Racial Disparity in C-Section Rates

Emily Oster: 

Why are you interested in C-section rates as an important outcome to study?

Molly Schnell:

C-sections are a procedure that of course can be lifesaving for mother and baby, if it’s necessary. But we have very high rates of C-sections in the U.S., rates that are so high that many would say [they] are too high. And of course it’s a major abdominal surgery that comes with the potential for a lot of risks and complications, both for mom and for baby. And so when we see really pronounced differences in rates of C-sections between black and white moms, it leads to questions of whether or not that might be contributing to persistent disparities in health outcomes between those two groups.

Emily: 

So what does the C-section rate in the U.S. look like now, on average?

Molly:

It’s in the low 30s, so somewhere between 32% and 33%. For white moms it’s going to be just below 30%, and for black moms it’s closer to 35%. And actually a lot of the data that we’re going to be using for this study is going to come from New Jersey, and New Jersey has a really high C-section rate, so it’s going to be about 45% for black moms and about 40% for white moms.

Emily: 

Wow. What is the correct C-section rate, and why is that a hard question?

Molly:

It depends on who you ask, right? The WHO is going to say somewhere between 10% and 15%. That’s a very difficult question to answer because it’s going to depend on a mother’s appropriateness for the procedure.

Emily: 

So there’s a racial difference in C-section rates, and you’re focusing here on unscheduled. These are people who come in with an expectation of laboring and having a vaginal birth. With that sample, what’s the kind of headline racial difference in C-section risk?

Molly:

We’re going to see that among mothers with unscheduled deliveries in New Jersey, black mothers are going to be 25% more likely to have a C-section than white mothers.

Emily: 

So there’s a set of people you have, they came in, you thought maybe they were unlikely to have a C-section. And then for black moms, that group is just way elevated, and their risk relative to the group that came in. And it seemed likely they’ll have a C-section where there’s an elevation, but it’s smaller.

Molly:

I think this is one of the most striking findings that just comes out of the data. We talked about how we get that measure of risk for moms, and if we group moms into risk, we have very low-risk moms — so thinking the lowest quintile — and then moms all the way up to the highest-risk; there is a disparity, but it’s very small for mothers once you get to the highest-risk quintile. Why? Most moms then are getting C-sections right as they should, and so there’s not much of a disparity. What is driving the underlying racial disparity we see? It’s low-risk moms. So if you show up to the hospital, these low-risk moms essentially have nothing in their record that would suggest that they’re going to need a C-section. We see that black moms are going to be almost 150% more likely to have an unscheduled C-section, so 1.5 times more likely, and that’s driving that overall disparity.

Emily: 

We started by talking about understanding the problem being core to developing policy solutions. So what do you take from your results about what should be done to try to help fix this?

Molly:

There are a couple of potential policies here. One set of policies I would group in the information bin: giving moms more information about C-section rates across different hospitals, or even racial disparities in those C-section rates. We know that patients tend to take this information into account when choosing their provider, and that also hospitals and hospital systems tend to respond to this sort of information in terms of trying to provide the services that patients want. And so that could help mothers choose where they want to deliver, and it could also put some pressure on hospital systems to change these numbers.

Emily: 

Doulas. I was hoping you would get to doulas. Otherwise I was going to have to get to doulas.

Molly:

There’s great evidence showing that birth outcomes are better when people have doulas, and there’s many reasons for that. But part of it is probably having somebody in the room who knows the process and who can advocate for you. And there’s been changes in reimbursement policy to make sure that women from many different backgrounds can have doulas with them.

Emily Oster: 

Molly Schnell, thank you so much for joining me on ParentData.

Molly Schnell:

Yeah, thanks for having me.

Emily Oster: 

We’re going to spend today talking about your new paper on race and C-sections and inequality. But I would love if you would start by just introducing yourself and saying what you do, where you are. You could tell us about your kids if you want to.

Molly Schnell:

Yeah. So I’m an assistant professor of economics at Northwestern, and so most of my work focuses on the causes and consequences of provider behavior, so thinking about how individual clinicians make decisions in the presence of incentives and constraints. And C-sections is one of those decisions, so I’m interested in digging into that in the data.

