Learn more about Ariadne Labs' response to COVID-19

Atul Gawande Asks Katherine Semrau: How We Can Improve Maternal and Newborn Health

Ariadne Labs teamed up with Project Syndicate to produce a podcast about our BetterBirth study in India. Our executive director Atul Gawande took over the reins and interviewed BetterBirth Director Katherine Semrau about what we learned after one of the world’s largest maternal newborn health studies.

Ariadne Labs Executive Director Atul Gawande, MD, MPH

Gawande: Today I’m taking over the reins from Greg as host for a special edition of this podcast. I going to get to talk to Dr. Katherine Semrau, who leads the BetterBirth program at Ariadne Labs. I want to talk to her about a startling fact: 99 percent of deaths in childbirth—deaths of moms and babies—are known to be avoidable. We have the knowledge. We even have the will. But although we’ve made progress in reducing those deaths, we still don’t know how to make sure the right thing happens.

Katherine is a maternal and newborn health epidemiologist who has spent more than 15 years working on ways to improve care and health outcomes for women and babies in Africa and Asia. For the last three years, she has lead the BetterBirth study in Uttar Pradesh, India, one of the largest studies ever conducted in maternal and newborn health, which I got to be a lucky part of the study as well.

It is the first study to demonstrate large-scale improvement in care during the 48 hours of child birth when women face the greatest risk of death and complications. But those improvements were not enough to reduce the maternal and perinatal mortality rate in Uttar Pradesh. It made the care better, a lot better, and yet, deaths did not fall. Our puzzle is why. So let’s dive in.

So Katherine, women die in childbirth all over the world—and in the U.S. We believe that 99 percent are avoidable. But why do moms and babies die from childbirth? What is the cause?

Katherine Semrau, PhD, MPH, BetterBirth Program Director, Ariadne Labs

Semrau: Thanks so much for the opportunity to talk a little bit about this work. There are seven big killers of moms and babies and these causes of death are the same around the world, whether they’re in the United States or in India. For women we see that the leading cause of death around the time of childbirth are: hemorrhage or too much bleeding; hypertension, also known as eclampsia; and sepsis, or infection for women.

For newborns we see the real causes of death are asphyxia or difficulty breathing right in the first few minutes after delivery. Prematurity. And also, children die of sepsis.

The thing is we know how to address these causes of death through proper clinical care when a woman is pregnant and all the way through the post-partum period. And especially during labor and delivery, we know that monitoring blood pressure and temperature can address those concerns. We know that handwashing to prevent infection and making sure that the proper equipment and supplies are available at the ready to actually deal with these causes of death.

Gawande: So if I were to take an example here, it’s astonishing to me how childbirth is a kind of brutal thing. 10 percent of babies are born with difficulty breathing. So let’s just take that as an example. What makes death from that difficulty breathing avoidable? What has to happen from that side?

Semrau: At the bedside, the birth attendant has to, first of all, recognize the child is having difficulty breathing. The interaction or the intervention is really about, first of all, ensuring that the airway is clear. If it’s not, using the neonatal bag and mask to help that baby start breathing on their own. So in order for all that to happen, there has to be a chain of events: recognition by the staff that there’s a difficulty breathing; second, having the supplies and equipment available at the bedside. Not somewhere else in the clinic or the facility, but making sure the bedside tray is ready with the supplies and equipment needed. And then the staff have to have the right skills and capabilities to be able to actually conduct appropriate newborn resuscitation.

Gawande: It’s fascinating to me how prosaic this is, right? So the 10 percent of babies born with difficulty breathing; it seems like it should be simple to recognize that they are struggling, but it’s not always totally obvious. And then the very first thing you do is you dry the baby with a clean cloth and really stimulate, vigorously. That gets 90 percent of them going, but that sterile towel may not even be there. And that–that’s a problem that you had to figure out how to solve in this study.

Then you had to go farther and have this more complicated thing; you know, a bag and mask and use it properly. So the approach that we took here, how did we achieve major changes, where those things started to happen more commonly?

