“Our evidence suggests that mainstream obstetric science follows mainstream obstetric practice. A patient and expectant approach to birth…where all is considered normal until proved otherwise, produces a science that proves intervention to be unnecessary. Alternatively, an aggressive approach to birth…, where birth is regarded as normal only in retrospect, generates a science that demonstrates the need for monitoring and intervention.” — De Vries 2006, p. 2704
The Washington Post sums up a debate that took place at this year’s annual meeting of the American College of Obstetricians & Gynecologists (ACOG) on whether all women should be induced at 39 weeks’ gestation—except it wasn’t really a debate because the debaters, Drs. Lockwood and Norwitz, were on the same side. Both agreed that routinely inducing labor at 39 weeks would reduce stillbirths and decrease the cesarean rate.
Their joint argument was persuasive. Prior to the non-debate, 63% of the audience were opposed, 20% in favor, and 17% were unsure. Afterward, 70% were in favor, 21% were unsure, and only 9% were opposed. Said one obstetrician, “What changed [my opinion] . . . was the really convincing argument that there is nothing to be gained for the baby to stay in beyond 39 weeks. I would personally recommend that any woman with a favorable cervix at 39 weeks be encouraged to get induced.” Her reaction is hardly surprising since what she heard was a sales pitch, not a debate—not that a case for an opposing view can’t be made. Rebecca Dekker, of Evidence-Based Birth , watched a live stream of the session and, according to the Post, she found the presentations to be misleading and the research flawed. I don’t know what specifics she heard, but my reading of the studies and the opinions expressed on 39-week induction line up with hers.
The stakes are high because, as Dekker notes, a policy of routine 39-week induction potentially affects at least 3 million women per year in the U.S. alone. And while Lockwood acknowledged that “expectant mothers should have the final say as to whether to let nature take its course or undertake this potentially salutary but clearly artificial intervention,” few women will refuse 39-week induction when informed by their obstetrician that it prevents stillbirth and has no adverse effects. So let me do what ACOG didn’t and present the “Against” side of the debate.
Does Inducing Labor Decrease the Cesarean Rate?
For decades, the only impediment to reopening the floodgates of “daylight obstetrics,” the practice of routinely inducing labor for obstetric convenience, was the finding that it increased cesarean rates—not that it has been much of an impediment, seeing as nearly half of women who labor have labor induced (Zhang 2010). A flurry of studies published in recent years, however, have concluded that the opposite is true. Since I’ve written in the past deconstructing some of these studies, in the interest of brevity, I’ll summarize their points here and link to my analyses for those interested in the nitty-gritty details. Let’s begin with “Does Inducing Labor Decrease Likelihood of Cesarean Surgery?”, a piece published in the International Doula in 2014.
- Studies comparing cesarean rates with elective induction at term, which ensures that you are isolating the effects of the procedure, consistently find that more women end up with cesareans compared with similar women beginning labor on their own, even after adjusting for factors such as gestational age and birthweight. The effect is especially strong in 1st-time mothers, although it is also seen in women with prior births and no prior cesareans. The use of cervical ripening agents doesn’t eliminate the difference. This makes biological sense because initiating and sustaining effective labor involves a complex cascade of feedback mechanisms that mutually reinforce and limit each other. You would expect simplistic efforts to bypass that process to fail.
- Randomized controlled trials, meaning participants are allocated by chance to one form of treatment or another, of planned induction at 41 weeks or planned expectant management in healthy women free of pregnancy complications report lower cesarean rates in the planned induction group. However, sizeable percentages of women assigned to induction go into labor before they can be induced and vice versa. When analyzed according to what actually happened, cesarean rates are higher in induced women. In addition, in what is by far the largest of such trials (Hannah 1992), investigators combined data from 1st-time mothers with data in women with prior vaginal births, which would mask any effect of induction on cesarean rate since women with prior births would be much less affected.
