Subscribe
Share
Share
Embed
Each week, Health Affairs Editor-in-Chief Alan Weil brings you in-depth conversations with leading researchers and influencers shaping the big ideas in health policy and the health care industry.
A Health Podyssey goes beyond the pages of the health policy journal Health Affairs to tell stories behind the research and share policy implications. Learn how academics and economists frame their research questions and journey to the intersection of health, health care, and policy. Health policy nerds rejoice! This podcast is for you.
00;00;00;00 - 00;00;35;00
Alan Weil
Hello and welcome to “A Health Podyssey”. I'm your host, Alan Weil. There have been longstanding efforts to reduce the incidence of preterm and low weight births, since they're associated with adverse outcomes for infants with effects that can last a lifetime. There have also been long documented disparities by race and ethnicity in these adverse infant outcomes in the United States. For example, rates of preterm birth for non-Hispanic black infants are more than 50% higher than rates for non-Hispanic white infants.
00;00;35;00 - 00;01;04;24
Alan Weil
But categories like black, Hispanic and Asian are very broad. They encompass large numbers of people from very different backgrounds. What do we gain and what do we lose when we report and use data on adverse infant outcomes that groups people into these large categories? That's the topic of today's episode of “A Health Podyssey”. I'm here with Clare Brown, assistant professor in the Fay W. Boozman College of Public Health at the University of Arkansas for Medical Sciences.
00;01;05;08 - 00;01;32;02
Alan Weil
Dr. Brown and coauthors published a paper in the February 2023 issue of Health Affairs, examining rates of adverse infant outcomes by racial and ethnic categories and subcategories. They find that disaggregating broader racial and ethnic categories can reveal important intra-group variation in rates of low birth weight and preterm birth. We'll discuss these findings on today's episode. Dr. Brown, welcome to the program.
00;01;32;19 - 00;01;34;17
Dr. Clare Brown
Thank you. I really appreciate the opportunity.
00;01;35;00 - 00;01;55;17
Alan Weil
I'm looking forward to talking with you today about these important findings. This is something we don't often do with the data, and you're going to tell us what you did. So let's start with the topic. You looked at adverse infant outcomes. Can you just give a little more of a sense than I was able to in the introduction of what those are and what do we know about why they happen?
00;01;55;29 - 00;02;15;05
Dr. Clare Brown
So in this particular study, we looked at low birth weight and preterm birth. Low birth weight is a birth that's less than about five and a half pounds. It's technically 2500 grams. In a preterm birth it's a birth that’s less than about 37 weeks gestation. There are a number of different factors that can increase risk for having a low birth weight or preterm birth.
00;02;15;21 - 00;02;37;07
Dr. Clare Brown
Some of these factors are demographic and aren't easily able to be changed. So like maternal age, someone who's very, very young or has increasing age, the risk of low birth weight and preterm birth is higher. There's also a number of clinical co-morbidities that may increase risk for low birth weight and preterm birth, the things like diabetes and hypertension.
00;02;37;19 - 00;03;00;17
Dr. Clare Brown
Having those conditions, particularly if they're not under control, can certainly increase your risk. We also have a growing understanding of how mental health conditions can impact low birth weight in preterm birth risk. Individuals who are exposed to chronic stress or may have experienced anxiety over the long term may have higher cortisol levels, which may increase risk for these outcomes.
00;03;01;07 - 00;03;24;22
Alan Weil
So there is a lot going into this, and it's really a priority at a population health level to try to bring these adverse event rates down. Now, anyone who's sort of worked in health care has heard about differences across race and ethnicity and things like infant mortality, life expectancy. What do we know about racial and ethnic disparities in adverse infant outcomes?
00;03;24;22 - 00;03;29;23
Alan Weil
What do we know about the disparities and what is, what do those disparities tell us?
00;03;30;03 - 00;03;51;07
Dr. Clare Brown
These are one of the largest health disparities infant and maternal outcomes, both. Black infants have around twice the rates of low birth weight and preterm birth. We have higher rates among Asian and Native Hawaiian and Pacific Islander groups. Hispanic populations have around 10% higher rates. But ultimately, what this indicates is lower population health among those groups, as you indicated.
00;03;51;17 - 00;04;16;06
Dr. Clare Brown
Infant outcomes are one of the primary metrics to measuring population health. It's something that's theoretically preventable. There are a lot of components that go into increased risk for adverse infant outcomes that we can mitigate and we can reduce those risks. So if a population has high rates of these outcomes, that's an indication that that population isn't as healthy as a population that has lower rates of those.
