WHEN LESS IS MORE: SHIFTING RISK MANAGEMENT IN AMERICAN CHILDBIRTH

Abstract

As maternal mortality increases in the United States, birth providers and policymakers are seeking new solutions to address what scholars have called the “C-section epidemic.” Hospital cesarean rates vary tremendously, from 7 to 70 percent of all births. Based on in-depth, semi-structured interviews with 47 obstetricians and family physicians in the United States, I explore one reason for this variation: differences in how physicians perceive and manage risk in American obstetrics. While the dominant model of risk management encourages high levels of intervention and monitoring, I argue that a significant portion of physicians are concerned about high intervention rates in childbirth and are working to reduce cesarean rates and/or the use of monitoring technologies like continuous fetal heart rate monitors. Unlike prior theories of biomedicalization, which suggest that health risks are managed through increased monitoring and intervention, I find that many physicians are resisting this model of risk management by ordering fewer interventions and collecting less information about their patients. These providers acknowledge that interventions designed to mitigate risks may only provide an illusion of control, rather than an actual mastery of risks. By limiting interventions, providers may lose this illusion of control but also mitigate the iatrogenic effects of intervention and continuous monitoring. This alternative approach to risk management is growing in many medical fields and deserves more attention from medical sociologists.

INTRODUCTION

The United States is one of only a few countries in the world, and the only developed nation, where maternal mortality rates are climbing (GBD 2015 Maternal Mortality Collaborators, 2016; MacDorman, Declercq, Cabral, & Morton, 2016). As a result, both American obstetrics and the broader US healthcare system have faced intense public scrutiny over the past few years. In 2017, a popular series of articles from ProPublica and National Public Radio argued that women had been routinely ignored in the childbirth process, and found that the vast majority of federal and state funding for maternal and child health focuses on improving outcomes for infants rather than mothers (Martin & Montagne, 2017b). Major childbirth organizations like the American College of Obstetricians and Gynecologists (ACOG) and the American College of Nurse-Midwives (ACNM) have released statements recognizing maternal mortality as a crisis (Christopher, 2018; The American College of Obstetricians & Gynecologists, 2015; The Joint Commission, 2010).

Researchers have identified many reasons why women are dying during childbirth in the United States, including obstetric hemorrhage, severe hypertension, venous thromboembolism, primary cesareans, and racial disparities during pregnancy (Creanga et al., 2014; The American College of Obstetricians & Gynecologists, 2015). Racism and racial disparities are particularly apparent – black women in the United States are three to four times more likely than white women to die during childbirth (Martin & Montagne, 2017a). Rising cesarean rates are also a serious concern, and one of the main foci of this chapter. Cesarean rates have risen over 50 percent in the last 15 years, and now account for almost one-third of all births (Menacker & Hamilton, 2010). These high rates have been linked to increased maternal morbidity and mortality (Caughey, Cahill, Guise, & Rouse, 2014). Sociologists of childbirth have written extensively about why intervention rates are so high in American childbirth. Morris (2013), for example, argues that birth providers are incentivized to prioritize the health of the fetus over the mother:

It seems easier and less risky to the provider to perform a C-section. The question that should be asked is whether a C-section is easier and less risky for the woman […] women must demonstrate that they are capable of having a vaginal birth that will not put their baby in peril or their maternity providers’ livelihood at risk. (pp. 106–108)

American hospitals and birth providers want to reduce their legal liability, and are incentivized to perform more interventions. But while the American childbirth system is designed to encourage interventions, there is significant variation in intervention rates across the country.

In this chapter, I explore one reason for this variation: differences in how physicians perceive and manage risk in American obstetrics. I argue that a significant portion of physicians are concerned about high intervention rates in childbirth and are working to reduce cesarean rates and/or the use of monitoring technologies like continuous fetal heart rate monitors. While prior theories of biomedicalization suggest that health risks are managed through increased monitoring and interventions (Clarke, Shim, Mamo, Fosket, & Fishman, 2003), I find that many physicians are resisting this model of risk management. Instead, these obstetricians and family physicians are intentionally choosing to manage the health risks of their patients by ordering fewer interventions and collecting less information. These providers acknowledge that interventions designed to mitigate risks may only provide an illusion of control, rather than an actual mastery of risks. By limiting interventions, providers may lose this illusion of control but also mitigate the iatrogenic effects of intervention and continuous monitoring. This alternative approach to risk management is growing in many medical fields and deserves more attention from medical sociologists.

