. Fix that and they can achieve their goal of minimizing the home birth population, but there will always be women who choose to birth at home. Trampling liberty is not going to solve our problems.
Legislative Chair, North Carolina Friends of Midwives
Hey Meg, CNM,
I absolutely agree with what youve said.
Dont get me wrong, no one is saying if we have a bunch more home births, and home birth midwives, outcomes will improve. At the same time, if every mother had great midwifery care (coupled with great OB care when they are sick) irrespective of the planned setting for birth, then maternal health would improve, the C/S rate would improve, outcomes would improve. This is just one step in that direction; however, I know CPMs who would immediately volunteer their time to give free prenatal care to impoverished women (planning to birth in hospital), but cant given their status in NC (of course they are licensed in CA). These women spend 2 hours in county health department waiting rooms for a 10 minute diagnostic analysis, to then birth with people theyve not met before (with all of the deficiencies youve noted).
I wish I could say that the environment has improved for CNMs in NC and entice you back, but according to the ACNM model that attempts to describe the goodness of a state for professionals, we havent improved in 10 years (and I think it is even worse for OBs).
Spoken like a true engineer (I thought so)! Well done! Ya just gotta love engineers (every team ought to have at least one).
Im pretty happy with where you have landed and if everyone believed that birthing in hospital was 99.9% safe and birthing at home was 99.8% safe (for healthy women experiencing normal pregnancies), then I would be a happy boy. My adversaries prefer to think in terms of percent differences, and only consider this one particular variable.
Now Im not trying to move your opinion (because I am perfectly fine with it), but I need to point out a few things (and I will complicate safety analysis for you just a little bit later).
1) You dont get to vote on the data. Your instrument says there were 5 intrapartum deaths and 9 neonatal deaths. You need to just accept that.
2) In your evaluation, you have to do one of two things. Either compare the neonatal mortality rate in J&D to the neonatal mortality rate in the NIH data, or take the combined (intrapartum + neonatal) mortality rate in J&D and compare it to the NIH data plus some value for the intrapartum component. You have found a reference for intrapartum mortality of 0.3/1000. I have seen references as high as 0.7/1000 for large populations. I would tend to favor comparing neonatal to neonatal in which case you would compare 0.9/1000 (hospital) and a range of 1.2/1000 to 1.8/1000 (home) depending on what you do with the anomalies if you make no adjustments based on race.
3) You dont have to embrace that the congenital anomalies would be counted differently to consider treating them separately. Several, if not all, of the anomalies in the home birth group were associated with the Old Order Amish and Mennonite families, which have a higher incidence of such things. These are not the same populations.
4) Amy is comparing datasets taken with different instruments. She is comparing retrospective data (the NIH data based on a review of records) to prospective data (the J&D study). This is considered inappropriate in the world of epidemiology - at least if the intention is to compute percent differences, particularly given the large differences in sample size.
5) Here is something to consider just looking at the setting, http://www.collegeofmidwives.org/Political_Action_2006/ACEO_%20Statement_Policy_OOH_Final_06.pdf. Reference 6 (Schlenzka) is a significant work that shows better outcomes for midwife attended Out-of-Hospital birth, even with elevated risk factors (you can find the complete 200 page thesis with the help of Google).
There are other considerations as well that clearly favor midwifery care (e.g. frequency of low birth weight babies - the setting for labor and delivery is irrelevant, but the care provider during the prenatal period is important here).