Semrau: So we really focused on a systems-wide approach that used the WHO Safe Childbirth Checklist, which is an itemized list of 28 actions or interventions that are evidence-based practices that should be done for every mom, every baby, around the world, during labor and delivery. We paired that checklist with a peer coaching model that focused on nurses coaching nurses, physicians coaching physicians about how to improve the care that’s being conducted at the bedside. It was about strengthening the communication in the clinic; ensuring that supplies and equipment were available; and making sure that the skills of the staff were appropriate.

Gawande: What’s radical about that is normally we’d say, “Hey, go get training, go take a course. You will figure out how to do this.” And then you come in. What makes it a systems approach by putting someone at the bedside?

Semrau: First of all, it’s taking the training out of being an external place where people often get trained and doing it in the room where nurses, birth attendants are actually conducting their labor and delivery work. It is focused on the reality on the ground and what these birth attendants are having to deal with.

Second of all, it’s focused not just on one intervention alone or one birth attendant alone, but rather focusing on communication with leaders–focusing on the district-level staff, the state-level staff. So it’s not just about one individual birth attendant, but the whole health system as an approach to tackle this problem.

Gawande: The nurse can’t solve the ‘there’s-no-sterile-towel-available’ or ‘the baby mask isn’t working’. So that sounds daunting. How big is this trial?

Semrau: We worked in 120 facilities. 60 of the facilities got the intervention, 60 were just control sites that had standard of care. But we followed 160,000 mother-infant pairs over a two-year period to find out whether we had made a difference. First of all the care that was being provided to these women and children but also their mortality rates.

Gawande: So how much did the care change?

Semrau: The care changed dramatically. In the standard sites, or the control sites that did not receive the checklist coaching, their adherence to practices was about 40 percent overall, but less than one percent of the staff were washing hands before delivery, only 25 percent of the women were receiving the right medication to prevent hemorrhage. And that was in the control sites.

Where we see a big change is with the intervention coaching plus the checklist. We saw that the intervention sites has 70 percent adherence to these practices. So 80 percent of the moms were receiving the right drugs to prevent hemorrhage after labor and delivery.

So we know that there can be a shift in these behaviors. We did see dramatic improvements in the quality of care that’s being provided at these front-line facilities.

Gawande: And then?

Semrau: And then we looked at mortality and we did not see an actual impact on perinatal mortalities, so those were stillbirths and early neonatal deaths. We didn’t see a difference in maternal mortality either. I think what we are coming to understand out of the trial is that it really has to do with, yes, we achieved behavior change, but we didn’t solve all the underlying difficulties or problems. We see that the gaps in supplies and equipment weren’t fully addressed. We see that skills and capabilities of the staff or the leadership wasn’t sufficient to overcome challenges. We also see that there’s a lack of connection between this frontline facility and higher level facilities that can manage complications appropriately.

Gawande: So that means that context matters a lot. Give us a picture here. What does childbirth look like in typical primary health care facility where we were measuring five percent death rate for the babies. What do these places look like? Where are the deliveries occurring if there are other problems that have to be fixed?

Semrau: Think about a concrete block building. It typically has three to five rooms. One room is dedicated to labor and delivery and in that room is one to two beds. The beds are typically metal beds that are raised up off the ground to help the birth attendant to be able to catch the baby or deliver the baby well. But we often see that these facilities don’t have curtains or privacy between the delivery beds. You may have several women delivering at once in these facilities. There may not be electricity, or constant heat, or running water in these facilities.

Women often come to the facilities quite advanced in labor, so that the time that you have to intervene and catch a complication early during the laboring period is short. We also see that women stay for less than 24 hours after labor and delivery, often because it may be a crowded room or a crowded space; there may not be a recovery room available to those women.

So that’s another time period that’s very important in the life cycle for a woman and her child.  We see that within the first 24 hours is the period when you can catch most of the complications. It’s the highest risk period for the baby. So if a woman is not actually staying at a facility after labor and delivery for that 24-hour period, we’ve missed the opportunity to catch those complications.