DATA AND METHODS

The data presented in this chapter come from semi-structured, in-depth interviews with obstetricians (37 participants) and family physicians (10 participants) practicing in American hospitals. I limit my analysis to physicians in hospitals because this is where the majority of births occur and because other researchers have previously explored differences between obstetrics and home birth or midwifery birth cultures (e.g., Rothman, 1982). These data are part of a larger research project exploring practice variation in contemporary American childbirth. My sample thus includes participants from five different states with a range of childbirth cultures related to average malpractice costs, cesarean rates, midwifery integration, and out-of-hospital births: Louisiana, New Jersey, Oregon, Wisconsin, and Massachusetts. Participants had a range of institutional affiliations, including small private practices, large teaching hospitals, and small community hospitals. Approximately half of the participants were female, and half were male. Approximately one-quarter of the physicians identified as a racial minority. Interviews ranged from 30 minutes to 2 hours, with most interviews lasting about 1 hour. I transcribed the interviews and analyzed and coded them using NVivo Version 9 software. I used an iterative approach to look for key themes surrounding the risk perceptions and practices of birth providers. Among many topics, I asked providers to describe the major risks they perceived during the birth process, and how they managed each of these risks. Discussions of these risks and management strategies serve as the main source of data for this chapter.

I address the following research questions: (1) How do physicians perceive and manage risks during childbirth? (2) How do physicians assess whether interventions will reduce or increase health risks to women and fetuses? In the next section, I first provide background on existing practice variation and birth cultures in American childbirth. I then detail how my research is informed by prior sociological literature on childbirth, risk, knowledge, and ignorance. In the main data analysis, I examine how physicians discuss what it means to be a risk averse or risk tolerant provider, how risks should be monitored, and what types of practices are seen as particularly risky. I focus specifically on different ways that providers frame the riskiness of interventions like cesareans and monitoring technologies like continuous fetal monitors. Finally, I conclude with thoughts on how this research updates prior theories on childbirth and risk in medical sociology.

PRACTICE VARIATION IN AMERICAN CHILDBIRTH

Birth practices, including cesareans, vary greatly across the United States. In Utah, cesarean rates hover around 22 percent of births. In Mississippi, they account for over 38 percent of births. States in the West tend to have lower cesarean rates, and states in the South and East tend to have higher rates (Martin, Hamilton, Osterman, Driscoll, & Drake, 2018). Cesarean rates vary even more by hospital than by state: from 7 to 70 percent of all births depending on the hospital (Kozhimannil, Law, & Virnig, 2013). When comparing only low-risk births, rates varied fifteen-fold, from 2.4 percent to 36.5 percent of all births (Kozhimannil et al., 2013). This means that the hospital can be a significant predictor of whether or not a woman gives birth vaginally or by cesarean (Haelle, 2018).

This variation in mode of delivery has gained significant public attention. In 2016, Consumer Reports began rating over 1,200 American hospitals based on their cesarean rates. Hospitals with lower cesarean rates were ranked higher. In a corresponding report, the organization said:

A number of factors can increase a woman’s chance of having a C-section […] But the single biggest variable may be where a woman chooses for delivering her baby (Consumer Reports, 2016).

The report warned consumers that cesareans are often overused, with health risks for both women and babies. It suggested that consumers should take a proactive role in protecting their health by researching hospital cesarean rates in advance. Other popular news sources have similar concerns about cesarean variation, like this article from the New York Times:

What’s likely to be the biggest influence on whether you will have a C-section? (A) Your personal wishes (B) Your choice of hospital (C) Your baby’s weight (D) Your baby’s heart rate in labor (E) The progress of your labor. The answer is B (Rosenberg, 2016).

Physicians I interviewed expressed worry about this variation even in places where cesarean rates only vary by small amounts between hospitals:

[Even small differences] are big enough. I mean, a five percent difference, those are real people getting potentially unnecessary surgery and like, a place that delivers a couple thousand people every year, those are hundreds of people in one city. So I think those differences are real […] (Boston obstetrician).

The increased concern about cesarean variation reflects more general unease about practice differences across the country (Timmermans & Berg, 2003). While cesareans may be the most visible marker of practice variation in childbirth, my interviews with physicians offer many examples of differences, even for seemingly simple decisions like when to admit a woman to the hospital:

If someone thinks they are in labor, we tend to admit people right away. If they broke their water, we bring them in right away. I err more on the side of sending patients over [to the hospital]. (New Orleans obstetrician)

The less [low-risk women] can be monitored, the better. I dissuade low-risk patients from coming into the hospital until they absolutely have to come, because I don’t want them to be on the [continuous electronic fetal] monitor longer than they need to be. (Milwaukee family physician)

