For this episode of Karger’s The Waiting Room Podcast, we spoke with Emily Lewis about Rare Disease Day on February 29, and how artificial intelligence (AI) and digital technologies could help patients living with a rare disease.

Emily (Kunka) Lewis, MS, CCRP, CHES is an opinion leader in the digital transformation of healthcare. She is driven by a desire for the democratization of modern medicine. Emily received a Master of Science in Clinical Research from Northwestern University. She is currently working as a Digital Transformation Project Lead at UCB.

Note: The statements and opinions contained in this podcast are solely those of the speaker.

Podcast Interview

To access further The Waiting Room Podcast episodes and/or subscribe to the podcast channel, visit The Waiting Room Podcast landing page.

Transcript

Hello Emily, and welcome to The Waiting Room podcast.

Thanks for having me.

Now, first of all, please tell us a bit about your work and your connection with rare diseases.

Sure. So, it starts quite a while ago, and like most cases, it’s a pretty personal case, which I’m willing to share. I lost my uncle from ALS, Lou Gehrig’s disease, back in 2009. And from then I really decided to turn my professional life into kind of a dedication to his life, which meant that I started working at Northwestern Medicine in downtown Chicago in a multidisciplinary neuromuscular clinic. There I became really heavily involved in a lot of advocacy organizations like The Les Turner ALS Foundation.

I remember I worked in the lab as well as in the clinic, so I carried around a lot of donated brains and spinal cords in buckets to the lab and truly immersed myself in what a multidisciplinary care clinic is. And if you aren’t familiar with what it is, it’s essentially a clinic where patients are there most of the day in really complex disease states. And we bring in other specialties beyond clinical care like social work, dietetics, research, and then also vendors, like wheelchair vendors or respiratory clearance vest vendors. And we take a really patient-centric approach to just getting everything to them that they need within a full day so that they don’t have to keep going into large, dense areas like downtown Chicago. And we really respect the patient’s time. So, it’s a long day for them, but they get to see so many different people all in kind of a one-stop shop.

And then after working at Northwestern, unfortunately, my father was diagnosed with frontotemporal dementia, which is a related genetic rare neuromuscular and degenerative disease. And so I really got interested in genetic mutations and genetic mutations causing disease and the relatedness of genetic mutations across several disease states. And so from there, I also had a lot of professional experiences. I was already working in research and worked in Cushing syndrome, Tay-Sachs disease and Sandhoff disease. I worked in decentralized trials, which meant that we were bringing clinical trials to the patient’s home and those were ripe breeding, oh, not breeding grounds, but ripe environments for us to really test out these models of interesting research delivery within rare disease.

So, we got lots of really complex trials of small numbers of patients in which we were able to try going to be patients homes and doing these research procedures within that home environment. That meant doing physical exams on the patient’s bed and all sorts of things. But just being in this digital space was really kind of an important foray into rare disease because that’s where people were willing to really try things. This is about six years back.

And then most recently at UCB, my current employer, we are working within hidradenitis suppurativa, a rare condition within immunology for skin, and then also myasthenia gravis. It’s a condition within neurology that we, it’s actually called a snowflake disease that presents many different ways across patients. And so those are big focus areas for us. We’re really pivoting into rare disease as a company. And so as a part of, some of those rare disease efforts have been building software as a medical device for these disease states, for specifically myasthenia gravis. I’ve also worked with academia. We’re partnered with Stanford University and working with them to see how we can utilize generative AI to improve the clinical care delivery of rare patients. And with that partnership, I’ve also produced and co-chaired some rare disease symposia in which we brought in caregivers and thought leaders to talk about how the collective we can tackle some of these thorny challenges for rare disease patients.

Excellent. Now, to get into more detail, since you mentioned AI already, and that’s the topic of today’s session, is how can artificial intelligence or AI for short and digital technologies help improve the diagnosis of rare diseases?

