 Hi, everyone. Welcome to the third talk in this session. This is by Joy Payton and Paulette McRae. The title is Data is Not Neutral, Biomedical Data, White Supremacy, and What You Can Do. I'm excited to be here. Paulette McRae is having some technical difficulties. So I am going to try to do her side of the talk as well, and hopefully she'll be able to join us shortly. So Paulette McRae is a neuroscientist at CHOP who works in research administration, and I am a data educator who works in the ARCIS program, which you may have just been hearing about in the previous talk. So I'll quickly go over this slide. This is my colleague Paulette as a child, and I'll let her talk about it when she's able to get into the meeting. But what's interesting about this photo is that Paulette represents a lot of the underrepresented groups that we know are not fully represented in science. As a woman, as a black person, as someone growing up in a family in which college was not typically expected, she really represents a lot of these groups. And this is a drawing that she did as a young child, as you can see with all of the body parts. And what's interesting is that she later went on to become a brain scientist, and the brain is not even depicted there, which we both found sort of amusing. So Paulette and I want to share with you a little bit about why data is not neutral. There is the expectation that we have sometimes that our data is somehow more pure and more representative of science than we are, right? That somehow data is neutral. But data is not neutral. And we're going to talk about two aspects. Unequal, yes. Oh, perfect. I'm here. I don't know what's going on. I apologize. Joy, thank you for getting us started. So I, you know, I think you did this slide already. So I will just jump in right where you were. Perfect. So we will be talking about why data is not neutral. And specifically, I want to talk about two out of the many reasons that we could cover why data is not neutral. So one is representation. So that's really the talent around the table not being representative of the diversity within the population. Therefore, we're developing research questions, designing research studies, recruiting research subjects, and executing and analyzing studies where there is a lack of representation. The second point that I will be talking about is the lack of trust between the black community and the medical community. So we can go to the next slide. So here we're looking at data from the NSF, the National Science Foundation, for science and engineering positions. On the right hand side of your screen, you can see the percentage representation for the US population. On the left hand side, you can see the percentage within science and engineering positions. And there are a couple of things that stand out that certain groups are not proportionally represented, such as women in general, black men, black women, Hispanic men and Hispanic women. So there's certainly an urgent need to close these gaps. And these gaps are present not just in science and engineering, but across all STEM fields. So the next slide, please. So in terms of unequal representation, I think that's something that most of us on this call are familiar with. So I'm not going to spend a ton more time talking about that. I really want to go and spend some time talking about the lack of trust between the black community and the medical community. And this is where I'm going to spend the bulk of my time. And in part, that's because the issues around trust are often less identified as a driving force behind the underrepresentation that we see in medical studies. However, this mistrust is real, and it is justified. So we can go to the next slide. So one thing that we do know is there is lower participation of African Americans across studies, and that is regardless of the study type, as well as the type of disease that's being studied. So this includes studies on AIDS, on Alzheimer's disease, on different types of cancers, stroke, cardiovascular disease, and the list goes on. So what's really driving that is a mistrust of academic and research institutions and the investigators driving these studies. And that is one of the most salient and significant barriers for African Americans and their lack of participation in these types of studies. So we can go to the next slide. So one of the driving factors of mistrust is unethical research practices. And here I'm just laying out two examples. One is the Tuskegee syphilis study that took place between 1932 and 1972. The other is the case of Henrietta Lacks. Hopefully, most of you are aware of these examples, and in the interest of time, I'm not going to go into detail on either one of these. But if you're not familiar, I do encourage you to look them up and get familiar with them. But these are two prominent driving factors as to why African Americans tend to not be involved in research studies. We can go to the next slide. But what I would like to do is spend a little bit of time looking at some less well-known examples of unethical research practices. And first, I want to point out the term iatrophobia, which is the fear of the healer. So this fear within black communities is real and it's justified. And hopefully, as I go through a couple of examples, you'll understand why the fear is real and why there is true justification for this fear. So if we take Dr. Hilton, he actually did a number of studies on a slave named Jim Fed Brown. Jim was subjected to various experiments, including being put on top of a hot pit of smoldering coals to study the effects of sunstroke and ultimately potential remedies for that. He was also subjected to blistering of his. I think there was a little hiccup, but I think we're all back, right? Okay. Are we good? Okay. When we look at the cases of Dr. McDowell, he is credited as the father of abdominal surgery. He discovered the overreactomy and performed a number of these procedures experimentally on enslaved women. And the first case done on a white woman is the one that's well documented. So that was done on