Episode Transcript
[00:00:00] Speaker A: Hey, everybody. Welcome to the Zora Talks podcast. On today's episode, we're taking a closer look at sickle cell disease. It's a condition that affects millions, but is often misunderstood, and how new innovations like clinically guided AI are changing the way doctors predict and manage care. On today's episode, our guest, Dr. Rushing, will share patients stories and her vision for how AI could transform the future of treatment, not just for sickle cell, but for other health as well. It's an episode you won't want to miss, so let's get started.
Hey, Melinda. Welcome to the Zora Talks podcast.
[00:00:41] Speaker B: Hello. How's it going?
Good.
[00:00:44] Speaker A: I'm really excited to have you here today. And before we get started, why don't we kick off with you telling us a little bit about yourself and what brings you to the Zora Talks podcast today?
[00:00:54] Speaker B: Okay. All right. So a little bit about myself.
Well, I describe myself as an applied data scientist, and I say applied instead of just not a data scientist because I can't get deep into the weeds of the technical side of things. But when it comes to utilizing these tools in, like, real world settings, that's where my expertise comes into play. And so currently I am an assistant professor at Rutgers University in the health administration program.
And so what brought me to Zora Talk's podcast today is that I am working on developing an innovation we're calling clinically guided AI.
And I'm just excited to be able to talk about what this new approach that we're developing and just start to really start this discussion of how do we utilize AI in the healthcare system.
[00:01:46] Speaker A: Man.
[00:01:46] Speaker B: Awesome.
[00:01:47] Speaker A: And like, exactly what we're looking to talk about on the podcast. So I'm excited to dig in, but before we dig in, I would really love for you to explain what particular applications you're using the clinically guided AI. I know you said sickle cell, and I would love to give our listeners a little bit of explanation on sickle cell disease in simple terms.
[00:02:11] Speaker B: Okay. Yeah, so this is a great place to start because it doesn't really make sense unless you understand the disease that we're talking about.
So sickle cell disease is a genetic blood disorder. And what that means is that it changes the way that red blood cells are developing.
And so typical red blood cells, they're circular and so they move through the body easily. They're able to carry O throughout the body to different organ systems. But for individuals that have sickle cell disease, the cells are not developed circular. They have these sickle like shape, hence the name Sickle cell disease.
And because they're not developing correctly and they're taking on this sickle shape, they actually don't carry oxygen throughout the body as well, which is why it's also called sickle cell is sickle cell anemia. And they also increase. They're very sticky and so they clot easily. And they result in just a lot of complications for those individuals that are living with that disease.
And really, if you just think about where does the blood go? That's every organ, every area of the body that this, that these sickle cells can impact.
And so people that are living with this disease have an increased likelihood of stroke, organ damage, of course, like I said, anemia, because it's not carrying oxygen throughout the body as well.
And a slew of challenges that they're facing. One of the main complications of sickle cell are these crises called vaso occlusive crises. And this is. These are severe pain episodes to where these individuals often have to go to the emergency department in order to receive an IV pain medication.
And if you think about just kind of like how things have come along when it comes to opioids and like being prescribed because this disease has, is the main complication is pain. And because it is so severe debilitating there, this patient population often are needing to utilize opioids. It places them. There's another risk that's coming that plays a role when it comes to their health. So the first one is just the disease itself, and the other one is a system that does not want to overuse opioids. And then the last factor with this, as I talk about like kind of the foundation of the disease is that this disease is predominantly found amongst people of color, specifically among people of African descent.
And there's like, there's theories around kind of why that is the case. And it kind of lies with like malaria and where malaria cases were, were at more prevalent. Because malaria can't actually survive in cells that have this genetic sickle, the sickle gene.
So anyways, it predominantly impacts individuals with color. So on top of just having severe pain, epis needing to utilize a treatment that has a lot of concern around overuse. It also is affecting populations that are marginalized in western societies. And so all this plays a role in why it's so challenging to manage this disease and why for me, sickle cell has become the center of my research. Because I desire to support this patient population and really identify ways that we can improve the way we're treating the disease.
[00:05:35] Speaker A: I did also want to Talk about the fact that we missed Sickle Cell Awareness Month. That was in September. When I was doing the research and I heard it was in September. I was like, darn it, I wish we were actually having this conversation in September. But I also know, just like Black History Month, it is important to continue the conversation. I know it's a lot different in Black History Month, but it's still very important. As much as important as Black History Month, if not more important than. So why is it important to continue the conversation all year long and not just wait for one dedicated month?