I think the other thing to point out is that we’re seeing through some new research that respectful maternity care is lacking in many parts of the world. Up to a third or a quarter of women are disrespected or abused. They may be hit or slapped at the facility or have an episiotomy without consent.

Many women don’t know necessarily what to expect when they go to these facilities. But we see that women vote with their feet. They come to the facility late maybe because they don’t trust the care that they are going to receive and they are leave early because it doesn’t have the accommodation that they need.

I would like to point out though that from my experience that birth attendants around the world really do want good outcomes for the moms and babies. Just like a family wants to go home with a bundle of joy. And so, they really are working, these birth attendants, with what they have–at their hands, at the bedside. And I think making sure that the facilities can actually accommodate and provide high quality care is where we need to go globally.

Gawande: Yes the interesting thing to me — having gotten to visit these places with you and the rest of the team to see how birth really works, what that system looks like — that picture is stark. These places are really primary health care clinics. They are not hospitals. Right? The outpatient department, the OPD, as everybody calls it, that is where the doctors–THE doctor, usually–is most busy. They come in and they have some people they are seeing around the clock for all primary health care issues. And then they’ll see a woman who might be pregnant and do some prenatal care then they will come back to deliver at that clinic.

And so the clinic will have that side room where there are deliveries and then a recovery room. But it will typically be the nurse who will be the one who will do the delivery.They’ll call the doctor if they are having any problem, but it’s really expected that the nurses are doing the delivery. Can you give us a picture of the nurse: how experience are they? What kinds of capacity and capabilities do they have if the woman needs a C-section? Does the doctor have those capabilities? Is there capacity for those things here?

Semrau: It’s a great question. In this setting, it’s important to differentiate the nurses that are providing labor and delivery care are not midwives. Those are different cadres of health facility staff. In the BetterBirth trial, we see that the birth attendants that were providing the majority of care are nurses, staff nurses, who typically have two to three years of theoretical classroom training. And then they come into the facilities where they have been trained to be a general nurse, not necessarily a labor and delivery nurse. There’s a additional training that about half of the birth attendants had received called a Skilled Birth Attendant training, where they are taught about the most appropriate and better ways to provide labor and delivery care. But you see that these staff nurses work in labor and delivery rooms, most of the nurses in our study had about nine years of experience, on average, but hadn’t received additional training in about four years.

When it comes to the physician interaction at these frontline facilities, as you mentioned, there’s one physician. They may or may not have C-section skills. And often these facilities actually cannot conduct C-sections. There’s not an operating theatre. So women are referred onto a higher-level facility.

Gawande: You and I were part of a global study, where we recognized that about 19 percent of deliveries need a C-section, or at least, when C-sections are up to that level you get significant improvements in mortality. Because it addresses babies who are stuck and therefore having difficulties with asphyxiation. Or major maternal hemorrhage; obstruction that can lead to sepsis for the mom as well. What was the rate of C-section given that they couldn’t deliver here and then had to move onward to the next facility?

Semrau: Our C-section rate in the trial was two percent. Which is much lower than what we had expected. As you just pointed out, WHO recommends anywhere between fifteen to twenty percent to actually see an impact on maternal mortality, is where we see that break point.

This speaks to us in a couple of ways. One, women are coming to these facilities but aren’t necessarily being referred up to the higher level facilities for that C-section or may have been laboring for quite some time.

Gawande: The study came out of the New England Journal of Medicine just this past December. An editorial was written alongside it. It’s always disappointing to have to publish a trial that says we made a big change in delivering the thing that we thought would make the big difference, but we didn’t get the death reduction.

And the editorial that was written that said, look, it could not have been more rigorously, more carefully done, and it seems evident that these kinds of primary health facilities where the majority of care is being delivered, simply can’t be improved enough to get to high quality care.

Women need to be delivering in hospitals where they can get a higher capacity for C-section. Severe hemorrhage is about three percent and you could have access to blood transfusion.

We recognize another three percent had eclampsia, high blood pressure, and in most cases, still not being adequately treated however much we’re putting the checklist into place and following through on it. So what do you think, is that right?