00;04;16;23 - 00;04;41;17
Alan Weil
So this really captures a lot about inequities in our society and now let's go into what you focused on because you did something I think really important and relatively unique in the literature. You instead of just looking at these broad categories, looked at subgroups. So instead of just Asian, for example, which is how we often categorize a population, you look separately at Filipinos, Vietnamese and Chinese and the like.
00;04;42;13 - 00;04;46;22
Alan Weil
Why did you do this? Why is it important to subdivide these large groups?
00;04;47;14 - 00;05;08;28
Dr. Clare Brown
So the reason that it's important is that within these broader categories, there are dozens of subpopulations. They’re the ones that I was able to look at, the ones that data are captured for. So I had seven different Asian groups, four different Native Hawaiian, other Pacific Islander groups, five Hispanic groups. But there are a number of other populations that are more granular that could fall into these categories.
00;05;09;07 - 00;05;31;24
Dr. Clare Brown
But what happens when we use these broader categories is we're masking potentially very large disparities and certainly masking cultural differences that may exist within these groups. While Hispanic populations may all largely speak Spanish, the Asian category and the Native Hawaiian or the Pacific Islander category literally have hundreds, if not thousands of languages and cultures that are within those categories.
00;05;32;09 - 00;05;55;21
Dr. Clare Brown
You think of the continent of Asia and it's very large. And I think that something that made me think about this was during the COVID 19 pandemic, there were some Asian subpopulations that maybe were more discriminated against than others. So those who were Chinese maybe face different levels of discrimination than those that identify as Asian Indian, although we would capture both of these under this broader Asian category.
00;05;56;02 - 00;06;16;01
Dr. Clare Brown
And so I think that that kind of opened my eyes to this and maybe recognize that we're masking a lot and we're not able to capture the underlying rates of these conditions. And outreach isn't going to be as good. If you provide educational materials to someone not in their native language, you're not going to be able to utilize that information as well as if it was in their native language.
00;06;16;12 - 00;06;41;00
Alan Weil
So you've really set this up nicely. Let's get to the findings in the paper. You showed levels of intra race and ethnicity variation and also across the broader racial and ethnic categories. Tell us a little bit about what you found and maybe in this instance, because there are a lot of charts and graphs in the paper, a couple of examples that might drive this home.
00;06;41;08 - 00;07;03;26
Dr. Clare Brown
Yes there are some relatively large charts, and that highlights exactly the point that there are lots of different subcategories within these broader categories. So the largest difference, let's take low birth weight just as an example. Again, I looked at both lower birth weight and preterm birth, but just for simplicity, for low birth weight, the largest disparity was in the broader categories was between white and black populations.
00;07;03;26 - 00;07;25;12
Dr. Clare Brown
So that ranged about 2.2 fold. So about twice the rates among black versus white individuals. I do want to point out that in this study I was looking at singleton birth, so this was looking at when a mother had only one child, so not twins and triplets and we do know that the rates of these adverse events are much higher when you have a multiple birth, but between white and black that the rate was about double for black women versus white.
00;07;26;00 - 00;07;57;03
Dr. Clare Brown
However, the disparities were the variation within categories was larger for the multiple race group. That multiple race category was about 2.3 fold, which is not incredibly larger, but it is 2.3 is bigger than 2.2. And there was 21 multiple race categories that we looked at. When you stop and think about it, it makes sense. We are combining into this multiple race group, someone who is, say, Asian and white, as well as someone who say black and native Hawaiian, other Pacific Islander.
00;07;57;04 - 00;08;20;13
Dr. Clare Brown
Those are people who literally live in different hemispheres. And so it's interesting that we, that is something we regularly do as researchers or as policy advocates, and we can talk more about potentially the implications in next steps for that. But we did see larger variation in terms of within the multiple race category versus what we saw between the lowest and highest rate when we were looking at just the category rates.
00;08;21;11 - 00;08;43;06
Alan Weil
So just to capture that, you can look at these broad categories and you see big differences across them, the ones that most of us have heard about historically. But when you look inside a category, in some instances in many instances, you actually find more differences between the subgroups within that category than you find between one category and another.
00;08;43;06 - 00;08;45;11
Alan Weil
Is that the right way to understand this?