- A body of studies comparing cesarean rates in women induced at term with women continuing beyond that week have also found lower cesarean rates in the induced population. But 3 studies using this design limited the induction group to healthy women undergoing elective induction while comparison women could have medical complications. Little wonder, then, that cesarean rates were higher in the comparison group. A 4th study confined both groups to low-risk women and found an excess of cesareans at every gestational week in the induced group. Furthermore, comparing outcomes with induction in a given week with outcomes of all women after that week biases results in favor of induction because it leaves out women who began labor spontaneously in that same week. When women induced in a given week are compared with all other women giving birth in that same week or beyond, results flip to favor expectant management.
- Medical-model management confounds all studies of cesarean rate because cesarean rate depends largely on care provider judgment. In support of this, studies show that while stable in in weeks 37 to 40, the cesarean rate leaps upward in week 41 and again in week 42 in low-risk 1st-time mothers. It seems highly unlikely that status changes so radically over a few days. Much more likely is that their care providers’ perception and management changes. In Hannah et al., mentioned above, 26% of ultra-low-risk 1st-time mothers admitted to the hospital in spontaneous labor had cesareans. This tells us something was terribly wrong with their management. Finally, home birth and birth center studies report induction rates ranging from 2-14% and cesarean rates of 4-8% yet equally good outcomes compared with similar women planning hospital birth. That fact negates any argument that high induction and cesarean rates are needed to ensure good outcomes.
Next is a guest blog post I did for Science & Sensibility (2015) entitled “Elective Induction at 40 Weeks? ‘Decision-Based Evidence Making’ Strikes Again.” It analyzed a review of trials of induction vs. expectant management at 40 weeks, not 41. (Note that having succeeded in pushing back the “deliver by” date from 42 to 41 weeks, obstetricians have been lobbying since that time to move it still earlier. The 39-week minimum now being proposed is based on research showing that more babies end up in intensive care with breathing difficulties with earlier delivery, but even that limitation is now under attack [Salemi 2016].)
The review’s title, “Induction of labor at full term in uncomplicated singleton gestations: a systematic review and metaanalysis of randomized controlled trials,” includes all the buzzwords that give the impression of providing incontrovertible evidence: “systematic review,” “metaanalysis,” and “randomized controlled trials”; however, 2 of the 5 trials were deemed inadequate to include in the Cochrane systematic review of elective induction, and a third trial appears only as an abstract. Quality systematic reviews exclude abstracts because they don’t provide enough information to adequately evaluate the study.
This leaves us with the two remaining trials, one published in 2005 and the other in 1975, a time when obstetric practices were quite different, and the national cesarean rate was around 5%, which makes it dubious to include in light of its applicability to current day obstetrics. That being said, neither trial reported an increase in cesareans, but the 2005 trial only induced labor in women with a favorable cervix, which greatly increases the chances of success. Furthermore, half its population was women with prior births, and the cesarean rate for progress delay was only 7%, which means results cannot be generalized to hospitals with higher rates. In the 1975 trial, investigators assigned women to induction at either 40 or 41 weeks. Half the women in the “expectant management” group were induced, and again, the population mixed 1st-time and repeat mothers. Neither of these trials, therefore, makes a compelling case that routine induction has no or a beneficial effect on cesarean rates.
None of these weaknesses and flaws are particularly subtle or obscure, but Lockwood, who was tasked with the “against” side of the debate, and who must have run across these studies in his researches, failed to notice them. Why? Because he and Norwitz start from the premise that, as Norwitz put it, “Nature is a terrible obstetrician,” with its implicit corollary that they, the competent obstetricians, must rescue women from her. When likeminded researchers report findings that fit comfortably with that belief, Lockwood would not be disposed to question them. The worst of it is that they aren’t even aware of their bias. “This is my interpretation of the published literature,” Norwitz says. “This is not a personal opinion based on belief or ideology,” but of course, it is because neither he nor the researchers start from a neutral position.
Does Elective Induction Reduce Stillbirth?