Now, here is where I complicate things just a little. This whole conversation is indicative of the deficiencies in the way the clinician views safety analysis (at least when they argue this argument from the other side of the table). Without question, there is an excess of maternal mortalities due to the C/S rate, albeit the absolute rate is lower than that for the baby. At the same time, it is largely agreed that a maternal mortality is a much more traumatic event for the family than the loss of the baby (and therefore weighted higher), and the rate is grossly underestimated. If a woman gets an abdominal infection due to C/S and dies 6 months post-partum, she is not counted as a maternal mortality. Also, women are losing their fertility (i.e. hysterectomy) due to our C/S rate. For talking purposes, you can use the following analysis basis probabilities:
Peripartum hysterectomy for vaginal birth = 1/30,000
Peripartum hysterectomy for 1 C/S = 1/1800
Peripartum hysterectomy for 2 C/S = 1/1300
Peripartum hysterectomy for 3+ C/S = 1/200
So, do you view an inadvertent hysterectomy as simply whatever discomfort is associated with the event (even if life-long), or do you view it as some number of lost opportunity dead babies? My personal opinion is that women will evaluate this differently when evaluating safety. There are myriad other considerations (some of which clearly favor planned C/S over vaginal birth in any setting, such as breech presentation and hypoxic brain damage). I think you belittle safety when you equate a small number of (postulated) excess neonatal mortalities to a big difference in birth experience.
So, my friend, what is your definition of safety in the analysis of birth models?
Finally, when you consult with your OB friend on intrapartum mortality, please view his/her response as anecdotal. I have observed OBs say the probability of suffering a uterine rupture in trying to birth vaginally after a C/S is 1 in 10, or 10%, and you need to just go ahead and have another C/S. Overall, the probability is more like 1/200, or 0.5%, and much of this is due to obstetrical practice of induction and augmentation of labor (increases the likelihood by 2x or 3x). Also, please shake his/her hand for me for his/her life-long service to women and families.
Again, I am real happy with where you have landed. I say Cheers to you, or rather Kampai in light of my setting.
Hi John D,
Please read the link I posted previously to see the counter arguments to these claims on mortality rates (prepared by epidemiologists).
Right from the get-go, you see that Amy is comparing the (intrapartum+neonatal) mortality rate for the entire home birth population in J&D to the neonatal only mortality rate for white women in hospital. Also, as it turns out, some of the bad outcomes were in non-white mothers and so an adjustment is needed. All this, and more, has been pointed out in the past, and so it goes Again, please read the link. This has been debated at length for 2 years.
John, Im an engineer, and I would like to tell you a little story. One of my senior colleagues in the area of analytical and experimental thermal hydraulics once told me Russ, as a manager, if you want to get to the bottom of some controversial experimental data and if you get a room full of experimentalists together with differing opinions, at the end of the day you will end up cross-eyed, and have to stand up and get cookie. As I said before, all this lobbing of rocks related to the numbers is irrelevant. Planned home birth is a safe choice and these women deserve access to care. Its as simple as that.
On the subject of releasing data (I hate all this mudslinging), nobody except governmental agencies whose charter is to collect and release data, do in fact release raw data. Universities dont do it (unless they are funded to do just that), industrial companies dont do it, this little marshmallow is just silly. As I understand it, the results will be release along with a proper study.
Hope that helps.
(Currently in Yokohama and not able to give this my full attention)
Those are very thoughtful questions. I cant engage in a multi-post/multi-day discussion just now, but I will try and be responsive here.
The epidemiological analyses that strive to compare similar populations and look for differences have lots of challenges. There are many things that can influence the outcome, including unforeseen complications. In scrutinizing the CPM2000 study, we observe that over 10% of the planned home birth population were Old Order Amish and Mennonite families that do experience an increase in congenital anomalies incompatible with life. We observe that the setting was likely irrelevant for a number of the bad outcomes (e.g. SIDS death, undetected GBS death in hospital, etc). Nevertheless, the outcome is assigned to the home birth group. We also know that known fetal demise would not likely be brought to spontaneous labor and delivery in our standard model of care (affecting the comparison groups). There are many studies on planned, attended Out of Hospital birth, and they tend to all tell the same story that it is really hard to see a difference relative to planned hospital birth for healthy women experiencing normal pregnancies. The objectives of the midwife are to 1) coach the mother to stay as healthy as she can and 2) refer to a physician when the pregnancy is judged not normal.
In a nutshell, there is no comparison group with which to compute 2-sample t-tests and percent differences between planned, attended home birth and planned hospital group for healthy women in the CPM2000 study. The probability of a bad outcome is very low (but it is not zero). Schlenkzas thesis actually showed better outcomes for OOH birth, even for elevated risk factors.