00;08;45;11 - 00;08;56;29
Dr. Clare Brown
Yes, that's exactly correct, that there are these extreme variation within these broader categories. So within Filipino, Asian Indian, Chinese, etc., within that broader Asian category.
00;08;57;16 - 00;09;39;29
Alan Weil
Well, this is really interesting. I want to talk a little bit about the implications of these findings. We'll do that after we take a short break. And we're back. I'm speaking with Dr. Clare Brown about rates of preterm birth and low birth weight. Looking not just at racial and ethnic groups, but at subgroups. Before the break, we learned the overall findings, which is that the within group variation in many instances is larger than the across group variation, which suggests that we're hiding a lot of variation when we use these larger categories.
00;09;40;12 - 00;09;59;22
Alan Weil
Before we talk a little bit more about the implications and what to do about it, there is one other dimension that you looked at in the study, which has to do with U.S.-born and foreign-born parents or mothers. So can you just say a little bit more about the difference you found there? And then we'll talk a little bit more about what this all implies.
00;10;00;00 - 00;10;22;05
Dr. Clare Brown
So this was all derived because we weren't able to easily differentiate between subgroups within the black population. That is something that I hope to do with some future research to look more into whether there are differences among those who are sub-Saharan African versus Caribbean and so on, so forth. But we did want to at least look at those who were U.S.-born versus foreign-born, and we did that for each of the broader categories.
00;10;22;16 - 00;10;31;14
Dr. Clare Brown
And what we saw across all of the categories were that there are higher rates of low birth weight and preterm birth among women who were born in the U.S. versus those that were not.
00;10;31;29 - 00;10;57;26
Alan Weil
So that's yet another dimension where you can mask differences if you just look at the racial or ethnic group as a whole. Okay. So you're finding these variation, maybe before we talk about the implications, you've mentioned, for example, wanting to do more disaggregation in those who categorize as black. You mentioned that even in the Asian group you've got some categories, but there are hundreds that are missed.
00;10;58;04 - 00;11;06;27
Alan Weil
Can you just say a little bit about what data you did have them and how were you able to capture these subgroups, but not all subgroups.
00;11;07;24 - 00;11;30;28
Dr. Clare Brown
So this study used data from the National Center for Health Statistics that use the birth certificate data. And within the birth certificate data, there are defined categories that are collected on the birth certificate. So there is a standardized birth certificate. And it's we're currently on a version that was established in 2003, and each state is advised to use that standard as birth certificate per CDC (Centers for Disease Control) guidelines.
00;11;31;12 - 00;11;54;26
Dr. Clare Brown
And so each state collects those standardized options for race and ethnicity. There's an ethnicity question that does ask for, those have breakdowns of different Hispanic categories. There's also a racial variable that has different racial categories as well. There's technically other information in that data as well, such as where the mother's country of origin is if she is foreign born.
00;11;55;03 - 00;12;11;29
Dr. Clare Brown
But other than that, there's limited information. So if someone marks if they were, say, African-American, if that particular individual wasn't from a foreign country and indicated that she was born in a different country, you simply know that she marked African-American.
00;12;12;11 - 00;12;23;21
Alan Weil
I see. So you have more granularity than the broad racial and ethnic categories that we often hear reported. But there's still limits to that. And I guess in any data collection, they're going to be limits.
00;12;23;27 - 00;12;51;03
Dr. Clare Brown
Correct. And one example that's relevant to me is here in Arkansas, we have the largest concentration of Marshallese individuals in the continental U.S. by Northwest Arkansas and the native Hawaiian, other Pacific Islander does have a few subcategories. And there's an option for individuals from Guam, there's a Samoan option, etc. But unfortunately for us, there's not a Marshallese option in that classic standardized questionnaire.
00;12;51;03 - 00;13;01;12
Dr. Clare Brown
So we do have a way for us to identify that here in Arkansas for more local efforts. But the national standardized form doesn't have that Marshallese option in the standardized questionnaire.
00;13;01;25 - 00;13;25;04
Alan Weil
And that, of course, limits what you can do, which is sort of where I wanted to take our conversation. So you earlier talked about sort of what got you thinking about these issues. Now you have some results and you see these intra group differences. So how do these results affect your thinking about the racial and ethnic disparities we've heard about for a long time in adverse birth outcomes?
00;13;25;04 - 00;13;29;13
Alan Weil
What do you think differently now, now that you know what you know than you thought before you did the work?
