Unpacking Racial Biases in Healthcare and Data Science

Lauren Brown is an assistant professor at the USC Leonard Davis School. Her research uses publicly available data to uncover the unique difficulties Black Americans face in maintaining physical and psychological well-being as they age. Her lab both challenges the methods used to study older Black adults and strives to increase diversity in data science research with the goal of increasing the visibility of Black and Brown people via data and storytelling.

On the role of racism in biomedical and statistical sciences and disease prediction

If you think about the history of statistics and where it starts from, the earliest statisticians were actually also eugenicists. And a lot of it stemmed from the fact that Black people at the time that the census had started were property. And it was a way to count and keep up with property until we get to a point in the early 1900s when we start recording actual race in the census and colored being one of the options that you could check. And that being a way we kept track of Black populations, unfree, Black populations in particular, but also freed as well. And that transition of having Black people in the census started what was eventually used as studies that were confirming or trying to confirm biological and genetic inferiority among Black people.

So once Black people were started to be included in the census and started included in medical research, clinical research, that research was usually often to compare Black people to white people with the innate goal to say Black people had more muscle mass biologically and genetically or smaller brain circumferences and justify it would a way to justify slavery by suggesting that the biological and genetic inferiority was a part of how Black people became slaves and would justify their continuation as slaves. So you fast forward to today that legacy of, of genetic and biological inferiority in medical, and statistical analyses has now manifested in things like race norming, where we're actually saying like, there are adjustments we use for Black patients in the clinic to justify whether they do or do not qualify for care strictly based on race. And a lot of it is based on false statistics that eugenicists had originally been pushing in the early 1900s.

How injustice through data and storytelling affects the health and wellbeing of Black Americans

When you think about like an individual, how this may affect one individual Black person, like for example, if we think about George Floyd's killing in 2020, his death originally was considered in the autopsy report performed by the medical examiners due to prior health conditions. They originally blamed his underlying health conditions and drug use as the cause of death. It was only after the family got an independent autopsy that they were able to show that the death was a homicide that then implicated Derek Chauvin and the Minneapolis Police Department, as responsible for the death and the knee on the neck. So this idea of blaming Black biology, is something that persists, I think, in larger society and that the biological inferiority is the cause and the precipice for Black death, and that it's not at all the function of society when actually now we know, you know, based on a lot of great research that the social environment is much more responsible for the fact that Black and Brown people often live shorter lives than white people or any other race and ethnic group in the US. We often live with more disease and disability at the end of life. And a lot of that we know is now it's social conditions, it's discrimination, it's racism, those are at the forefront. But the research doesn't always follow that line of thinking because of the history and the legacy that still exists that we're still combating. And this new level of science is trying to push up against this idea.

On diversity in population studies

It's been really obvious that a lot of the measurement and the things that people use to measure the wellbeing of Black life is really centered in white populations. And it's not innate or particular to the lived experiences of Black and Brown people. And so I think oftentimes we miss the real story that's happening up underneath a lot of Black health and aging specifically because those studies weren't designed just for Black people. They were designed for all aging populations and to monitor the aging of populations over time.

The ethical considerations if you're leaving a whole group of people out or if you're not intentional about measuring their aging, is that you're not able to predict their clinical progression or able to assist their aging process in a way that's meaningful for them. We're doing everything much better for white populations than we are for minoritized populations. And so that the injustice is embedded in the structure of how these studies often come about. And the intention around what I want to do in this work is to help magnify the voices of Black people in these studies so that they more accurately represent the aging experience so that we can get better at predicting disease, preventing disease, and ensuring better aging process.

On the Linked Fate Data Collective

Linked Fate Data collective is a group of activists, of scholars, of students, of people who are interested in expanding their data science tools in order to promote the accurate depiction of the aging and the living process or the lived experiences of Black and Brown people. The idea being that, you know, most of the data science spaces are very white and male and often then reflect the values of people who are white and male. And I am very passionate about creating a space that looks and feels different for the people that I would love to bring into the data science realm. And you know, how we do that, I think, you know, there's a lot of argument about the pipeline issues of how we get people into data science or how we get people the skills to be able to do this on how we get Black and Brown people interested in data work.