Emily Oster: 

Awesome. All right. So we’re going to talk about this paper of yours, which is about racial differences in C-section rates. But I want to sort of start at the beginning and set the stage, both about what are those racial differences if you just sort of look out at the world and the question of why this matters. So let’s start actually with the second question. So why are you interested in C-section rates as an important outcome to study?

Molly Schnell:

Yeah. So C-sections are a procedure that of course can be lifesaving for mother and baby, if it’s necessary. But we have very high rates of C-sections in the US, rates that are so high that many would say are too high. And of course it’s a major abdominal surgery that comes with the potential for a lot of risks and complications, both for mom and for baby. And so when we see really pronounced differences in rates of C-sections between black and white moms, it leads to questions of whether or not that might be contributing to persistent disparities in health outcomes between those two groups.

Emily Oster: 

So what does the C-section rate in the US look like now, on average?

Molly Schnell:

So it’s in the low 30s, so somewhere between 32 and 33%. For white moms it’s going to be just below 30%, and for black moms it’s closer to 35%. And actually a lot of the data that we’re going to be using for this study is going to come from New Jersey, and New Jersey has a really high C-section rate, so it’s going to be about 45% for black moms and about 40% for white moms.

Emily Oster: 

Wow. What is the correct C-section rate, and why is that a hard question?

Molly Schnell:

So it depends on who you ask, right? The WHO is going to say somewhere between 10 and 15%. That’s a very difficult question to answer because it’s going to depend on a mother’s appropriateness for the procedure. So that’s going to differ across populations, that’s going to differ across countries, that could differ for a given mom throughout her life cycle. She might be a good candidate for a C-section for one of her births and not for another based on her underlying conditions. And so it’s ultimately closely tied to medical risk, and that’s something that changes. And so one of the goals that we’re going to really look at in this… One of the goals of this study is to use a lot of information on medical risk to see whether or not that’s one of the reasons why we see these big differences in C-section rates.

Emily Oster: 

Yeah. If we sort of establish that a C-section, while safe and can be lifesaving, that the rates may be too high and that it is, at least in the short term, a much more complicated recovery and may have contributions to morbidity. The way I’d often like to say this is like, “This can be a great option if you need it, but it’s not likely to be the first choice for delivery method”.

Molly Schnell:

Yes.

Emily Oster: 

So when we look out at the world and we see differences in C-section rates, either across space or across racial groups, what are the possible set of reasons we could see that? So I would say you said New Jersey is very high and that there are big differences across racial groups. So as a researcher, as a person who studies it, what are the set of things that might be going on?

Molly Schnell:

Yeah, that’s a great question. So the high C-section rates in general and the disparities by race are really well-documented and quite persistent, and so that’s led a lot of people to hypothesize about what could be driving that. It’s going to be important to know that to be able to point to different policy solutions. So what have people thought? Well, people have argued that it could be differences in what we would call maternal preferences. So it could be, for example, if we see that black moms are much more likely to have a C-section than white mothers, maybe it’s that black mothers prefer to have C-sections. So maybe they’re requesting them from their providers and that’s why they’re higher up rates. That of course would go against a lot of other things we know about how black Americans tend to interact with the healthcare system given a history of racism within the healthcare system. In the US, we tend to see that black Americans want less intensive treatment as opposed to more intensive treatment, and so it would be surprising if C-sections sort of went in the opposite direction. But that’s one of the hypotheses that you’ll hear of maybe its differences in maternal demand.

Emily Oster: 

Of course that would lead the question, why does that hospital have-

Molly Schnell:

Exactly.

Emily Oster: 

[inaudible 00:06:00] explanation.

Molly Schnell:

Yeah. This is just going to kick the can a little bit in that we’d still have to figure out why those rates are different within those hospitals. But maybe once you look in the same hospital, or within the same provider, you see no racial differences between C-section rates and it’s really just selection into different hospitals. Of course, it could also be something about provider discretion, either that providers are more likely to do C-sections for black moms because they’re more worried about something going wrong, or maybe there are just general biases in the healthcare system that we observe in a lot of other settings in the US.