If inducing labor does, in fact, increase the likelihood of cesarean, then the argument for routine induction collapses because, among other harms, cesarean surgery increases the risk of death in women, babies, and most especially, babies in subsequent pregnancies. Ironically, this includes stillbirth. Some of the excess risk comes from complications related to prior cesarean surgery, which include cesarean-scar ectopic pregnancy, placenta accreta, placenta previa, placental abruption, and scar rupture. Indeed, cesarean-scar ectopic pregnancy (the embryo implants within the scar) and early placenta accreta (the placenta implants in the scar, but the embryo is within the uterine cavity [Timor-Tritsch 2012]) are exclusive to cesarean surgery. Prior cesarean is also associated with an excess of unexplained stillbirths. A systematic review reports that 4% of stillbirths and 7% of unexplained stillbirths can be attributed to prior cesarean surgery (population attributable risk) and that women with prior cesarean have a 23% increased risk of subsequent stillbirth compared with women with only prior vaginal birth (O’Neill 2013).
We could stop here, but it is still worthwhile to examine the relationship between gestational age and fetal mortality to gauge the magnitude of the problem. MacDorman (2015) provides us with an analysis of stillbirth by gestational age in the United States. The traditional method of calculating stillbirth rates is to divide the number of stillbirths at a given gestational age by the number of live births plus stillbirths at that same gestational age and multiply the results by 1000. Here are the 2012 stillbirth rates from 37 to 42 or more weeks calculated by this method:
Traditional Stillbirth Rates per 1000 by Gestational Age and Race-Ethnicity
|42 or greater||1.69||1.46||2.19||1.87|
These rates, however, are in all women, including women and babies with medical indications for delivery, babies with congenital or genetic anomalies, multiple gestation, and women with risk factors such as smoking, and, as you can see, race/ethnicity influences rates as well undoubtedly because it serves as a marker for socioeconomic deprivation and therefore the ills connected with it. In evidence of this, a study of factors associated with stillbirth in high-income countries found that the population attributable risk of low socioeconomic status was 9% (Flenady 2011). Women and babies free of risk factors, the ones ACOG is talking about inducing, would be at lower risk of stillbirth than these numbers indicate.
That’s not all, though. MacDorman and colleagues argue that the traditional method is not the best because it is vulnerable to shifts in the distribution of births by gestational week. A more useful statistic, and one more relevant to the issue of curtailing vs. continuing pregnancy, is the prospective stillbirth rate. It is calculated by dividing the number of stillbirths at a given gestational age by the number of live births and stillbirths at that gestational age or greater and multiplying by 1000. This is preferable because the denominator is the number of women who are pregnant and therefore at risk of stillbirth with ongoing pregnancy. Judging from a graph in the MacDorman paper, the prospective risk for all women at 39 weeks is 0.4 per 1000, or 4 per 10,000, rising steadily to a little below 0.8 per 1000, or 8 per 10,000, at 42 or more weeks. As before, non-Hispanic white women run numbers below the rates overall and numbers in black women run considerably above it, and again, these are rates for everyone, not just women with no factors predisposing to stillbirth.
A couple of studies by the same group of investigators use California data to analyze the risk of stillbirth and infant death according to each additional week of expectant management. One study stratifies results by maternal age and the other by race/ethnicity. Both, like MacDorman (2015), calculate stillbirth rates per week using ongoing pregnancies as the denominator. Investigators in these studies exclude multiple gestation, diabetes, chronic hypertension, and congenital anomalies and genetic abnormalities, although they don’t, exclude prior cesarean delivery or pre-eclampsia. As with MacDorman, rates are low and vary according to race/ethnicity. In white women, rates per 10,000 ongoing pregnancies increase from 4 per 10,000 in week 39 to 5 in week 40 to 8.5 in week 41 to 22 at 42 completed weeks or more (Rosenstein 2014). Black women fare worse with rates at 8 per 10,000 in weeks 39 and 40, rising to 15.5 in week 41 and to 32 at 42 or more completed weeks. Similarly, in women younger than age 35, rates are 4 per 10,000 ongoing pregnancies in week 39, rising to 7 in week 40, 8.5 in week 41, and 28 after 42 completed weeks while for women age 35 or older, rates increase from 5 per 10,000 in week 39 to 10 in week 40 to 15 in week 41 and to 32.5 at week 42 or more (Page 2013).