[00:06:11] Speaker B: Well, yeah, I think it's. With any kind of disease. Like, if it's chronic, you know, sickle cell doesn't only, like, flare up or, you know, need to be treated during September. Like, this is. This is a life. This is a lifelong illness that individuals are living with.
And so we may be aware of it during September. And also, you have one more opportunity, though, because in June, or in June, Juneteenth, actually, which it does, it happens to fall on the same date. But that's. This is not why this was developed. But Juneteenth is also World Sickle Cell Day, or World, but I first heard it was World Sickle Cell Education Day, but it's a World Sickle Cell Day. So we have two opportunities per year to, like, raise awareness to focus on sickle cell. But anyways, why isn't, like, we just, again, like, these are individuals that are living with this disease day in and day out. And so whether we want to focus on it or not, they are still having to navigate these systems and deal with biases in our healthcare system, deal with the barriers they have to, like, living the type of lives that they want to live. Even, like, the bias associated with having the disease as well, on top of just, you know, having to deal with the different complications and challenges of daily medications, all these type of things. So advocating for them year round, supporting them, listening to their stories, be learning about this, contributing to the cause is so important because it has to be a continuous effort in order for us to start seeing a change in what's happening. Like, we're. We have a national.
I'm sorry, we. Sickle cell is included in the newborn screening across all 50 states now because people continue to advocate and push to have the screening included when newborns are. When we have newborns in the hospital so that we can identify from birth who has this disease and get them into treatment and get them to support the parents and the child needs in order to increase their likelihood of survival.
Throughout childhood, into adulthood. And so now we need that advocacy during that transition stage. We're getting them to adulthood, which is great, but life doesn't stop at 18 or 22 or. I mean, life is kind of just beginning, you know, like, it's. You're starting to have this independence. You're starting to, like, you know, pursue, like, what you want to do, like this. Like, this is when, like, life is starting to begin.
And it's. It's heartbreaking to see that this is also when mortality spikes for this patient population. Like, that's where that's not okay. And. And the reason why we're able to get the babies into childhood and young adulthood is because we continue to advocate. We got to continue to advocate for those in this young adult stage. And that's just one component we need to advocate in order to get gene therapy out to where it's more universal and available to individuals that qualify for it. We need to continue to look for more treatments that can modify the disease for those that don't qualify for or don't meet the criteria in order to get these different curative methods. We need to advocate for every hospital system throughout our society to know what sickle cell is, be willing to treat sickle cell, and have the expertise needed in order to understand and capture or identify these different complications that can crop up beyond just a pain crisis. So, like, it's. This is a. For me, it's like a whole career being dedicated to it. Like, there's many others, but there's a component that each person can take on when it comes to, like, contributing to supporting our brothers and sisters. And I'm not just saying it's because it affects people of color. I mean, just as. As humans, these are our brothers and sisters.
Excuse me. And we don't want to see our brothers and sisters suffering because we wouldn't want somebody to sit there and watch us suffer. So if we can contribute in some way, we should, just like we do for childhood cancer, breast cancer, diabetes, cardiac disease, drug prevention. You know, like, people advocate it and care because it touched their families. And this is a disease that's just touching our loved ones.
[00:10:39] Speaker A: Yeah, thank you for sharing that. And I, you know, I share your sentiment, but, yeah, I appreciate the emotion behind what you said, and it's super important for us to embrace this and see what we can do to support it.
And I did want to say, also, I'm glad you mentioned the two dates that we can spread awareness, because this podcast lives on.
[00:11:05] Speaker B: Right.
[00:11:05] Speaker A: It's not just A once and done kind of thing and I can resurface it during September and also on Juneteenth. So I'll put that on my calendar and hopefully be able to share this podcast again on those dates.
I'm glad we were able to get that awareness out there about the actual disease, sickle cell. Hopefully some of our listeners will advocate for it and also spread the awareness around it. But now I'd like to pivot and talk specifically about something you developed called clinically guided AI.
So for a non technical audience, how is it different than traditional AI?
[00:11:48] Speaker B: What this approach is doing is a pre summarization of information before the AI models access it and start drawing insights or inference from them to explain that a little bit more. When you're looking at clinical notes or a provider's way to look at a patient's clinical notes, there's a lot of information that's there.