Researchers have tested many hypotheses about the causes of this practice variation, including differences in the health of patient populations, patient preferences, malpractice laws, professional training, and institutional environment (Dubay, Kaestner, & Waidmann, 1999; Kozhimannil et al., 2013; Morris, 2013; Rothman, 1982; Vadnais et al., 2017; Yang, Mello, Subramanian, & Studdert, 2009). While all of these factors likely influence birth practices in the United States, this chapter focuses on another source of practice variation: risk cultures. While many researchers have called for more data on how culture impacts practice variation in obstetrics (e.g., Marcus, 2017; Moore, 2005; Rydahl, Declercq, Juhl, & Maimburg, 2019), few have actually examined variation in how risk cultures are conceptualized and enacted in American obstetrics (for an example of a quantitative approach to measuring variation in culture, see Plough et al., 2017).

MULTIPLE RISK CULTURES

American obstetrics is one of the few medical fields where two robust models of health risk operate simultaneously. The first and most predominant model of risk values early testing, monitoring, and intervention (Clarke et al., 2003; Morris, 2013). This approach can be traced back at least to the beginning of obstetric intervention in American birth. Joseph DeLee, a prominent early twentieth-century obstetrician, argued that birth was an inherently dangerous event necessitating high levels of obstetric intervention. In the first issue of the American Journal of Obstetrics and Gynecology, he noted that labor was:

A pathological process […] only a small minority of women escape damage during labor […] So frequent are these bad effects, that I have often wondered whether Nature did not deliberately intend women should be used up in the process of reproduction, in a manner analogous to that of salmon, which dies after spawning? (DeLee, 1920, pp. 40–41)

Barbara Katz Rothman labels this active management of birth the “medical model,” where technology and surgical intervention is common (Rothman, 1982). This model is usually associated professionally with obstetricians and spatially with hospitals. Many of my interview participants said that American labor floors are designed to be high intervention spaces:

In childbirth, every single baby gets delivered in an [intensive care unit (ICU)]. We don’t think about labor floors as ICUs, but everything that defines an ICU occurs on a labor floor. We have one-to-one nursing, we have telemetry or monitoring of heart rates, we have medications we have to titrate minute-by-minute. The only difference between a cardiac ICU and a labor floor is that we have [operating rooms (ORs)] attached, so we are actually a higher treatment intensity facility than even a cardiac ICU. That’s how normal moms get delivered. So, of course, we overdo it, because I mean, if you have scalpels around, you are probably going to use them. (Boston obstetrician)

Despite incentives to intervene early and often in childbirth, this is not the only method of risk management present on American labor floors.

The second risk culture, which views medical intervention in birth as potentially harmful, is equally old. In 1933, the New York Academy of Medicine noted:

The increase in the use of instrumentation [in labor] brings with it an increased hazard. Clearly a reduction of the [maternal] mortality rate can be achieved through a reduction in operative interference. (New York Academy of Medicine, 1933, p. 127)

This language is strikingly similar to recent criticism of high cesarean rates in the United States. Following feminist and women’s health movements in the late 1970s and 1980s, “natural” birth and anti-interventionist movements gained momentum both publicly and academically (Kline, 2010; Michaels, 2014; Wolf, 2012). These movements coincided with a resurgence of American midwives following their relative absence in prior decades. Thus, the risk culture that advocates passive management is generally referred to as the midwifery model (Rothman, 1982). As I will argue throughout this chapter, this model of risk management valuing fewer interventions is not found solely in midwifery. Obstetricians and family physicians can also adopt this approach. The perseverance of both risk cultures hints at the difficulties of consensus-building in childbirth and the never-ending realignment and coproduction of medical and social knowledge (Jasanoff, 2004).

Prior sociological literature on birth cultures tends to focus on either cross-national comparison or on the differences between the midwifery and medical models of care. Cross-national comparative studies are important because they highlight the ways that birth is constructed in different social contexts (e.g., Jordan & Davis-Floyd, 1993). These studies have demonstrated that childbirth in the United States is particularly medicalized, placing high value on technology and markets (De Vries, Benoit, van Teijlingen, & Wrede, 2001). Cross-national comparisons allow us to critically examine practices in the United States and to uncover the contingent histories of American birth. More recently, Theresa Morris (2013) describes how American childbirth is designed to avoid unpredictable outcomes through planning and conservative decision-making, in an effort to have “perfect births.” Studies like these highlight the ways that American institutions shape the way birth cultures are enacted, practiced, and supported through health policies.