In other words, we aren’t talking large numbers of stillbirths even in women overall—let alone healthy women with uncomplicated pregnancies—until 42 weeks or more, but degree of risk isn’t really the issue. The thinking behind cutting pregnancy short at 39 weeks is that it eliminates stillbirths beyond that date. Every stillbirth is a tragedy, the reasoning goes, and, however rare, the odds are 100% if it happens to you. This, however, is a kind of tunnel vision that only counts the results of not intervening but not the results of intervening. It counts the woman who has an unexplained stillbirth in week 40 but not the healthy mother and child who die as a result of an elective induction at 41 weeks . And if, as I argue is the case, inducing labor increases cesareans, it doesn’t count the woman who loses her baby to cesarean-related complications in a subsequent pregnancy. Shouldn’t the “rare but 100% if it happens to you” argument apply to these women too?
What Are the Flaws in the Medical-Model Approach?
As stated, medical-model thinking imposes tunnel vision, and that has consequences. For one thing, it focuses on the harms of expectant management and ignores or discounts the harms of induction. In addition to excess cesareans, these include increased likelihood of postpartum hemorrhage, increased likelihood of uterine hyperstimulation and its follow-on effects on the baby, and increased possibility for rare but severe outcomes such as amniotic fluid embolism and rupture of an unscarred uterus. It also fails to consider harms that don’t fall within the canon of adverse effects accepted within the medical model but which may be profound all the same. These include interference with the interplay of hormones and other factors that orchestrate the onset of labor, labor progress, birth, breastfeeding, and the formation of attachment between mother and baby. They also include the psychological fallout of inculcating that women’s bodies are lemons, incapable of safely carrying and birthing their babies on their own timetable. In addition, induction increases pain, which leads to increased use of epidural anesthesia and therefore its potential harms.
Indeed, the willingness to expose healthy women and babies to the potential harms of induction could well be seen as a violation of the fundamental principle of medical treatment: Primum non nocere: “First do no harm.” It is hard to think of another circumstance in which doctors advocate derailing a physiologic process that is proceeding normally in a healthy person on the remote chance that doing so will prevent an extremely unlikely event, and it arises from the medical-model belief that women are unreliable incubators from which babies must be rescued at the earliest feasible moment:
“Once delivery occurs, the ongoing risk of stillbirth disappears.” — Caughey 2006, p. 700
The pervasiveness of the medical-model approach distorts the research as well. Medical-model thinkers point to rising induction and cesarean surgery rates with advancing pregnancy at term as a rationale for early induction, but as we have seen in studies by adherents of the physiologic approach, this isn’t inevitable. It’s a self-fulfilling prophecy. However, because medical-model researchers are conducting studies in medical-model environments, what is actually an echo chamber appears to be objective reality to those inside the room.
Medical-model beliefs also give rise to illogical thinking. For example, the investigators who analyzed California data chose infant mortality, the risk of death in the year after birth, instead of the more usual neonatal mortality in their calculation of the effect of an additional week of expectant management (Page 2013; Rosenstein 2014). They did this on the grounds that birth-related issues such as aberrant fetal growth can have a long tail, but the point of the study was to determine if there were benefits to inducing labor in uncomplicated pregnancies. The benefits of inducing labor when the fetus is undersized is a different research question. Meanwhile, using death within a year introduces a large number of confounding factors having nothing to do with timing of delivery while at the same time tilting the playing field in favor of intervening because it inflates the absolute numbers of deaths, especially in disadvantaged populations. In another example, Flenady and colleagues (2011) set out to evaluate risk factors for stillbirth in high-income countries with special attention to modifiable factors. They even selected “previous cesarean section” as a factor to investigate, yet while they report on a multiplicity of factors, previous cesarean isn’t one of them despite making a considerable contribution to potentially preventable stillbirths (O’Neill 2013). These logic gaps are invisible to researchers, reviewers, and readers, however, because their biases render them unable to see them.