With large language models like ChatGPT, they're able to find these different patterns because there's a whole lot of data that's being given to them. But when you have a small patient population like sickle cell, it's more challenging because there's not as much data. So a provider, if you have the LOM and or and you're putting these notes in for you have these notes around sickle cell, it's not enough information to really gather what's going on. And so it may not be as accurate the inferences, insights that the model provides.
So to improve this, the approach, the clinically guided AI approach actually goes in, uses like a bunch of predictive models to go in and extract the specific things that related to sickle cell that providers and nurses and clinicians typically need in order to inform the way they treat the disease or treat the patient. And then the AI pulls from that information to then start drawing out, drawing inference from it. So instead of saying here you go Chad gbt, here's all these notes, tell me what they mean.
What we're doing is saying is we'll take the notes, start pulling out what we know is important and then say, okay, Chad gbt based off of this information, what's going on with this patient? How severe is this patient's disease? What does the severity mean? How can we intervene?
Is supporting or guiding the AI? Instead of saying having the AI look for these different patterns on its own, we are giving ChatGPT a summary of the information that we want ChatGPT to look at so that ChatGPT can do a better job at telling us what we Want to know with the clinically guided AI, what we did is we have three parameters that we want to pull from the clinical notes. The first one is disease severity. The second one would be a summarization of key lab trends. Then the third one will be these five year risk forecast. So we decided on these three parameters because when it comes to the disease severity, clinicians are wanting to understand at this point in time, how severe is my patient or how good or severe is my patient's disease and where is their health, so that they're able to then determine what interventions or what treatments they need to continue or discontinue or to seek out next with the lab trends, because you're looking at months or years worth of clinical notes, but providers can miss little changes that are there that our model can pick up. So providing these summarizations, clinicians will be able to see, oh, you know, six months ago they say their creatinine was this level, but even though it's still the normal range is starting to tick upward. So maybe we need to look into that and see if there's something going on with their kidneys.
And then the last one with the five year forecast, the goal with that was to give a give provider or clinicians an idea of what could happen, what negative outcomes could happen in the next five years so that they can intervene earlier and hopefully mitigate or prevent these outcomes from taking place, because they know this ahead of time. And so all this was the goal is to support this kind of preventative approach and things, let's not wait till it gets severe, let's start to identify this early on.
And I think that when it comes for the patients, their providers having this type of support when they're approaching their treatment plans or developing their treatment plans, in my view, will only enhance the way that they're able to manage and treat their disease, which ultimately will improve health outcomes for those living with the illness.
[00:15:53] Speaker A: Wow.
[00:15:54] Speaker B: Yeah.
[00:15:54] Speaker A: So it's so great because it sounds like you will be able to see if there's anomalies in the labs.
You're able to give people a better sense of where their disease is. With the five year predictions and knowing how severe their disease is.
It seems like there's wider applications of this beyond just sickle cell. And I don't want to say just sickle cell, but it seems like this is an opportunity to really expand your research beyond one disease and maybe other diseases that may benefit from understanding these anomalies and predictions as well. Is that you think is a possibility Absolutely, yeah.
[00:16:37] Speaker B: And that, that is the long term goal. Like for, for me my research is around sickle cells so that's my domain expertise and so I'm applying it here in this, it's kind of like a pilot at this stage and so we want to, we want to, we want to build this out to. Because we have the insights as to the things we need to look out for. But we, I also foresee this as a template for other diseases to apply to how they're approaching managing well integr to the clinical workflow and how they manage their patients in a way that supports them as opposed to just like I said, like traditional way methods of here's the clinical notes or here's information, you make inference, you draw inference from it. ChatGPT but guiding it so that it's actually clinically relevant.
We actually are, have started the process with, at Rutgers University to pursue protection for this, for this approach and we also, within, within my lab the goal is for us to develop this product and to get it, take it to market and to get it out there so that it's not just like a paper that we, that's published. I mean the goal, yeah, we will publish papers on this innovation but we want it to be used as largely as possible after we validate it and we're able to show that okay, yes, this is, this has a positive impact in a clinical setting. So at this stage we are at the validation. So we built the code in order to, to do this and like the algorithm and so now we need to validate it using a clinical expert panel. And so yeah, it's, it's exciting to be at this stage but also, I mean it's going to take some time. We have to go through it in order to make sure that we're actually contributing in the way that we want to contribute. We are looking to partner and collaborate with venture capitalists or whoever really that would see this idea and want to champion it or collaborate with us in order to get it out on a larger scale or maybe integrate it into something that's already being done. But we're answering another problem that this, this system or this model or this software or whatever is produced or addressing.