Semrau: It’s not that rural primary facilities can never provide labor and delivery care. I think that kind of blanket statement could be challenging. A rural facility in India may deliver more than 2,000 babies a year. A rural facility in Namibia may deliver 500 or 600 babies a year.

It’s really about the capability of the health facility and the health system functionality. We know from other global work that health facilities that deliver less than 500 babies per year probably should not be providing labor and delivery services because the staff can’t keep up their skills to be able to manage complications or manage difficult situations appropriately. And we know that about 15 percent of women have some kind of complication around the time of childbirth. But it’s really been about making that frontline facility capable to handle the complications appropriately and have connections to a higher level facility that can provide C-section in a very short time.

I don’t think we’re ever going to get to the place where we build hospitals everywhere that can provide C-sections capability for all women everywhere. But it really is about strengthening the health system and connecting these lower-level facilities to a hospital, but it has to be fast, it has to have the transport system, it has to have the referral systems that can function.

Gawande: All right, I want to dig into this. This is a really interesting fundamental question.

My parents are from India. My father is from a rural village that is the kind of village we were trying to help in this program. They’re both doctors–they settled in rural Ohio in Athens County, the poorest county in Ohio in the Appalachian foothills. And where I grew up there, in the 70s, our first obstetrician came in 1968. We had no anesthesiologist, no formally trained anesthesiologist, until 1983. So, this was a rural community. It did not have 500 deliveries a year. When the obstetrician went out of town, it was just a nurse available to do deliveries. And we had to rely on transport–the next place was 45 miles away, that would have medical capability.

And achieving that death reduction required getting to the place where you had at least midwife-level capability, obstetrician or surgeon immediately available to do C-section, have the blood transfusion capacity, have the anesthesia capacity, that kind of thing, for a town of about 20,000 people in a county of about 120,000 people, which is not that different from the communities that we’re talking about in India.

The striking thing to me is around the 1920s in the United States, we actually said, we have to build that hospital. The Hill-Burton Act was all about building hospitals all across the country and then figuring out and staffing that capability and lo and behold, part of the reason why my parents came to that kind of place, was because we as a government started paying for and saying we have to drive that capability.

If the reality is that these are all the pieces you got to put together to save moms and babies, isn’t this then the direction you have to create?

Semrau: I think the argument of having women access tertiary care facilities or facilities that can conduct C-sections is right in the long-term. But we have to have a process to make labor and delivery safer now, and not in 10, 15, 20 years.

Gawande: It took about 50 years for our county.

Semrau: Right. Yes, the goal is to get everybody to well-equipped, well-staffed, capable facilities that can provide the right levels of complication management, C-section capability. Yes, that’s where we need to be aiming. But we have to have a strategy that can work in many different settings that may not have the economic resources that are going to be able to build hospitals rapidly. And I think that has to do with the cadre of staff. So making sure that nurses have the right level of training but also making sure that midwifery is an important component in health system for labor and delivery.

Gawande: So we can train more midwives?

Semrau: Yes, but again, that still takes time, right? All of these interventions we need to think about what can we do for the long-term–which is bigger infrastructure, bigger hospitals, capability, midwifery training–but also what can we do immediately about ensuring that supplies and equipment are available, ensuring that the staff who are conducting labor and delivery actually have the skills to be able to deal with a complication when it comes up. Because they come up fast. It’s not like they are always readily predictable.

Gawande: One thing that really struck me was that the mortality rate for newborns, we could see it by facility. Maternal death rates are lower and you couldn’t quite make out the variation. By facility, there’s really wide variation. How high did the mortality rate get and how low did the mortality rate get, because maybe we can drive everybody towards at least being down to that level?

Semrau: Exactly. I think that’s a great point, Atul. The variation in perinatal mortality was as high as 104 deaths per 1,000. So 10 percent of babies at one of these facilities was dying before seven days post-partum.

The lowest rate of mortality we saw was 14 deaths per 1,000. Which is just 1.4 percent, which is much closer to a middle-income country’s rate of perinatal mortality. I think what we have the opportunity now is to really learn what were these low-mortality settings doing that was so different than what the high mortality sites were doing.