00;13;30;04 - 00;13;46;14
Dr. Clare Brown
Honestly, the way I primarily think differently is that I have to do better. I mean, I have to do better in my research, in my advocacy, in my teaching. You know, we oversee a number of master's and PhD level students, and I'm going to have to teach them to do better than what I've done in the past. And I hope to do better in the future.
00;13;46;28 - 00;14;11;00
Dr. Clare Brown
I think that if the data are available, we have to do our very best to use that data in research. There's obviously going to be limitations with sample size and certainly making sure that we don't indicate any information that may allow for targeting of a certain population. We have to be careful in the way we display information and phrase things so that certain groups don't get targeted or ultimately face more discrimination because of what we present.
00;14;11;17 - 00;14;37;11
Dr. Clare Brown
But if the data are there, we really need to do better at using it as a journal reviewer or someone who advises different types of research. I think that I'm going to ask questions about that in the future. When I'm reviewing a journal article, if someone is making an argument for racial disparities, I'm going to say, well, were you able to look at more granular levels, maybe as a secondary analysis or an appendix or something, that that's something that should be thought about. In terms of advocacy
00;14;37;29 - 00;14;57;29
Dr. Clare Brown
I think most of my work would be able to be done locally, either here, I'm at a university institution, so is there anything we can do in the data we collect on our students or on our patients here at the hospital that could help elucidate more granular disparities that are faced by our patients and in our students, in the community members that we work with regularly.
00;14;58;18 - 00;15;35;07
Alan Weil
Yeah, I was really struck by your example of sort of language that, you know, we say, okay, for this particular indicator, people who we identify as Asian are not enrolling in the program as much, and so we need to target. But of course that category represents literally billions of people who don't all speak one language. So how do we think about not just as you already discuss the sort of the research and analysis side, but also on the policy design and advocacy side?
00;15;35;18 - 00;15;57;15
Dr. Clare Brown
So I think that a lot of this may come at the local level. Of course, we like or there may be options for federal policies that encourage more granular data collection. I do think that some of these policies can at least start at the local or state levels. I do know that there's at least one state, I believe, New York state does require collection of more granular data for Asian and Native Hawaiian.
00;15;57;15 - 00;16;21;09
Dr. Clare Brown
Other Pacific Islander groups, and that's specifically for that reason, so that they can have better surveillance. And I think that's a potential place to start, is to encourage states to identify what groups may be within their state and develop policies to better capture data, to have surveillance for the populations within the state and also to improve outreach.
00;16;21;26 - 00;16;57;22
Alan Weil
Yes. And, you know, there's been so much work to try to improve our collection of race and ethnicity data. We still have a long way to go, but we're going to have to go even farther if we want to be able to look at subgroups and do real analysis. That's helpful. I'm also mindful of the notion that for descriptive statistics, your sample sizes, you know, you can live with a certain sample size if you're trying to measure a program impacts where maybe the incremental effect of a program is quite small, to then be able to measure those by subgroups is going to get even harder and harder.
00;16;57;22 - 00;17;06;04
Alan Weil
So probably our ability to use these, what we can use these data for is going to depend in part on what questions were asked.
00;17;06;09 - 00;17;26;16
Dr. Clare Brown
Certainly. And I want to similarly highlight that I did use five years of data. So I mean, this is something I did have to combine lots of years so that I could look at these outcomes. There are some populations, particularly in that multiple race category, where someone could have clicked three or four different categories in their multiple race designation.
00;17;27;02 - 00;17;56;22
Dr. Clare Brown
Some of those populations are very small. We actually had to exclude five because they were too small. And so I think that's something to think about. An excellent point, and there are analytic methods to compensate for some of that, and that's probably out of scope for this discussion, but it's something to look into that there are potentially ways that we can mitigate small sample sizes, analytic challenges, etc. and that may be a good next step is to have a standardized way to do that so that we can use more granular designations.
00;17;57;01 - 00;18;20;01
Alan Weil
Well, Dr. Brown, thank you so much for doing this work and for providing a very precise quantitative picture of how much we lose when we use these broad categories. You were able to do it in this one domain, but I think it will get people thinking about it in others. It's really important work that I think will have long standing implications, not just in this area.
00;18;20;05 - 00;18;24;08
Alan Weil
So thank you for doing it and thank you for being my guest today on “A Health Podyssey”.
00;18;24;08 - 00;18;26;07
Dr. Clare Brown
I really appreciate it. Thank you so much.