Another strand of research in the sociology of childbirth focuses on the differences between midwifery and medical models of care in the United States (e.g., Rothman, 1982). In particular, scholars have emphasized differences in the way that birth is conceptualized, managed, and practiced in the hospital versus at home. In the hospital, Rothman and colleagues argue that:

much of the battle to maintain personal power, autonomy, and control may well be lost long before the birth, as the fetus itself is constructed as a needy patient and the woman as little more than a maternal barrier to its care. (Simonds, Rothman, & Norman, 2007, p. xxii)

They argue that the most important way for women to regain choice and control over their birth is to change the setting and move out of the hospital:

Setting – place, location – counts. The differences between medical and midwifery models of birth are not just about ‘attitudes’ and not just a set of guidelines for practice. Different bodies of knowledge are produced in different settings. (Simonds et al., 2007, p. xx)

My interviews with maternity care providers confirm that there are significant differences in epistemology, ideology, and practice between the American midwifery and physician communities, and between home and hospital births. But since this variation has been well documented, I focus instead on differences among physicians within and across hospitals. A significant number of obstetricians and family physicians I interviewed, especially in places with low intervention rates like Portland, Oregon, have adopted language from the midwifery model, even though they practice in hospitals. It is important to document different risk management philosophies and practices within physician communities, especially because I find that many physicians believe that they can promote the health of women and fetuses by ordering fewer interventions and less continuous fetal monitoring. My data suggest that birth cultures may not be as spatially or professionally divided as they have been reported in the past.

THE SOCIOLOGY OF RISK AND THE POLITICS OF KNOWLEDGE

In biomedicine, managing health has increasingly come to mean managing risk. Rather than simply treating illness and disease, biomedicine now treats risks themselves as legitimate health problems requiring surveillance and treatment. Maintaining health is viewed as an individual moral responsibility, necessitating frequent monitoring and intervention (Clarke et al., 2003). Because everyone has the potential to be “at risk” of becoming ill, we are all subject to health discourses, technologies, services, and procedures. In childbirth, the highly-litigious American health system tends to value physicians who do “all they can” to mitigate risk and prevent poor outcomes (Dreger, 2012). But recently, a growing number of medical institutions have suggested that the treatment of risk can lead to unnecessary medical intervention and possible harm (ABIM Foundation, 2019). Clinical guidelines, for example, now suggest that physicians should no longer routinely place stents in patients with stable coronary artery disease, because stents do not decrease a patient’s risk of related health problems (Epstein, 2017). Many medical institutions are also calling for less frequent health monitoring, as evidenced by shifting guidelines for mammograms, pap smears, and prostate exams (e.g., Siu & U.S. Preventive Services Task Force, 2016). While concerns about overtreatment and overmonitoring in medicine have grown particularly over the past ten years (ABIM Foundation, 2019), these concerns have a much longer history in American childbirth (e.g., Enkin, Keirse, & Chalmers, 1989).

A related emphasis on “intentional non-knowing” has started to garner attention in the social sciences. In the 1990s, Robert Proctor popularized the field of “agnotology,” which studies the cultural production of ignorance. Proctor uses cases such as the tobacco industry’s campaign to manufacture doubt about the health effects of cigarettes to impart the idea that the study of “how or why we don’t know” is just as important as the epistemological question of “how we know” (Proctor & Schiebinger, 2008). Other scholars in the sociology of ignorance have highlighted how ignorance can be an organizational resource (e.g., Gross & McGoey, 2015). Ignorance becomes an organizational resource because:

it is distinguishable from deception or the suppression of data by virtue of the fact that unsettling knowledge is thwarted from emerging in the first place, making it difficult to hold individuals legally liable for knowledge they can claim to have never possessed. (McGoey, 2012, p. 559)

Nearly all studies of ignorance and non-knowing focus on deception. These negative cases are important to recognize but do not tell the whole story about how we use non-knowledge as a resource. Only a small number of scholars have pointed to positive uses of intentional non-knowledge (e.g., Beck & Wehling, 2012). In this chapter, I will show that some physicians view intentional non-knowledge as a moral imperative, as it can lead to better health outcomes for mothers and fetuses. This positive revaluation of intentional non-knowing has been relatively ignored by scholars of knowledge, risk, and ignorance, but deserves more attention as it can help us understand how risk societies are resisting the pull towards increased surveillance and testing. This trend to manage risk through ordering fewer interventions and less monitoring conflicts with dominant accounts of risk management in previous theories of biomedicine (Clarke et al., 2003). It also provides evidence of a growing risk culture that does not behave in the way Ulrich Beck or Anthony Giddens would have expected in their initial theory of risk society, which posits that modern societies manage risk through increased reliance on scientific instruments and expertise (Beck, 1992; Giddens, 1990; Giddens, 1999).