Yeah. So, we’re in an interesting age, right, where we bring around digital devices with us wherever we are, whether it’s a smartphone or a smartwatch. There’s certainly this continuous high-volume data that we’re now generating just in our personal lives, right? And so this is now called digital exhaust. And so that’s kind of the term that’s been coined to really talk about this massive amount of data that every day we are generating in essentially everything we do, whether it be on computers or, like I said, in our personal lives, if we’re on social media or whatever. There’s just digital data being generated essentially in most of our daily lives. We’re lucky in that with rare diseases, although they’re complex and they are variable, as I mentioned with the myasthenia gravis, each myasthenia gravis patient essentially can present differently.

We can now use AI to integrate this clinical, biological, social data and really connect the dots for us to recognize patterns in large data sets and also combine diverse data sources. So, it could be the electronic health record, genetic testing, imaging, patient-reported outcomes. Lots of different data sources can really come together for the AI to then identify correlations in that data that are likely missed by clinicians. And then this is especially true in imaging, but it’s also possible in other data modalities. We also know that with natural language processing, which is a specific type of AI, we can also process and analyze unstructured clinical notes and medical literature to really extract relevant information that might be buried in that text. So again, it’s sort of, it can be a superhuman power almost, where it really augments the clinicians to really see things within information they may not have put to put together on the front end. It really happens on the back end.

The other thing that’s interesting about this era of AI is that it’s really optimized for human computer interfaces. The AI can then surface insights and suggestions in clinicians’ workflow, for example, at the right time. In terms of diagnostics, which was sort of the nature of the question, you know, clinicians can really see these insights in real time and use them as diagnostic tools, especially in differential diagnosis of a patient, to have these quicker, more accurate, either medical, the latest medical research or similar case histories pop up and really prompt them to think through in a really organized fashion how they might approach, like I said, kind of the differential diagnosis process. So, yeah, I really think of AI as a copilot. I think Microsoft has a pretty good naming of their better latest product because I really do think that we should be thinking of AI as a copilot in medicine and otherwise.

Okay. So, now that we’ve spoken about the diagnostic part, how about the management of rare diseases?

Yeah, yeah. I think there’s definitely a role for AI in kind of personalized health monitoring for patients. As I mentioned, we kind of worked on software as a medical device. I see a place for apps to really help patients track the progression of their disease by tracking their symptoms. So, they can, in their own time, record, whether it be a selfie of their face or them doing certain movements with video recording happening. And AI can be in the background really looking at how the disease is progressing, if it is progressing, and then really help the patient communicate with their doctor about this progression of symptoms to really talk to their physician or their clinician about how they can optimize treatment. It could be that they adjust medication based upon the progression of disease, and they can really track the effectiveness of that treatment when they have made the adjustments and then go back and see how their symptoms are getting better or worsening or what happens afterwards.

So, it really allows for better, more efficient communication between patients and health care teams to really fine-tune the treatment according to how symptoms are progressing. I would also say that algorithms can analyze symptom data for really looking at proactive management. Rather than reactive management, as I just mentioned, we should also be thinking about kind of identifying this worsening of symptoms earlier so that we can essentially try and prevent a worsening of disease. In the example of myasthenia gravis I gave earlier, what we’re trying to do is predict flares of disease or worsening of symptoms before they even start by looking at things like weakness or level of activity. So, we want to get to a place where we’re more proactive and use is data-driven, more responsive integration of people’s exams, labs, scans and daily symptoms to really think through how we can be more proactive about treatment, or even just rest. You know, it doesn’t even mean treatment. It can be an avoiding of behaviors or lifestyle changes.

Okay, that sounds like a great approach. Now, do you see any challenges and possibly barriers for patients with rare diseases when trying to use and access these digital solutions?

Yeah, absolutely. I mean, as with anything, there are challenges, but specifically within rare disease because of the fact that they simply get less attention and are less prevalent than more common diseases. There’s just less availability of solutions for rare disease patients. And a lot of digital health solutions are just not designed for rare disease. So, that’s one disadvantage. But also we know that a lot of tools, even when they are created and put on the market, aren’t necessarily clinically validated and reliable. Clinicians often look for this to make sure that they’re giving, recommending solid treatment options for their patients.