Emily Oster: 

I think the reason that thinking about these explanations and trying to distinguish between them is so important is that we can’t really think about solutions without understanding the problem. So if you said, “I’d like to lower this disparity, I’d like to lower the rates, I’d like to lower it in general”, just observing this very robust correlation or differences across groups is actually going to tell you nothing about how to fix it. And those explanations you just gave me all suggest pretty different approaches to whether this is a fixable problem or something we could affect and how we would do it. I mean you’re going to approach this pretty differently if it’s something about differences across hospitals, versus something about differences across provider discretion, and certainly if it’s something about differences in risk factors.

Molly Schnell:

Yeah, absolutely. I think also, when you hear about this disparity, you sort of naturally think that that’s probably a bad thing. But if it’s differences in risk or if it’s differences in preferences, well then maybe we shouldn’t be so worried about it. We should want to address those differences in medical risk, but at the point that a mother shows up to deliver we don’t want clinicians to reduce C-sections among black moms if they really just are more appropriate for the procedure.

Emily Oster: 

So your goal in this research paper, which is entitled Drivers of Racial Differences in C-sections, is to try to unpack a bit what in fact is going on. So you’re going to start this paper by trying to understand whether the differences across racial groups are about differences in pre-existing risk. Whether there’s something that we could point to and say, “People are more likely to have this complication that makes the C-section more likely”. And so do you do this in a very specific way using very fancy methods. So could you talk me through a little bit what you do and why it was important to approach it in this quite comprehensive complicated way?

Molly Schnell:

Yeah, absolutely. So basically when people have looked at this in the past, where do we get these statistics on C-section rates by race and over time, those tend to come from birth certificate data. So you can get really good information on when you’re at the hospital and you fill out that birth certificate after your child’s born, those are digitized and researchers get access to those and you can track things like whether or not people had C-sections.

Emily Oster: 

Just to be clear, people, this doesn’t have your name on it.

Molly Schnell:

Yes.

Emily Oster: 

So people would not be able to find you with this, the information from the birth certificate. Sorry. Just don’t want [inaudible 00:08:58] panic and go remove themselves from-

Molly Schnell:

Yes. These are anonymized, but they’re very useful for researchers in order to track these trends over time. Now that’s useful in that it has a lot of information, it’s not very useful in that it doesn’t have a lot of information about whether or not we think… Or what the mother was presenting with that would cause her to either need a C-section or not. So what we’re going to do is we’re going to get really detailed administrative data from the state of New Jersey. So again, this isn’t going to have individual’s names, so this will be anonymized on the patient side, but we’re basically going to see everything that would be in the mother’s record. And so we’ll know all of our health conditions leading up to birth. We’ll actually also be able to follow mothers and babies following the delivery in order to look at how this might influence their health later on.

Emily Oster: 

This approach, it shares some features with a regression approach of sort of trying to think about basically holding constant characteristics across people.

Molly Schnell:

Mm-hmm.

Emily Oster: 

But doing it in a way where the thing you’re holding constant is in some ways way richer and much closer to holding constant the actual thing you care about, which is when this person showed up what was the likelihood, based on everything you would see about them when they arrive, what was the likelihood they could have the C-section? They would have a C-section and you sort of have that number produced by this. You guys use a random forest?

Molly Schnell:

We used a random forest. Exactly. And so it’s exactly as you describe it. We can think of it similar to our regression, it’s just going to have a lot of nice properties that we tend to like in research and that’s going to do a very good job of predicting whether or not you actually were going to have a C-section.

Emily Oster: 

Do you have a sense of how good these random forest models have metrics of how predictive they are? How good is your random forest?

Molly Schnell:

So what we can do is we can sort mothers in the testing sample into groups based on whether or not they need one, and we’re going to basically be able to predict it for you. So the women who are not in our sample that we’re using to train the model, we can kick out this number of whether or not we think you’re going to have a C-section. We can group you into different groups based on that average risk. And then we can compute the true chance that mothers in that group actually had a C-section, and they’re going to line up, they’re going to match almost perfectly. So I think this also tells you there’s a lot of discretion in the decision of whether or not to have a C-section, that I’m sure we’ll get into in just a bit, and we will see some patterns in the data that are going to really highlight the role of discretion. But at the same time, there are a lot of characteristics that are going to really predict whether or not you’re a good candidate for the procedure or not.