Jane Crawford, but no one mentions all the enslaved women that were subjected to experimental procedures before her. The other example is Dr. Sims. So that's the father of modern gynecology. He pioneered surgery that really treats some of the complications related to childbirth. But again, there's very little conversation about the way he got to those successful surgeries. There were a number of enslaved women that he did experimental surgeries on, one of which was anarchist. He was a 17-year-old girl. He did over 30 experimental surgeries on her alone. Those surgeries were done without anesthesia. And the driving force there wasn't because they didn't have anesthesia or access to anesthesia. It was that there was a prevailing belief that blacks did not feel pain or anxiety. So there was no need to use anesthesia on them when you're doing these types of procedures. He also did experimental surgeries on black children. So we can go to the next slide. So you can say, okay, well, that was the 1850s. That was a long time ago. So let's look at things a little closer to modern times. And we can look at some experiments that were done on prisoners. So Dr. Stanley did a number of experiments between 1913 and 1951 on prisoners in San Quentin in California. Those experiments that he did included experimentation with sterilization, potential treatments for the Spanish flu, and most disturbingly, some testicular transplants that he did between prisoners. We can also look at Holmesburg Prison in Philadelphia, where Dr. Kligman, a very well-known and prominent dermatologist, conducted a number of studies between 1951 and 1971. He's actually the founder of Retin A or Retinol A, which you guys may be familiar with. But he did a number of skin tests for very shampoos, foot powder, deodorant, and that may all sound very benign. But when you think about the types of experimentation that were done on these prisoners, to test these types of products, a lot of times the skin is exposed to heat lamps and other types of manipulations to the skin to test how these products work and if they're safe. He also later in his career experimented with mind-altering drugs, as well as dioxins, which are a group of persistent chemical pollutants. So we can go to the next slide. And to move even closer to present day, there was a number of studies looking at, and I'm going to talk about one in particular, the genetic link to aggression in black boys. This study in particular required the withdrawal from all medications that included asthma medication, asthma is something that's very prevalent in the black community. It included ingesting a low protein diet and overnight stay without parents. These are boys that are ranging from 7 to 12 years old. With holding water, hourly blood draws, and administration of fin fluoramine, that's part of the drug that got pulled off the market fin fin. This drug is known to increase serotonin levels and be associated with aggressive behaviors. So these are just a few examples of some of the unethical research practices that have been going on and this study was done in the 90s. So we can go to the next slide. The problem and the driving force with this mistrust runs deeper than just impacting black individuals that participated in research studies because you can naturally say, well, just don't participate in research studies, but it's deeper than that. There is and there was and there still is rabbit mistruths and biases in medicine. So we can look back to 1851 where Cartwright was commissioned to write a report about the cultural nature of the black race. So naturally at that time, he's talking about enslaved black people and the way he did visits was observing them. And what he concluded was that black people have a lower lung capacity and having them forced to work by white people was good for their health. He also came up with the psychiatric diagnosis, drapetomania, which is the disease that causes neighbors to run away. So really saying that any slave that was trying to escape was doing that because they were suffering from a mental disorder. So again, that's the 1850s, but we can come much closer to present day and see a number of examples where there are mistruths and biases in medicine. Some including the undertreatment of blacks as it relates to pain. So 43% of blacks versus 26% of whites receive no pain medication for long bone fractures. Those are known to be extremely painful. In addition, you may think this is just related to adults. Black children are also subjected to this lack of pain treatment. So black children with appendicitis and severe pain were less likely to receive opioids than white children. The silver lining here is this is probably the driving force why the opioid epidemic is not as rapid in the black community as it is in others. And if we look at a study from 2016, they looked at medical students and residents and found a number of them hold false beliefs about biological differences between blacks and whites. Over 50% believe that blacks have thicker skin and this actually goes back to the 1850s studies where they were trying to understand the difference in black skin and white skin and subjected slaves to blistering experiments. There's also a belief among over 25% of medical students and residents that blacks have less sensitive nerve endings therefore don't perceive pain the same way. We know both of these are not true, but again this is still very prevalent in medical training in medical society. So we can go to the next slide and with that these are two reasons out of a number of factors that are not covered in this talk that are really driving part of why data is not neutral. So we raise the case that it's not neutral and at this point I will turn it over to Joy who can talk more about where we go from here. So first of all I want to speak about the term white supremacy which can be a very scary term for people and I just want to highlight that white supremacy is the term that I use and many people use because whether or not I am an intentionally overtly racist person I have been shaped