The inception of the name Linked Fate comes from a term that was originally used in African American studies. And the term was referring to block voting in Black populations where African Americans vote primarily Democratic with this idea that, you know, their fate is connected to the fate of the larger group. And so, there's an interest in finding a collective voice in order to impact change and power. And that's really what I named this space after is that we have collective voice in data and it's the power of an individual magnified by many that gets people something that's powerful with the data work. And so that's really what this Linked Fate Data Collective is trying to do, is bring underrepresented groups and people and their ideas into a space that will honor the data science that we want to see in the world. And that is not perpetuating scientific racism, that's not perpetuating a lot of the genetic determinism and the things that some of the current science and medical and clinical spaces are perpetuating.

On the Black mental health paradox

One of the things I like to do in my work is move away from these disadvantaged narratives that really plague the aging story of Black Americans. Most people are very interested in the weathering and accelerated aging of Black Americans, when really there's a lot of trends that suggest that's not the only way that Black Americans are aging. That it's not just weathering stress aging faster, that there are also a lot of other processes that don't act so linearly. One of them is that mental health paradox, which is this data artifact that has been found in like five nationally representative samples now that despite having higher stress burdens, despite facing discrimination, despite having lower socioeconomic status, so lower education, income and wealth and despite having worse physical health, Black Americans have lower rates of depression relative to white Americans.

So this could exist for many of reasons. It could really be a data artifact and it just could be that we are not measuring either mental health and depression in Black people in the way that it manifests so that we can measure it. Or it may be that we're not measuring the stress that's most impactful for Black Americans. And so we're not really capturing the stress burden. And so, we don't understand how that translates to mental health. And a lot of the work that I'm doing on the paradox is in that exact realm, which is that the stress experience is not being fully captured for Black Americans. And it's not acknowledging the coping response that Black Americans can use in order to fight the adversity that they're facing. So, my idea here is to restore agency to Black people. That you're not just the sum of your stress exposures, you're also able to react and respond to those. And we have a lot of agency in responding to that and a lot of historical agency and a lot of lessons generationally passed down. And that's a really important way to acknowledge both the incredible hardship that Black Americans face in this country in growing old, both psychologically and physically. But it's also acknowledging our ability to fight back at the same time. And it's already happening. You know, it's not like we need an intervention for it or something else to do for it. Black people are already doing this and you can measure that. So yeah, it's a cool project.

On the Fatal Encounters research project

So motivated by the George Floyd murder in 2020 me and a colleague, Dr. Terrence Keel at UCLA recently got a RSF, Russell Sage Foundation grant. We're basically going to this data source called Fatal Encounters. It is a data source that crowd sources all of the police involved deaths that have happened in the United States. So, we are going to this data source and we are looking in LA County and we are finding the names of people who've been involved in police related deaths that have not involved firearms. That's because firearm deaths are very straightforward, can typically labeled the death as a homicide because you know, the act of shooting. But for non-firearm deaths like George Floyd, those are more arbitrary and harder to prove homicide and the autopsy reports can be very misleading, especially by the medical examiner and the coroners. And those autopsies are public in LA County. We're taking the names going to get all of the autopsies from the medical examiner/coroner. So, we have like 320 autopsies from 2000 to 2020, and we're trying to create a data set that represents how people are being classified in terms of cause of death and if there's any other indication of, you know, markings on the body some type of conflict that happened during the process. So, it's any interaction with the police out on the street or in LA County Jail. So, we have both of those data sources and we're able to try to say something about what's happening to a lot of these people, especially Black men whose moms are also very interested in understanding what happened to their kid. And so, the project is really motivated from that space.

On the lack of diversity in genomic data

In genetic sciences, you know I think 80% of our genetic and genomic data is from European ancestry populations, even though only 16% of the world is European ancestry. So, there's this huge imbalance in what we know about genetics because we only know what's happening among European populations. It's not, they're even telling people right now to not do genetic work in Black and Brown populations because we're not sure what we're finding is accurate because we don't have good training data. And the way genetic sciences work is that training data, everything is based on a reference population and a training population. It's not dissimilar from early eugenics where everything is compared to whites. You're constantly comparing Black and Brown people to white people. And if that's the way you're starting, it's going be a story that's rooted in inferiority and rooted in comparison and not necessarily rooted in the true story that should be unfolding that you can unfold when you're not trying to make those comparisons. So that's happening really horribly in the genetic sciences where you have dominant European frameworks and genetic data. You're trying to say something about other types of people and it's really not working, and scientists know that, but they're continuing to just do work on European populations.