Emily Oster: 

Yeah. Okay. So let’s start. Let’s get into what you find. So I want to start with the baseline difference. So there’s a racial difference in C-section rates, and you’re focusing here, it’s probably useful to say, on unscheduled. So we’re already taking out the possibility, although I agree it’s remote and not true, but the possibility that there are differences in what people request or even what is sort of happening at the doctor beforehand in terms of how you negotiate for things. These are people who come in with an expectation of laboring and having a vaginal birth, and then there is a C-section or not afterwards. So with that sample, what’s the kind of headline racial difference in C-section risk?

Molly Schnell:

Yeah. So we’re going to see that among mothers with unscheduled deliveries in New Jersey, black mothers are going to be 25% more likely to have a C-section than white mothers. And I think that’s already quite striking. Why? I agree with you that I didn’t think this was going to be differences in maternal demand. I don’t often hear a lot of moms sort of begging to have a C-section. But if you really want to have a C-section, you can certainly find a provider who’s willing to schedule one for you. And so that’s the reason we’re focusing on unscheduled deliveries, and is also something that people haven’t been able to do in previous data. So again, going back to that birth certificate data, again is anonymized, it will just say whether you had a C-section or not. It’s not going to say whether that C-section happened after a trial of labor.

Emily Oster: 

So you see an increase in… We should say it’s about 25%. It’s not 25 percentage points, it’s 25%.

Molly Schnell:

No. So it’s 4.2 percentage points over the base for white moms is 25%.

Emily Oster: 

So I think a thing that’s really striking for me, even in the baseline statistics, is that this difference is in percentage terms much larger for people whose ex-ante coming in risk of a C-section is lower. So there’s a set of people you have, they came in, you thought maybe they were unlikely to have a C-section. And then for black moms that group is just way elevated, and their risk relative to the group that came in. And it seemed likely they’ll have a C-section where there’s an elevation, but it’s smaller.

Molly Schnell:

Yeah. I think this is one of the most striking findings that just comes out of the data. We talked about how we get that measure of risk for moms, and if we group moms into risk… So we have very low-risk moms, so thinking the lowest quintile, and then moms all the way up to the highest risk, there is a disparity but it’s very small for mothers once you get to the highest risk quintile. Why? Most moms then are getting C-sections right as they should, and so there’s not much of a disparity. What is driving the underlying racial disparity we see? It’s low-risk moms. So if you show up to the hospital, these low-risk moms essentially have nothing in their record that would suggest that they’re going to need a C-section. We see that black moms are going to be almost 150% more likely to have an unscheduled C-section, so 1.5 times more likely, and that’s driving that overall disparity.

Emily Oster: 

I mean this paper’s really, really good, but it’s also very upsetting and we haven’t actually gotten to the most upsetting part leading into it. The next thing you do in the paper is you talk about these controls. You talk about basically trying to rule out some of the differences that might be driving this, and what happens when you use your fancy system and do that?

Molly Schnell:

Yeah. So we’re going to start with that baseline disparity that we talked about. That if we just look at women showing up with an unscheduled delivery, that we see that black moms are 25% more likely to deliver by C-section. We’re then going to basically control for things that are closely related to those hypotheses that people have thought might be driving this disparity to see how it changes that disparity. So we can control for medical risk. So say think of two moms with the same medical risk, one who’s black and one who’s white. We’re actually going to see that, if anything, the disparity is then larger. Now that seems a little surprising because you might, as I said, that black moms are going to have on average some health conditions that would make them better candidates for C-sections. However, in New Jersey we’re going to see that black moms are significantly younger on average, and that’s going to be a very important predictor of whether or not you have a C-section. And so the average black mom is actually going to be a less appropriate candidate for the procedure than the average white mom.

Emily Oster: 

So in this, you’ve controlled for quite a lot of stuff, and I think it still shows up in this very striking way that there’s this difference. But a skeptic coming to this is going to say, “Well, it still could be unobserved differences. Yes, you see everything about their risk status, but you weren’t there. You didn’t hear them talk about what they wanted. There’s something that you’re missing”. And this is in some ways for me why this sort of last part of the paper is so crucial where you show us one more piece of evidence that I think really suggests that what is going on is not differences in unmeasured risk. And so can you just talk us through this part of the paper about surgical resources?