In the next three sections, I present interview data on how obstetricians and family physicians discuss what it means to be a risk averse or risk tolerant provider, how risks should be monitored, and what types of practices are seen as particularly risky. I focus specifically on different ways that providers frame the riskiness of interventions like cesareans and monitoring technologies like continuous fetal monitors.

WHAT IS RISK AVERSION AND RISK TOLERANCE?

When I asked physicians to discuss the major risks they perceived in childbirth, many of them discussed their levels of risk aversion or tolerance. This concept is initially confusing, because the type of risks one is averse to or tolerant of is unclear. When providers mention that their practice style is either “risk averse” or “risk tolerant,” to which risks are they referring? Which bodies or interests are they prioritizing? Childbirth differs from many other types of medicine because providers must balance risks to both women and their fetuses, even in scenarios where the interests of these two patients may compete (often referred to as maternal-fetal conflict). Accounting for and managing the risks of two patients at once is what drives many obstetricians to the field. As one obstetrician explained:

I was really interested in obstetrics – managing two patients, not just the mom, but the baby, too. That relationship changes constantly and everything is dynamic. Should we keep the mom pregnant? Should we deliver? That’s the challenge of this relationship between mom and baby. (Boston obstetrician)

While the separation of mother and fetus into two individual patients is a fairly new phenomenon, a large body of literature in medicine, law, and social science has examined the consequences of how maternal-fetal conflicts are enacted in various contexts (e.g., Casper, 1998; Chavkin, 1992; Markens, Browner, & Press, 1997; Townsend, 2012). Importantly, most have found that the health of the fetus is prioritized over the interests of the mother (e.g., Armstrong, 2003). While risk management in childbirth is often framed as a balance between risks for mothers and fetuses, deeper analysis shows that risks to fetuses and providers are more salient than risks to mothers. Below, are two examples from obstetricians in New Orleans and New Jersey:

The reason you do a C-section is never for the mom really, it’s to prevent something bad from happening to the baby. And we aren’t the best at predicting when the baby is in trouble. If we could know 100 percent for sure it would be a much easier job, right? A lot of things you have to do because you are worried that something bad might happen, and then you get the baby out and things were fine, but you never know what would have happened if you didn’t do that. (New Orleans obstetrician)

Does this patient need a C-section? No, there are a lot of babies that are going to be fine. But you need to watch. Is the ultrasound 100 percent accurate? No, it’s an estimate. A lot of times we think babies are big and they are totally average. But this is what we have and you have to do the best you can. I didn’t go into this field to do any harm. I’m here to try my best and I’m trying to make the best decisions I can with the information I have. (New Jersey obstetrician, emphasis in original)

Of course, there are plenty of cases where maternal status indicates cesarean birth. But physicians almost always discussed cesarean births as a way to prevent injury to the fetus. Physicians used the phrase “risk averse” only to describe people who liked to intervene early to protect the fetus. A Boston obstetrician discussed risk aversion when describing how she decides when it is appropriate to intervene in a delivery:

It’s hard to decide when to deliver someone, or induce someone. We try to follow guidelines somewhat, and they are nice to have, but it’s not always clear cut. You have to use your best judgment and try to integrate that with how the patient feels. Like, if you have a patient who very much doesn’t want to be induced, you can stretch it a little to prolong things. Or if you have a patient who is very risk averse, you can push a little bit in the other direction. Within a range, I try to match what the patient wants with what I am comfortable with as a provider. (Boston obstetrician, emphasis added)

This obstetrician sees value both in intervention and non-intervention, but assumes that a “risk averse” patient would prefer to intervene early. I heard many similar statements in interviews, like “Most patients are risk averse. They want to have a C-section.” This definition of risk aversion prioritizes potential dangers for the fetus over iatrogenic effects of intervention for the mother. The main reason to induce a woman who has not starting laboring naturally is increased medical risk to the fetus, most notably stillbirth. But it is equally plausible that a “risk averse” woman would not want to induce her labor, since inductions can lead to cascades of interventions, and potential harm from those interventions, later in labor (Rossignol, Chaillet, Boughrassa, & Moutquin, 2014). In this case, and many others in my sample, physicians and mothers described as “risk averse” are those who prioritize the absolute healthiness of the fetus. While many physicians described themselves as risk averse, they also frequently referred to laboring women as risk averse. Thus, some providers seemed to shift responsibility and accountability for the decision to have a cesarean birth to the laboring woman.