by societal influences that subtly tell me that white people are better than or superior to non-white people and this causes harm. So what do we do with our data? Sometimes we just have to reject data because it's been tainted by white supremacy. For example in this infamous Boston housing data set that I was given as a graduate student in which there is a single variable that carries a lot of predictive weight and I won't go into why this is an unacceptable data set but just keep in mind there's probably a lot of co-linearity there's problematic language and not all races are included so the data is impoverished. But if we want to keep our data and not just throw it out some of the things that we can do is ask ourselves if our findings are truly generalizable we can statistically describe our lab staff and we can normalize the disclosure of bias. And this is an R conference so I actually did use R to look at my own biases and my own complicity in white supremacy and realized that I did a Chi-squared and did a visualization and that information is right there for you to take a look at. So we can use R to take a look at that and I think that's all the time we have I think I just heard the bell ring. So there's a few more slides that you'll get to see when you take a look at the slides that are sent to you if you're interested in that. Thank you so much everyone. Thank you. So part of the reason that we stopped this a little bit early is to encourage questions and conversation. So I guess as a first question it seems so the unethical treatment so the unethical treatment that was shown in the beginning it seems like there's almost this movement from these egregious experiments to more this issue of neglect in terms of treatment. Is that a fair thing to say? And then that probably becomes a more difficult thing to be able to point out and to deal with. So what are the strategies for kind of compensating or dealing with that? Yeah so in terms of the strategies I think the first thing that has to be done is acknowledgement. You know acknowledgement that these things happened that the fear that people have is justifiable. There are also studies out looking at the treatment that blacks receive for the same exact symptoms relative to whites going to the same positions. So they're not treated equally. Their symptoms aren't treated equally. Their course of medications a lot of times is not the same. So there are a lot of inequities and we're seeing a lot of that actually play out now with the COVID-19 as well as the increase in maternal death that we're seeing in women of color. So how do you get over that? I think first you have to acknowledge that it's real and then you have to make sure that the people that are being trained to administer or be the medical professionals are aware of this history and are aware of the innate biases that they have or that they may have. And once you start to address that in addition the other thing you want to do is you want to make sure your medical community is representative of the community in general. So diversifying medical professionals, diversifying people that are doing the experiments who will likely bring a different perspective to the table. So I think those are some of the strategies that could be put in place. Okay. All right. So another question we have is how do you feel about this in terms of COVID where they're trying to push out the vaccine when minorities may not be well represented? Yeah. So that's a great question. Not only may they not be represented in terms of the trials. Again, with this distrust within the black community of the medical community, I think there will require a lot of convincing that this is okay, this is safe. We can look back at the polio vaccine where that was rolled out quickly and erroneously people were actually given active cultures of the polio disease instead of a vaccine. So we have this really turbulent past with certain things. So I think it really is going to require a lot of education and outreach on part of getting the black community on board with taking a vaccine. Okay. Great. Thanks. So another question is that when we do data analyses, especially in the medical context, a lot of times race is one of the variables that we use and we see significance. And this becomes something that we, a lot of times have to detangle, right? Because race is a lot of times a proxy for a couple of things. It can be behavior or sometimes it's even a, it indicates a genetic variant. So how do you think about those when you see that there is significance but it's not very well characterized? Joy, I would punt that one to you. A lot of the things that we can do are do school analyses of co-linearity. So which of our variables, which of our features are related to one another? So, you know, it's not surprising that the Boston housing data set includes race, problematic language aside, because we do know that race and housing prices are related. But what we have to ask ourselves is why is that? Is it practices that have led to the gradual and insidious destruction of non-white neighborhoods through neglect, lack of services, redlining, criminalization of blackness, criminalization of poverty? So instead of just looking at race and making race carry so much weight, let's look at police presence. Let's look at how many supermarkets there are. Let's look at college completion. I think we settle too easily for some demographic markers when we could go deeper into things like the Community Disorganization Index, which measures a lot of different socio-economic status indicators. So I think like statistically, don't be lazy. Do some feature engineering. And because we do know that race is important and race carries a lot of predictive power. But why does it carry predictive power? I think we need to go a little bit deeper and understand what are the variables hidden behind the variable of race? Okay, thanks very much, Paulette and Joy. We're out of time now. If you have more questions, you know, you can contact either one of them. Thanks. Thanks. Thank you so much, everyone.