[00:18:58] Speaker A: Can you walk us through one of your case studies? Like the 20 year old woman with the severe complications?
Can you explain how your AI model added insight that might have been missed otherwise?
[00:19:11] Speaker B: Yeah, yeah. So in this case it was a 21 year old female who had just had multiple like pain crises as well as some other complications.
And so when you're reading the case, you just kind of see what's going on. And so what our model did, when we ran this case study through our model, it was not only able to pull out these complications that you could see as you're reading, but then it was able to assign a disease severity score based off of the complications that were described at other facets. Like I believe that this young woman also smoked. And so she said that smoking was a trigger of hers. So like it pulls, it doesn't. It does better not to overlook or miss these different contributing factors that could lead to how severe their disease is at that moment.
But it also connected that disease severity to a five year prediction which show like things that the provider needs to be concerned with, which I think maybe one was like future transfusions or maybe it was like increased hospital utilization based off of what we're seeing at this point so that a provider can intervene early on. Now if we didn't have our model and you just kind of read the case study, your one provider's what is it? Rating of the disease severity could be different from the next provider. So it's kind of like ours is giving this standardized way to put like this objective measure on the disease severity, which again, we're at the validation part so we can do it in code, but we need to validate and make sure that we're doing it correctly. And then also with those predictors, predictions or those forecasts, as you're reading this, some of these things that could happen down the line, one provider may see, another provider may not see because maybe they don't have that kind of exposure, maybe they haven't seen these type of outcomes in sickle cells before or they may not be familiar with the disease at all.
So if you take our, what is it, if you take our model out or our approach out of it, then you just have this one provider that's approaching from only what they are, what they know at the time, which can be a lot of experience, or it could be a little experience or somewhere between. When you put our innovation in there, you now are getting a standard, more standardized way that's validated by clinical experts in order to get an idea as to what the patient's condition is at this moment, what could come down line so that they can then make decisions on how they can treat.
We are looking to partner and collaborate with venture capitalists or whoever really that are that would see this idea and want to champion it or collaborate with us in order to get it out on A larger scale or maybe integrate it into something that's already being done. But we're answering another problem that this system or this model or this software or whatever is produced or addressing.
[00:22:21] Speaker A: And if our listeners want to support your work or help move this forward, what's the best way they can get involved?
[00:22:28] Speaker B: Yeah.
So what's next? Again, like this validation piece for this project is the next stage. And so we're at this, we are applying for different funding, looking at traditional versus non traditional and non traditional methods of being able to fund this process as well as I've been talking with in different medical or clinical circles. And so we have some institutions that are interested in, which is exciting because being able to collaborate, not just for me, I'm in Jersey, so not just in Jersey, but seeing and applying this in other areas is only going to strengthen these, strengthen this approach and expand the region that it has.
So, like, if someone is interested, wanting to know more about it, wanting to again, like, like collaborate again, we are all, we are looking for collaborations.
Somebody's wanting to like, support it with resources. We're looking for that as well. Definitely link reach out to me on LinkedIn.
I, I'm on LinkedIn pretty regularly and yeah, and I would love to talk more about it.
[00:23:39] Speaker A: Wow. Well, I'm sure that there's somebody out there that you're going to be hearing from soon, so let me know. And I, you know, hope that you are to get some really good research and connections out there.
So final question.
If you could leave our listeners with one bold prediction or call to action around AI, what would it be?
[00:24:02] Speaker B: This. Yes, I like this question and I have a call to action.
View AI as a tool and use it as a tool.
Like all the doom and gloom around AI, don't personify it, see it as is. It is a tool that can improve things and use it accordingly.
[00:24:23] Speaker A: Awesome. Love that. Yes. And I feel like we're going to be healthier because of your research. And so thanks again for joining us on the ZORA Talks podcast and we'll be following your work.
[00:24:37] Speaker B: All right, thank you again so much. I really enjoyed the conversation and opportunity to talk about things that I'm very passionate about.
[00:24:44] Speaker A: Awesome. Thank you. Thank you again.
That's a wrap for the Zora Talks podcast. If today's convo put new ideas on your radar, be sure to follow the show and share this episode. New episodes drop on the 15th of every month.
For show notes and links, head to ZORA Digital Online and connect with us on social media.
Zora Digital is a digital marketing agency agency based in Chicago, delivering AI driven strategies with human centered results.
I'm Yawande. Thanks for listening and we'll see you next time.