At the same time, physicians also repeatedly referred to professional risks for themselves as providers:

There hasn’t been, to my knowledge, the case of someone getting sued because they did the cesarean too soon […] If you are a risk-averse medical provider, (you) do the C-section. (Portland obstetrician)

This was true even for obstetricians who said they placed high priority on vaginal births as opposed to cesareans:

There are varying degrees of risk aversion. I’m actually fairly risk averse, I would say, for someone who counsels her patients pretty thoroughly and allows them to make their own decisions. I don’t want to get sued. I know that if and when I get sued, I will be devastated by it. I don’t know that it will be something where I have to stop practicing, but I will cry about it. I will be very sad. (Portland obstetrician)

Instead of focusing on maternal-fetal conflicts, these data demonstrate that a more accurate model of risk management would examine potential maternal-fetal-provider conflicts. For example, a family physician described his use of continuous electronic fetal monitors:

I don’t think [electronic fetal monitoring is] appropriate for some patients, and I think more patients get C-sections or more interventions than they need because of it. But am I willing to be the one who sticks my neck out for a multimillion dollar lawsuit?…Where do I as a physician, where does that risk to me, versus the risk to the patient, come in? (Milwaukee family physician).

As other sociologists have demonstrated, providers in the United States are nearly always incentivized to intervene early and often in the childbirth process (Morris, 2013).

Because the dominant risk culture prioritizes early intervention, how do physicians preferring fewer interventions operate within this framework? Obstetricians and family physicians often described themselves as risk tolerant, but still suggested that intervention would be the less risky course of action:

I don’t have a problem with prolonged labors and letting people be in active labor for a long time, and as long as they are fine and the tracing is fine I don’t have a big issue with it. I’m told that I’m fairly risk tolerant, compared to my colleagues [….] I sit on [fetal heart rate] tracings maybe a little bit longer than some of my colleagues. I have some risk tolerance. (Boston obstetrician)

Despite having a higher preference for, and professional comfort with, non-intervention than her colleagues, this obstetrician still frames her discussion of risk tolerance and aversion in a manner suggesting that intervening is less risky. Most physicians I interviewed similarly suggested that they viewed waiting to intervene in labor as a risky maneuver requiring tolerance, even though they acknowledged that interventions also carry their own risks. Below, a young family physician who had recently moved across the country to Milwaukee describes why she liked her new practice better than her old practice:

I’ve had like 3 months of comparison time, but it seems like [in my old job] people went to a C-section more quickly. I feel like [in my new job] we sort of have a higher, a little higher risk tolerance of bad heart tones and weird things before calling for a C-section. (Milwaukee family physician)

Providers who want to reduce monitoring and intervention in childbirth are not trying to reframe what it means to be a “risk averse” obstetrician, but are instead advocating for more “risk tolerance.” No one suggested, for example, that they have such an aversion to the risk of maternal complications from a cesarean that they tolerate more signs of fetal distress. Discussions of risk are framed by the interests of the fetus, even by providers who are trying to prioritize maternal health.

HOW SHOULD RISKS BE MONITORED?

In addition to differences in how physicians viewed their levels of risk “aversion” or “tolerance,” they also differed in how they thought risks should be monitored during the labor process. One of the primary ways to monitor the health of the fetus during labor is an electronic fetal monitor (for a more in-depth discussion of fetal monitors, see Owens, 2017). These monitors are strapped onto women’s abdomens and provide a continuous tracing of the fetal heart rate and maternal contractions. With the exception of some wireless devices, the monitors almost always restrain women to their hospital beds. Nearly all obstetricians and family physicians I interviewed used these monitors, although many were skeptical about their usefulness. Some providers felt that monitoring the fetal heart rate continuously led physicians to intervene based on small changes in the heart rate that had no real bearing on the fetus’ health:

We thought by watching these babies we were going to intervene in the nick of time, have good brain function and good outcomes. Of course, what we found is that there is such a false positive rate with monitoring, that it is actually what caused the rise in the C-section rates and the resulting problems related to that. (Milwaukee family physician)

Another obstetrician echoed this sentiment:

There are a lot of false positives with fetal monitoring, and so you run the risk of unnecessary intervention. I can see why patients try to stay home or try to have midwifery care, because testing begets more testing, and the more you look, the more you have to look, right? Maybe it’s better not to know. (Portland obstetrician)

Providers who were trying to limit the use of the electronic fetal monitors, like this obstetrician, recognized that it was difficult to accept that managing risk could mean collecting less information:

I’ve been involved in a project on starting to be more reliant on listening to the baby’s heart-beat in labor without the monitor but with the Doppler, intermittent auscultation. And what we’ve found is that we have been so indoctrinated, for the last fifty plus years, that it’s hard for us to go back to trusting birth and accepting an inherent risk that I may not be able to control everything that happens in your baby’s birth. Even if I’m doing my very best job. (Milwaukee obstetrician)