Molly Schnell:

Yeah, absolutely. So as you said, we’re not there and there are certainly things that clinicians on the ground observe that we as researchers, even if we have the best data possible, not everything’s written down in the medical record. And so what we were thinking when this disparity still persisted conditional on everything that we saw is that there could be differences in what we would call unobserved health risks. So maybe black mothers are unobservably to us, to the researchers, but not unobservably to the clinician, just more appropriate candidates for the procedure and so that’s why they’re getting additional. And so what we do in order to look at that is we leverage variation throughout the day in terms of whether or not there’s a scheduled C-section at the same time. So why is that useful for us? Well, one thing that you see in the data is that when there’s a scheduled C-section at the same hour that a mother with an unscheduled delivery is delivering, she’s going to be significantly less likely to have an unscheduled C-section.

Emily Oster: 

And that’s true across races. That’s just a basic fact.

Molly Schnell:

That’s true across races. You just see that when the OR… These hospitals in New Jersey typically only have one or two operating suites sort of set up to do C-sections. And you see that when one of them is busy, mothers are significantly less likely to have an unscheduled C-section, which already I think highlights the discretionary nature of a lot of these unscheduled C-sections that it doesn’t drop to zero. It’s not that when there’s a scheduled C-section there’s no unscheduled C-sections. If mom and baby are going to die, they’re absolutely going to do that unscheduled C-section, and we still see some unscheduled C-sections happening. But you see that throughout the day they just sort of trend inversely with what’s happening with scheduled C-sections. So they’re significantly less likely when that takes place.

Emily Oster: 

I think the uncharitable view of this discussion is we have a free OR, C-sections pay a lot, let’s use it. Who should we use it on? Maybe we’ll use it on black moms. I mean that is one thing that would be consistent with your explanation, is it not?

Molly Schnell:

Yes. So that certainly would be consistent with what we’re seeing. And I’ve spoken to some clinicians since the study came out who actually sort of thought that that was the most likely explanation. But I think there are also more charitable explanations that could be happening. It could be the case that clinicians are certainly well aware of disparities and outcomes between black and white moms, and maybe when there is additional capacity to put additional resources towards the care of black mothers, maybe they’re more likely to want to just provide those extra resources to black moms to try and narrow those disparities. Now what we’ll see in the last part of the paper, and we can look at the health implications of that, is that C-sections aren’t costless procedures for mom or baby. And so if you’re very low risk and you get a C-section, that can lead to additional complications that you wouldn’t have had in the absence of that surgery. And so that’s not necessarily a good thing for clinicians to be doing, but you could envision that that would be part of the thought process. Although it could also be that they want to fill that room with a mom and they’re going to choose which moms to do that on.

Emily Oster: 

Before we get to the last part about health impacts, I’m curious if you have a sense from talking to clinicians about, in some ways even putting aside the racial differences, this suggests a fair amount of discretion in whether I go to a C-section or not. A reasonable amount of the time, I’m kind of like, “We could do it, we could not do it. If there’s a space, we do it”. Do you have a sense from clinicians of what are you doing instead? Is it just trying harder? What is happening that is allowing vaginal births in these cases in which the OR is busy and somehow we’ve given up when the OR is not busy?

Molly Schnell:

Yeah. I guess what we’ve heard from clinicians is that there’s often a lot of pressure, both internally, externally, from peers, from the system, also from themselves of just wanting to keep your schedule in order. And so if you know have a lot of births ahead of you for the day, you might want to move some women through and so that you can sort of clear up time for other mothers. There’s also evidence showing that towards the end of a shift, people are more likely to do C-sections. And so in addition to the financial considerations that are always going to be present with these different types of delivery methods, there is also just time constraints and just wanting to keep things moving.

Emily Oster: 

We did an interview a couple of episodes ago with a midwife, at which this sort of general discussion came up. And I think one of the things that she would articulate as being somewhat different about the midwifery model of care is you expect to be there for a while, and you’re just like you’re hanging out. There’s more waiting. There’s more waiting.

Molly Schnell:

Yes. No, absolutely. I think that it’s just a different approach to it.

Emily Oster: 

Yeah.