But, you know, health care providers are just hesitant to recommend these kinds of solutions. And insurers, at least in the US, are less likely to cover them if they don’t have the clinical studies and the clinical evidence behind them. So, availability, clinical validation and reliability. But also, you know, obviously cost is a factor. We know that digital literacy is sometimes a factor, especially in older populations. It’s getting better, but, you know, people may not be as familiar with digital solutions. And then also within rare disease, you know, with really complex patients, there may be actual physical or cognitive limitations with some patients.

I was wondering about one aspect, which would be the ethical part. For example, ethical considerations which need to be taken into account, like data privacy or something like that. Do you see any problems there?

I don’t see them as problems. I see them as watch-outs. Certainly one thing throughout any research and especially digital and AI is informed consent. Patients and caregivers should always be providing continuous informed consent for understanding how the data will be used, understanding who will have access to it, and what the purpose of the access is. If it’s for research or treatment optimization. You know, I think about how I share my location with my family through Google, and Google continuously sends me an email and says: ‘You’re still sharing your location with your family members. Is this okay?’ And I think of that as sort of what digital should be for or what this process should be like for all for patients and caregivers. It should be continuous, even if it’s just reminding that you are participating in a trial or that your data is being used. I think that’s an excellent way to keep this kind of thing front of mind. So, informed consent is huge.

I also think, as you mentioned, data privacy and security are important things to consider. This data is inherently sensitive and it can be stigmatizing. Breaches of this data can re-identify patients in some cases, and there are some serious discriminatory consequences if this data is leaked, such as in the US insurance eligibility. There’s things like employment prospects could be jeopardized or even educational or social interactions. So, we need to make sure that we’re keeping this data safe. And also bias and fairness. We know that rare disease patients are diverse in their disease severity and their presentation of symptoms. We know that we need to be gaining large amounts of this data to make sure it’s diverse enough to train the data appropriately so that we don’t inadvertently perpetuate biases and make sure that it’s as representative of the population as possible.

And then the last thing I would mention is really the idea of autonomy. We should still be thinking about how digital technology should really complement and not replace the patient-provider relationship. And so this is especially critical in rare disease patients where maybe there’s a caregiver involved who’s making decisions on behalf of the patient. But in either way, if it’s the patient making decisions for themselves or the caregiver making decisions, it should be a shared decision-making process with the clinician, and they should be actively involved in their care and treatment decisions. And so AI is certainly a way to augment both the patient and physician in being a part of the decision-making process and being apprised of everything happening along the way in their care delivery.

Absolutely, yeah. Now let’s have a look into the future and sort of look into the crystal ball. Which innovations are you expecting in the field of AI and digital health devices, which would help patients to possibly have a better quality of life?

Yeah, I sort of think about this in two ways. From a care perspective, I’m really hopeful for a truly personalized treatment plan. Ones that really consider a patient’s unique genetic makeup, you know, based upon markers and mutations in genetics as well as their phenotype. Considering their disease presentation. I really see a confluence of lots of different data sources that really bring both historical data together with this current patient data for this individual to really guide the selection of more effective treatment strategies and find treatments that have the least amount of side effects. So, really optimizing outcomes for patients based upon more informed treatment decisions.

And then in the research space, I really think about AI augmenting clinical trial matching. Finding patients for clinical trials that help us have better chances of finding effective therapies of disease. And then also in more of the rudimentary R&D process, drug repurposing. So, finding AI, using AI to identify new use cases of existing drugs, but then also discovering entirely new drugs so that we, we know there’s a lot of stimulation happening with AI. We know that people can use what’s called GPUs, graphical processing units, in the cloud, and they have unique instances of this that they can then test using generative AI, test virtual modifications of active molecules, using machine learning-based methods to really find, really narrow down the number of compounds that they are interested in and have this done all, you know, in the cloud with just virtual simulations of active molecules. We can actually find better targets and potentially new molecules that we never thought of using in generative AI.

You previously mentioned that we basically know that rare diseases are really diverse and there’s a lot of them, about 7,000. I was thinking that there’s a high degree of individual patient need because each and every case and each and every condition can differ vastly from any other patient. What is your take on this?