Most insidious, medical-model thinking justifies what benefits doctors and hospitals, as these quotes make clear:
“Although improvements in Intermountain’s appropriate elective induction rates [elective induction rate fell from 28% to 2%] saved the citizens of Utah more than $50 million per year through reduced payments, Intermountain’s costs fell by only about $41 million. Intermountain thus lost more than $9 million per year in operating margins.” — James 2011, p. 1190
“The best part about it is that you can block-schedule your nurses so that you have enough on hand. . . [I]f we start our inductions at 7 a.m., we know that we’re going to have X number of patients in labor being admitted by 4 p.m. That’s helped our hospital tremendously, . . . [Cytotec is] a great agent. It works very, very efficiently. . . . And it’s ungodly inexpensive: 27 cents per tablet.” — Jancin 2004
“‘Scheduling an induction can make everyone’s life easier,’ [Dr. Leveno] said. . . . ‘I am not capable of constantly doing my best work in the middle of the night.’” — Villarosa 2002
This necessarily sets up intense resistance, unconscious or not, to taking an alternate approach.
Finally, medical-model beliefs result in a “single arrow in the quiver” mentality. Imagine a conference that has come together to deliberate the problem of stillbirth in which, contrary to the ACOG conference, all stakeholders are present and all are committed to a physiologic approach to care and to minimizing the use of medical intervention whenever possible. All present understand that measures aimed at improving health beat medical intervention every time because unlike medical intervention, they have no downside. How would such a hypothetical conference proceed? They would examine the underlying causes of stillbirth across the full spectrum of gestational age and evolve a comprehensive, multipronged plan for redressing them. Even when labor induction was indicated, they would spell out the circumstances when it was inadvisable and hedge it with strategies to minimize its harms. They would also develop balanced, objective, patient-education materials to assist women in making informed decisions.
They would not, as the ACOG conference attendees did, end with recommending a blanket policy of inducing every woman, regardless of health status or perceived risk, at week 39 and leave it at that. They would not throw in, almost as an afterthought and the only caveat, likely to be ignored by many of their number, that inducing 1st-time mothers with a long, thick, closed cervix might not be a good idea. And they certainly wouldn’t consider that a woman’s agreement to induction at 39 weeks, after being told that it was harmless and could possibly save her baby’s life, constituted informed consent.
In the end, it comes down to “Choose your own adventure.” You must decide whether you agree that “Nature is a terrible obstetrician” or whether you agree with David Stewart (1998), editor of The Five Standards for Safe Childbearing: “When nature does work, it cannot be improved. Technology does not enhance a natural process that is working. It can only mar or destroy it.”
If the latter, choose a caregiver who prefers to support and facilitate the natural process and who refrains from medical intervention except when patience and lesser measures have failed to resolve the problem or there is an unequivocal need for it.
If the former, then here are some considerations and tips to minimize the potential harms of inducing labor:
- Wait for cervical ripening. Inducing before the cervix is ready for labor greatly increases the chances of cesarean if this is your first baby, and some studies find the odds are greater even if you have had babies vaginally before (Goer 2012a). Cervical ripening agents do not eliminate the excess risk (Goer 2012a).
- If you have had a prior cesarean, your best bet is likely to be waiting for labor to start on its own. Inducing labor in women with prior cesareans generally shows that it both increases the likelihood of symptomatic scar separation and reduces the likelihood of vaginal birth (Goer 2012b).
- Find out your care provider’s parameters for making progress. You want to make sure you will be given adequate time. Induced labors may take longer, but longer labors don’t result in worse maternal or newborn outcomes (Goer 2012a), and more time gives you a better shot at vaginal birth.
- Refuse having the bag of waters broken before you are established in active, progressive labor (around 6 cm dilated). If membranes are intact and the induction isn’t working, you can stop, go home, and try another day. Once membranes are ruptured, you are committed to delivery one way or another.
Regarding labor induction management, negotiate for these ahead of time because they may not be standard procedure.
- Request that your doctor or midwife use a physiologic oxytocin (Pitocin, Syntocinon) regimen. This means dosage levels in line with what your body would produce naturally with increases no closer together than 30 minutes so that any given dose can reach its full effect before determining if a stronger dose is needed. High-dose/short-interval oxytocin protocols increase the odds of oxytocin-related complications without decreasing cesareans (Goer 2012a).
- Request having the oxytocin drip turned off once you are in progressive labor. In many cases your body will take over and labor will continue under its own steam. If it doesn’t, the drip can always be turned on again (Goer 2012a).