Gawande: I want to talk about what we saw in Namibia where we had the partner there, a district hospital there, that adopted the checklist and the coaching model. They started at 70 percent adherence to the checklist, and then what happened?

Semrau: They started at 70 percent adherence to these basic practices and over a two-year period, achieved 90 percent adherence to the practices. But this is in a hospital setting, a district hospital setting. They had midwives that were providing labor and delivery care and they had a leader who was extremely passionate about quality improvement and ensuring that his staff had a skills lab, they did maternal death audits or maternal death reviews. Their approach was much more about adding the training, the supplies and equipment, and the review process with feedback at all levels of the system as well.

Gawande: So it seems like there are building blocks here, right. So they were able to get to 90 percent because they had midwife-level trained people, so they had the skills. If the baby’s not breathing, you could make sure resuscitation is there. They had enough organization that they could get the supplies and systems to be consistently there and that could pull them from 70 percent to 90 percent.

I was really struck by the fact that they reported not just a 50 percent reduction in stillbirths, but they went from having maternal deaths–usually three to four a year–to having none in two years. Now it’s been more than two years. It’s astonishing.

Semrau: Exactly. They really have been quite the success story. I think they have a good model for us to think about the package of intervention that came together to make that successful.

Gawande: Last couple of questions. One is, why after the trial, are we still nonetheless seeing 30 countries where people are now wanting to roll out the approach and test it themselves?

Semrau: I think that globally the maternal and newborn health community is really looking for interventions that can be successful. Not just using one-off interventions, but a systems approach. And I think the WHO Safe Childbirth Checklist provides guidance around what are the key practices and the guidelines that need to be adhered to.

It acts as a tool of where we need to get. The WHO Safe Childbirth Checklist in these 30 countries is being used in a myriad of ways. It’s being used as a training tool in pre-service training. It’s being used as a checklist at the bedside with labor and delivery. It’s being used to help identify where the gaps are in the health system, with respect to supplies and equipment or challenges that people are facing. So we’re seeing–and I’m excited to see–how are these different countries are implementing the checklist with different implementation strategies, all focused on achieving lower mortality rates for women and newborns.

Gawande: Awesome. Well, I was really struck by the fact we’re seeing it roll out in Indonesia and Chiapas, Mexico, in other parts of India, other parts of Africa–but then Italy. And the U.S. has got the highest mortality rate of major developed countries in the OECD data. Two of the biggest problems are not getting the blood pressure management right, not getting the hypertension right, and we’re even seeing that systems here adopt the checklist approach and so we’ll see whether, in the right setting, we have mortality reduction or not.

I’m going to throw you a wish list question. If you had unlimited funds right now, what would be the most important thing you would say can be done and should be done?

Semrau: With the long-term goal of getting women into well-equipped, well-staffed facilities, as was outlined in the commentary.  I think the pathway and my wish list would start off with ensuring that provider skills and the number of staff that are available at these facilities actually matches the need. So that the labor and delivery staff can actually manage the complications, that they can refer women appropriately when C-sections are required. And it’s more than just sending women or birth attendants to training. It’s actually an ongoing process of continuous medical education. And I think that needs to be greatly improved globally.

Second is this connection of the health system. We’re not overnight, as I said earlier, going to get women into hospitals, so we have to make the health facilities that are now currently doing labor and delivery be connected to the higher-level facility appropriately with the right tools of referral and that means dealing with transportation systems, but it also means dealing with communications systems between this lower- and higher-level facility. Finally, we really need to focus on community relationships. Globally we have pushed women to deliver in facilities. Those facilities are not always delivering on the quality of care that we have promised these women. And I am concerned that women vote with their feet.

They know where the high quality is, they know whether they have trusted relationships with the providers and I really think we need to understand more about what women expect, ensure that they fully know their rights and the care that they should receive in these facilities. So it’s really about building relationships of trust and respect with the community, as well.

Gawande: Well it’s such daunting work, but I think what you’re helping show is that it is tractable. There are ways you make progress. We can drive quality improvements and just have to figure out the leveraging to put it together.  Thank you for taking the time to join me.