I think you’re absolutely right. And I was actually at the symposium that we put on in September. This came up as a big theme of the conference. And so there was one quote that I want to attribute to Onno Faber is his name, he’s a key opinion leader in this space, but it really stuck out with me. And it was something to the effect of, you know, thinking of many individual patients and the diseases within these individual patients as one big problem. Although there are many of them, if you really consolidate to think of rare disease as one big problem, then it’s in some ways easier to tackle. But I really believe that we should take a really global perspective on rare disease. I think that we are stronger in numbers and so international collaborations are really important. Number one, because you can pool resources, but also you can pool patient populations and expertise as well.

And luckily, we’re seeing a lot of movement within like the government and regulatory sectors as well. We’re seeing many governments provide incentives for developing orphan drugs for rare diseases by doing things like incentivizing market exclusivity, tax credits and assistance with clinical research. So, this makes going into rare disease research and development just more economically viable for those companies who are interested. And then we’re also seeing, interestingly, some alternative funding models. And those are things like venture philanthropy and public private partnerships. We’re seeing more capital go into this space in areas that might not have been traditionally profitable.

It’s interesting to see how this space is evolving. There are certainly changes happening with regulatory agencies. They’re becoming more flexible in how they want to interact with companies who are doing rare disease research, and they really are promoting novel methodologies for research and development of rare disease drugs, using things like adaptive clinical trials, as well as the consideration of real-world evidence, and then also being in favor of expedited review processes to get these therapies to market quicker.

Thank you. Now, more of a hypothetical question: If you had all the money and means in the world, which ideas and plans would you realize first regarding AI and digital health and rare diseases?

Yeah, I mean, when I think of this question, of course, selfishly, coming from the US, but my greatest wish, for this country at least, is for a universal payer system in the US, which has been the bane of our existence for a long time. If we can just have some seamless interoperability and data integration, I’d be very happy. But stepping out of that selfish perspective, I would certainly say a global framework for the ethical development and use of AI in health care. We’re seeing lots of regulations pop up kind of in pockets. You know, there are certainly some in Europe right now. And there’s more guidelines in the US and other parts of the world. So, a harmonization of those frameworks would be nice.

And also just addressing issues like data privacy and confidentiality, security bias. Those are all things that are very top of mind for me. And the idea of transparency is super tricky in the field of health care AI because a lot of times more transparency means less accuracy in models. And so how do you balance those things and really keep the manufacturers of these models accountable? And the US has some frameworks around this to make sure that patients are accountable, and we’re certainly starting to see the same in other places, too.

Equity is also on my mind as well. We need to make sure that we have a responsibility really to get AI to every corner of the earth because I really believe that it can democratize medicine, much like telemedicine did during COVID. We have the capability of really transforming sick care into true health care by reaching some low-income rural areas and underserved regions. So, in the end, we can do our best to conduct the best R&D, discover the best products, and build the best products. But unless we get those to the right people at the right time in a personalized way, as we saw with a lot of digital therapeutics, if that last-mile delivery isn’t there, you know, it’s all for naught if we’re not getting it to the right people.

Definitely, definitely. Thank you. Now, as we are approaching the end of the interview, here’s my last question to you: What are your own plans for this year’s Rare Disease Day, which is on February 29 this year as we’re having a leap year, as well as for rare disease month?

Yeah, I mean, I’m pretty active on social media, so that’s a lot of where I do my activism. But I believe we can all contribute to kind of bringing awareness to the special month in special ways. It could be in your own way. You could go into the community and get involved in a local chapter of an advocacy group or a non-profit. You could raise awareness on your social media, attend events like a charity run or auction, but I think it’s all of our responsibility to really figure out how to empathize with these patients. If we really understand them, we can truly be in solidarity with them and understand their needs to best meet them, right?

I think all of us should be really finding ways to connect with somebody who maybe has a rare disease that you know of or share their story and make sure that you’re really understanding where they’re coming from. I know there’s one really great podcast that I can recommend. It’s, well, there’s several, but there’s one, Once Upon a Genie, or Once Upon a Gene, that’s really good, that actually interviews rare disease patients and allows them to tell their story. So, that’s one great place that anyone can start.

Excellent. Thank you, Emily. Thanks a lot for your time and for your insights, and all the best. Thank you. Take care.

Thank you. Likewise.