Molly Schnell:

With my first child, I had an obstetrician tell me that if I didn’t want more Pitocin, then I should go home because the beds are reserved for women in labor. And so I think he was really like, “We got to move this along or let’s just get you out of here”.

Emily Oster: 

Please contract faster, Molly. Thank you.

Molly Schnell:

Exactly.

Emily Oster: 

Your pelvis is too small and it’s moving slow.

Molly Schnell:

Yeah. He wanted things to just go faster. He seemed really focused on how long it had been and that he wanted to get the show moving and he wasn’t willing to wait.

Emily Oster: 

With my second kid I almost waited too long to go to the hospital, having been given a similar speech with the first one, and we arrived at the hospital like 15 minutes before the baby came out.

Molly Schnell:

There you go.

Emily Oster: 

That was too fast. I also got [inaudible 00:28:47]

Molly Schnell:

Too fast. Exactly.

Emily Oster: 

Too fast. So let’s talk about the last piece of this, which is what are the implications? So what are you finding when you’re able to look in the data on health impacts? What is the result of this?

Molly Schnell:

Yeah. So we can use that same variation in terms of… Or the variation in whether or not there’s space in the OR to basically try and isolate these what we would call marginal or discretionary C-sections. And so you should think of this as a mom who has a C-section because the OR is empty, but she wouldn’t have had a C-section if the OR was full. And so these are those cases that are on the margin. And what we’re going to see is that when those C-sections take place, particularly for low risk mothers, we’re going to see additional complications, both for mom and for baby. So for mom, what we’re seeing, we’re going to see additional complications basically of the surgical wound. Which makes sense, right? You’re not going to have complications of a surgical wound if you didn’t have surgery. And so I am not surprised that C-sections are leading to increases in complications involving those wounds. But again, this really just emphasizes the fact that these procedures do come with risks, and so that you really only want to be having one if it’s really necessary.

Emily Oster: 

So what do you take away in terms of policies? So we started by talking about understanding the problem being core to developing policy solutions. So what do you take from your results about what should be done to try to help fix this?

Molly Schnell:

Yeah. So I think there’s a couple of potential policies here. One set of policies I would group in the information bin. And so you could envision giving moms more information about C-section rates across different hospitals, or even racial disparities in those C-section rates. We know that patients tend to take this information in account when choosing their provider, and that also hospitals and hospital systems tend to respond to this sort of information in terms of trying to provide the services that patients want. And so that could help mothers choose where they want to deliver, and it could also put some pressure on hospital systems to change these numbers.

Emily Oster: 

Doulas.

Molly Schnell:

Exactly.

Emily Oster: 

Doulas. I was hoping you would get to doulas, otherwise I was going to have to get to doulas.

Molly Schnell:

Yeah. And I think to the extent that there is a discretionary part of this, making sure that you have somebody in the room who can help advocate for you. There’s great evidence showing that birth outcomes are better when people have doulas, and there’s many reasons for that. But part of it is probably having somebody in the room who knows the process and who can advocate for you. And there’s been changes in reimbursement policy to make sure that women from many different backgrounds can have doulas with them. In New Jersey, it was added to Medicaid in 2021. And so our study ends in 2018. I would love to go back into the data in more recent years and see whether or not that’s affected these racial disparities, and whether it’s changed whether or not people have somebody with them and whether it changes what we’re finding here. So advocates I think is important.

Emily Oster: 

Thanks for doing this work, Molly.

Molly Schnell:

Yeah, thank you for covering it.

Emily Oster: 

Thanks for coming and talking about it. I think it’s incredibly important, and also really high quality, which doesn’t always overlap. So I’m glad you’re doing it.

Molly Schnell:

I’m glad you think so.

Emily Oster: 

Yeah.

Molly Schnell:

Awesome. Thank you.

Emily Oster: 

ParentData is produced by Tamar Avishai with support from the ParentData team and PRX. If you have thoughts on this episode, please join the conversation on my Instagram, @ProfEmilyOster. If you want to support the show, become a subscriber to the ParentData Newsletter at ParentData.org, where I write weekly posts on everything to do with parents and data to help you make better, more informed parenting decisions.

Penelope:

Right, Mom.

Emily Oster:

We’ll see you next time.