- Request the ability to take breaks from continuous fetal monitoring, a practice supported by the Society of Obstetricians & Gynaecologists of Canada (SOGC 2020). Provided the oxytocin dose is stable, and the baby isn’t experiencing any problems, this will allow you to get up, move around, etc. which can help labor progress.
Finally, if you want a physiologic approach but only medical-model practitioners are available to you, you may have a difficult decision. It is always your right to refuse treatment, but some caregivers don’t respect that right and will pressure you to comply, which, as you may imagine, can be extremely unpleasant and stressful. It is also possible that you may be better off with elective induction. A care provider who believes that delay increases the baby’s danger or allows the baby to grow too large to be born vaginally may be more inclined to think that a cesarean is necessary with longer pregnancy. You may need to feel things out. If you decide that you are better off agreeing to an elective induction, I recommend negotiating for it along the lines laid out above.
Flenady V, Koopmans L, Middleton P, et al. Major risk factors for stillbirth in high-income countries: a systematic review and meta-analysis. Lancet 2011;377(9774):1331-40.
Goer H. Does inducing labor increase the likelihood of cesarean surgery? International Doula 2014;22(2):8-11.
Goer H. Induction of labor: Patience is a virtue. In: Goer H, Romano A, eds. Optimal Care in Childbirth: The Case for a Physiologic Approach. Seattle, WA: Classic Day Publishing; 2012a.
Goer H. The case against elective repeat cesarean. In: Goer H., Romano A., eds. Optimal Care in Childbirth: The Case for a Physiologic Approach. Seattle, WA: Classic Day Publishing; 2012b.
Hannah ME, Hannah WJ, Hellmann J, et al. Induction of labor as compared with serial antenatal monitoring in post-term pregnancy. A randomized controlled trial. The Canadian Multicenter Post-term Pregnancy Trial Group. N Engl J Med 1992;326(24):1587-92.
MacDorman MF, Reddy UM, Silver RM. Trends in Stillbirth by Gestational Age in the United States, 2006-2012. Obstet Gynecol 2015;126(6):1146-50.
O’Neill SM, Kearney PM, Kenny LC, et al. Caesarean delivery and subsequent stillbirth or miscarriage: systematic review and meta-analysis. PLoS One 2013;8(1):e54588.
Page JM, Snowden JM, Cheng YW, et al. The risk of stillbirth and infant death by each additional week of expectant management stratified by maternal age. Am J Obstet Gynecol 2013;209(4):375 e1-7.
Rosenstein MG, Snowden JM, Cheng YW, et al. The mortality risk of expectant management compared with delivery stratified by gestational age and race and ethnicity. Am J Obstet Gynecol 2014;211(6):660 e1-8.
Salemi JL, Pathak EB, Salihu HM. Infant outcomes after elective early-term delivery compared with expectant management. Obstet Gynecol 2016;127(4):657-66.
Society of Obstetricians & Gynecologists of Canada. No. 396-Fetal Health Surveillance: Intrapartum Consensus Guideline. J Obstet Gynaecol Can 2020;42(3):316-48 e9.
Timor-Tritsch IE, Monteagudo A. Unforeseen consequences of the increasing rate of cesarean deliveries: early placenta accreta and cesarean scar pregnancy. A review. Am J Obstet Gynecol 2012;207(1):14-29.
Zhang J, Troendle J, Reddy UM, et al. Contemporary cesarean delivery practice in the United States. Am J Obstet Gynecol 2010;203(4):326 e1- e10.
Caughey AB, Nicholson JM, Cheng YW, et al. Induction of labor and cesarean delivery by gestational age. Am J Obstet Gynecol 2006;195(3):700-5.
De Vries R, Lemmens T. The social and cultural shaping of medical evidence: case studies from pharmaceutical research and obstetric science. Soc Sci Med 2006;62(11):2694-706.
James BC, Savitz LA. How intermountain trimmed health care costs through robust quality improvement efforts. Health Aff (Millwood) 2011;30(6):1185-91.
Jancin B. Tips on labor induction using oral misoprostol. Ob Gyn News April, 2004:17.
Stewart D. The Five Standards for Safe Childbearing. 4th ed. Marble Hill, MO: NAPSAC; 1998.
Villarosa L. Making an appointment with the stork. New York Times Jun 23, 2002.