Conversely, other providers thought that using electronic fetal monitors was the obviously correct way to monitor risks during labor. As an obstetrician argued:

With the monitors you just get more information. I was taught that the more information you have, the better position you are in for an optimal outcome. (New Orleans obstetrician)

Even providers who worried about the use of electronic fetal monitors often felt most comfortable using them with patients:

The truth is, I will confess, I have a sense of security with the monitor. I understand that increasing fetal monitoring rates increases interventions. I totally understand that and believe it. But I’m also terrified of not knowing what is going on. (New Jersey obstetrician)

While the majority of physicians expressed opinions like the one above, and continued to use continuous fetal heart rate monitors, a significant minority were trying to limit their use. Many physicians suggested that continuous monitoring of the fetal heart rate was not just ineffective, but that it was actively hurting mothers by unnecessarily increasing the cesarean rate. They perceived collecting less information as a moral imperative and a mechanism to protect their patients’ health. The electronic fetal monitor may provide an illusion of security and control, rather than an actual mastery of risks.

WHAT COUNTS AS RISKY?

Another common difference I observed between physicians across the United States was how they viewed the severity of a cesarean in comparison to a vaginal birth. Some see very little difference between the two options. As an obstetrician noted:

I think most people do not see a very big downside to a section, patients or providers. Which is actually not untrue. I know we all want to say that [cesareans] are bad, and I’m a big fan of vaginal deliveries. [Cesareans] are worse, but very incrementally so […] It’s actually not bad, usually. (Boston obstetrician)

For this obstetrician, cesarean births were not inherently risky. She also suggested that her patients did not find cesareans particularly risky [although research suggests that decisions regarding cesarean sections are usually made by providers, not patients, (Declercq, Sakala, Corry, Applebaum, & Herrlich, 2013).] Another Boston obstetrician mentioned that she knew reducing cesarean rates was a priority at her hospital, but that she still found vaginal deliveries to be more risky:

So, not that we should be doing [cesareans] more, but there is very little incentive to get the rate down. You have to just sort of believe that vaginal deliveries are better for women, mostly. Not a lot of people are willing to take much, any risk with the baby. (Boston obstetrician)

Some physicians argued that while they thought cesareans were more risky than vaginal births, women should still want to err on the side of protecting their fetus by having a cesarean. This Portland obstetrician argued that cesareans are only slightly riskier for women than vaginal deliveries, which makes them a risk worth taking:

[A cesarean] shifts the risk towards mom, slightly. But she is already doing the single highest-risk thing: having a baby. People don’t understand, or they say ‘Oh, the C-section is this huge risk.’ And they compare it to nothing. Mom is already taking a big risk. Now, most of the time the delivery goes fine, but there is no guarantee that mom is not going to also have problems with vaginal delivery. So, you are comparing two risks, shifting it towards the baby. And we can’t afford to be wrong, in terms of babies. (Portland obstetrician)

While a small number of physicians felt comfortable with current cesarean rates, most were concerned about unnecessary complications from cesarean births. In opposition to the above physician, they did not find it appropriate for women to undergo risky interventions without clear evidence that a cesarean would help their fetus. For example, this obstetrician from Portland placed great emphasis on the value of a vaginal birth, and tried to avoid a cesarean whenever possible:

Most women will do anything for their babies, you know? They will die for their babies – they do it regularly. But the fact is that cesarean delivery is riskier for women. It just is […] It’s pretty rare to die in the United States, but it does happen, and in fact I had a sister-in-law who died due to a cesarean, so I think it’s real. Somebody who doesn’t have to have a cesarean, probably it’s good to avoid major surgery when you can avoid major surgery. (Portland obstetrician)

Similarly, a Boston obstetrician said he preferred a vaginal birth and was particularly concerned about the cesarean epidemic, noting:

Whenever you do surgery you get surgical complications. So, there’s three-fold higher rates of major organ injury, bleeding, infections, just from C-sections. Five billion dollars spent annually in the U.S. that we don’t need to be spending, longer recovery times, more suffering, so, you know, clearly overtreatment is a problem. (Boston obstetrician)

For these providers, the risks of cesarean births for women are real, specific, and potentially severe. Many physicians discussed the visceral and tangible unease they often felt when operating on women. This Portland obstetrician told me that while some of her colleagues view a cesarean as a minor operation, she disagrees:

[…] It’s a major abdominal surgery. When we cut into the woman’s abdomen for a C-section, I see her bladder, and her intestines, both large and small. And they don’t want to be touched […] Head to head, the risks of cesareans are greater [than vaginal birth], and down the line, the more cesareans you have, the higher the risk of abnormal placentation. (Portland obstetrician)

The above examples demonstrate that providers can have very different understandings of what a cesarean is and means for women. For some, the potential risk of harm to the fetus trumps nearly all other concerns. But for others, the risks of abdominal surgery for women are real and to be avoided whenever possible. These differences in risk perception seem to play a role in practice variation among physicians. A Boston obstetrician I spoke with thought this was the biggest source of practice variation in American childbirth:

I think different risk tolerance is probably the biggest [source of variation] between providers, and I think different providers place a different worth on a vaginal delivery. So it depends on how hard they are going to try for it. Some see it as a lot more valuable than others. (Boston obstetrician)

Variation in provider understandings of the value of cesarean versus vaginal births could have a significant impact on intervention rates across the country. In my interviews, it was difficult to pinpoint why these understandings varied. When I asked providers, they suggested that differences in training and personality affect how one views the risks of cesareans. Many also pointed to different regional birth cultures, like the emphasis on “natural” birth in the Pacific Northwest. Some hospitals are also placing heavy incentives on providers to reduce their cesarean rates. All of these factors deserve further research to determine how they may impact provider understandings of risk.

CONCLUSION

In this chapter, I explore variation in how US obstetricians and family physicians perceive and manage risk during labor and delivery. While my research confirms prior sociological accounts of the dominant risk culture in American childbirth, I find that some providers are resisting this dominant narrative by ordering fewer interventions like cesarean sections and calling for less continuous monitoring of the fetal heart rate. While prior literature in the sociology of childbirth suggests that spatial, institutional, and professional contexts drive differences in birth cultures in the US, I find that there is significant variation in how risks are perceived and managed among physicians practicing in hospitals. Thus, in addition to documenting differences between midwives and physicians, and between home and hospital births, it is also critical to examine differences among physicians and risk cultures. I argue that this variation in physician attitudes and practices has significant impact on intervention rates, and should be considered when designing programs to reduce cesarean rates or maternal morbidity and mortality in the United States [some approaches already take risk culture seriously (e.g., Main, 2017; Smith, Peterson, Lagrew, & Main, 2017; Vadnais et al., 2017)].

For medical sociologists, it may seem counterintuitive that some providers are seeking to manage risk through ordering fewer interventions and less frequent monitoring (For more on the turn towards less monitoring in medicine, see Owens, 2017). Classic theories of risk society and biomedicalization posit that risks are consistently managed through increased surveillance and technological intervention in the hands of expert scientific and medical groups (Beck, 1992; Clarke et al., 2003; Giddens, 1999). While this is still generally true, it is worth exploring the growing risk culture I identify in this research. In essence, these physicians argue that there are some health risks that they are unable to control. While monitoring and intervention may create an illusion of mastery of these risks, it can actually lead to worse health outcomes. There is some evidence that the movement towards “doing less” and “knowing less” is growing in medicine broadly. In 2012, a new initiative from the American Board of Internal Medicine called “Choosing Wisely,” in collaboration with professional medical societies, started publishing lists of “tests, treatments, or services that are commonly used […] and should be reevaluated by patients and clinicians” (Cassel & Guest, 2012, p. 1801). The recommendations suggest that physicians should stop performing many tests on average-risk patients, like pap smears and pelvic ultrasounds to test for ovarian cancer. The number of medical societies participating in Choosing Wisely has grown from nine to seventy over the past five years, and the initiative has been cited in nearly 300 journal articles and 10,000 popular media articles (ABIM Foundation, 2019).

At the same time that some providers are questioning the utility of medical interventions and continuous monitoring, others embrace the rise of big data and precision medicine. Increased collection of health data may sometimes lead to more personalized and effective care. I do not argue that physicians are seeking to know less and do less in all situations, but rather that they are interrogating the boundaries of responsible knowledge and intervention. Importantly, the impetus to collect less information and order fewer interventions in childbirth was sparked, at least in part, by increasing information about cesarean rates and risks. Ordering fewer interventions and gathering less information about patients may often be contingent on knowing more about the nature of health risks. As the amount of available health information continues to grow, it will be increasingly important for medical sociologists to examine how medical institutions and providers decide which types of tests and information are beneficial to patients, and under which conditions. I expect that there will continue to be vigorous debate among risk cultures, both in obstetrics and in medicine generally, as institutions and providers assess the boundaries of responsible knowledge and intervention.