All of this detailed analysis is irrelevant at this point. Planned home birth is a safe choice, but not without risk. The key is that it is attended and that the women who choose this have access to care. To keep it as safe as we can, we need to assure that there are no obstacles to access to care nor a transfer of care when indicated.
Without question, if I were Chief Engineer of maternity care, I would re-engineer the whole thing. I would have a large cadre of midwives attending all women in whatever setting they felt safest. Obstetricians would be engaged and nurtured. I would address the medico-legal environment that is resulting in OBGYNs leaving their practice like lemmings to the sea. This would improve maternal health, improve outcomes, improve the quality of care and reduce costs. For the time being, I will be happy if women who choose to birth out of hospital have access to care.
Welcome to North Carolina! I hope all is well with you up in Massachusetts.
Fortunately, Amy has propagated these same, tired arguments in so many places on the internet that the counter arguments are readily available. I think I will address a few things.
With regard to our excess bad outcomes, the Perinatal Period of Risk Methodology was applied in the analysis of NCs outcomes during the years 2000-2002. The conclusion was that the single biggest contributor to our excess bad outcomes was poor Maternal Health resulting in premature low birth weight babies and that the key interventions should be targeted at addressing this. We know that the Midwives Model of Care is much more effective at promoting health and the frequency of low birth weight babies in the CPM2000 study was half that of comparable women with very conservative assumptions. As such, we absolutely expect improved maternal health and outcomes with improving access to trained midwives in all settings. The states with the best policies and programs related to maternity care, reflected in the best perinatal mortality rates, license and regulate CPMs.
On the subject of safety analysis, Im very fond of Marsden Wagners quote in the BMJ Rapid Responses when the CPM2000 study was published (http://www.bmj.com/cgi/eletters/330/7505/1416):
When a study is published with scientifically valid evidence against an important position of a clinical group, clinicians have two common reactions: ignore the study and hope it goes away; torture the data until it confesses to what they want it to say.
Marsden G. Wagner M.D.,
Former Director of Women's and Children's Health, WHO
The attempts at torturing our analysis basis for safety have failed and a very nice extension to the CPM2000 study can be found here that addresses Amys claims.
Furthermore, with indifferent outcomes for healthy women experiencing normal pregnancies, coupled with fewer premature low birth weight babies and a dramatic reduction in harmful (albeit sometimes life-saving) interventions, planned home birth with CPM can be evaluated as safer.
From our perspective, however, this is not so much about which model is safer, or best. There are some tremendous people working to support women and families who choose to birth in hospitals. The work that has been done thus far clearly shows that planned home birth with a trained midwife is a valid choice and the women who choose it need access to care and it is important that the standards of care are maintained (two key functions of licensing). Should future work observe a difference in outcomes, then it is important to understand that to assure women can make informed decisions. The best way to monitor outcomes is by having licensed and regulated midwives in North Carolina.
Every argument Amy makes is actually an argument that favors licensure and regulation of the midwives that serve this community (that are also the national standard for care providers serving women who birth Out of Hospital).
Someday, with advancements in obstetrical process capability that address the risks associated cesarean section, a comprehensive safety assessment may conclude planned C/S is safer than vaginal birth. I think it would be great to have that kind of capability. Will we then declare women out of process who choose a vaginal birth? Perhaps what we will do is put every obstacle in their way to deny them access to care thats what we do to women who choose to birth at home in North Carolina.
We know what the right answer is. Planned home birth with a CPM is a valid choice and women who choose this deserve normalcy and access to care that is integrated into the healthcare system. We know that licensing is a key step toward assuring women have access to care and that the standards of care are maintained. We know breaking down the barriers in access to midwifery care is a key step in addressing our maternal health and access to care problems.
Trampling liberty is not going to solve our problems.
Legislative Chair, North Carolina Friends of Midwives
Indy Week • 302 E. Pettigrew St., Suite 300, Durham, NC 27701 • phone 919-286-1972 • fax 919-286-4274
RSS Feeds | Powered by Foundation