AI and Digital Health Accelerating Rare Disease Clinical Research Powered by Patients

Harsha Rajasimha, MS, PhD, Founder, Jeeva Informatics Solutions, Inc.; Founder and Chairman, IndoUSrare; Co-Director, Rare Diseases Systems Biology Initiative, George Mason University

After losing a child to a rare congenital disease, Dr. Rajasimha became determined to apply his clinical genomics data research experience to develop solutions to help accelerate clinical research leading to faster cures for rare disease. Dr. Rajasimha will discuss his efforts in fostering collaborative bridges between patient advocacy groups and researchers in the USA and their counterparts in India to help accelerate clinical research, trials, and therapy access across borders. The talk will include recent global initiatives to accelerate screening, diagnosis, and treatments of rare and undiagnosed diseases. He will also share work on the development of an AI-driven digital health platform to improve clinical trial operational efficiencies while significantly reducing costs and travel burden on patients.

Transcript:

John:

Welcome, and thank you for joining today’s S3 webinar. Our topic is AI and Digital Health: Accelerating Rare Disease Clinical Research Powered by Patients, presented by Harsha Rajasimha. Harsha is the founder of Jeeva Informatics Solutions, Chairman of IndoUSrare, and co-director of the Rare Diseases Systems Biology Initiative at George Mason University.

John:

We will be taking questions after the presentation, so please submit them using the Q&A button. There will be polling questions throughout the webinar. We encourage you to take part in those.

John:

I will now turn it over to Harsha.

Harsha Rajasimha:

Thank you, John and thank you Sanguine for the opportunity to present my perspectives today. My talk will focus on the question around whether it’s AI and digital health accelerating rare disease clinical research powered by patients, or patient focused drug development powered by AI and digital health technologies.

Harsha Rajasimha:

I’ll take a moment to introduce my background and where I am in my journey. I’m a computational biologist by training and I was going about my day-to-day life having transformed from a software engineer to a data scientist working at the National Institutes of Health since 2008.

Harsha Rajasimha:

By 2012 I already had over 10 publications and was looking at data analysis of retinal genomic data at the National Eye Institute, and a life-changing event happened then, that of a child born with the rare congenital disease called Edwards syndrome. The baby passed away at day four. That put me on a… I decided to apply my years of post-doctoral training and set on a mission to accelerate clinical development timelines.

Harsha Rajasimha:

Since then, what I have been learning is the journey that the patients go through, a majority of patients still remain undiagnosed, across the world and more so in the rest of the world than in the United States or in EU. Once they do get a diagnosis of a disease, and they are very lucky if there an FDA-approved treatment because there is some for, maybe about 5% to 7% of the rare diseases at the moment.

Harsha Rajasimha:

Once they have a diagnosis and do not have an FDA-approved treatment, then they need to look at what therapies are in development, in clinical trials. And currently, 1,971 clinical trials are associated with the rare disease that they could look into in the clinicaltrials.gov portal.

Harsha Rajasimha:

And some patients are diagnosed with a disease without any ongoing clinical trials. What’s the option there? Is it right to try based on ad hoc or some sort of a hope that a treatment for some other disease or a herbal product might come to the rescue?

Harsha Rajasimha:

So that’s the very heavy journey that patients who have or have a child living with a rare disease go through. And there are a number of initiatives in every part of this journey that provide a lot of hope and opportunity to advance these activities.

Harsha Rajasimha:

When we talk about, “How can we accelerate diagnosis and development?,” it’s not just automation, but a lot of other things that I will draw your attention to for this presentation.

Harsha Rajasimha:

So we look at in at least four stages, of diagnosis, discovery, research, where we are discovering candidates, candidate drugs or targets, and development of the drug itself, especially clinical development, and monitoring the patients and the disease after a therapy’s approved. And more importantly, a feedback loop that needs to go back into the early, undiagnosed situation and research it as well.

Harsha Rajasimha:

So when I set out on the mission, I looked at a map similar to this, which looked very similar, but a fewer number of trials back in 2012. What was very striking is a majority of clinical trials are concentrated in the western world, particularly in the United States and EU, probably 80% or so. And so that represents under 10% of the population globally, and not as diversely representative of the world’s population.

Harsha Rajasimha:

When I looked back at India, where I grew up and emigrated from and I’m currently based here in Virginia for last 20 years, when I looked back at India, there was probably about 2% of clinical trials that had any footprint there, and the awareness about rare diseases was very minimal or none, as there was not a single research article that spoke about the status of rare diseases in Indian subcontinent.

Harsha Rajasimha:

When we look at the situation today, it hasn’t changed a whole lot, but it has come a long way in terms of awareness in other parts of the world as well. If you look at where we are here in the US, we know the cases of Mila Makovec and Jaci. Mila with the Batten disease and Jaci with the ALS, both having a single patient with the ASO therapy, or antisense oligonucleotide therapy that was custom-made for one person, one patient.

Harsha Rajasimha:

And so that’s where the western world has progressed. So, that innovation exists. It happened in a very rapid pace; however, it’s not widespread and it’s not evenly distributed even here in the US. On the eastern side of the hemispheres, you’ll see that there are still millions of people remain undiagnosed and still do not have even a public health policy that covers the needs of patients with rare diseases.

Harsha Rajasimha:

So my journey as a social entrepreneur, and to distinguish that from serial entrepreneurship, is a social entrepreneur is motivated to solve social problems, and in my case it’s rare disease and this continuum, and speeding up the development of diagnostics and treatments. It does require both public and private sectors to come together for various aspects of this continuum, from raising awareness, educating, both on the patient side, the physician side, as well as the researchers, industry, policy-makers, all state quarters need better awareness and education about over 7,000 rare diseases that exist.

Harsha Rajasimha:

Then generating and collecting high-quality data. That needs to happen within the scientific communities. My own journey as I set out, brought together a group of key opinion leaders and experts in the field, while I was just getting started and led the review article summarizing the status of rare diseases in India, in particular, and tying it with global initiatives.

Harsha Rajasimha:

Of course, right here in the United States we had NORD, Global Genes, and EveryLife Foundation doing tremendous work across the awareness, education, policy, advocacy in particular, but that needs to reach beyond, as rare disease is a global issue, and given the sparse numbers, it was clear that engaging a densely populated eastern part of the world would be critical, particularly in the coming decade as we are seeing that we do not have enough patients, for example, with Batten disease, right here in the US alone, to be able to conduct all the clinical research and clinical trials with enough statistical power, et cetera. So we needed to vote for humanitarian grounds as well as scientific advancement and to accelerate drug diagnostics and treatments. That was a necessary aspect.

Harsha Rajasimha:

So in that regard, I have founded a couple non-profit organizations and currently focused on Indo-US Organization for Rare Diseases, a 501-C3 non-profit here, as well as my experience at the NIH and FDA gave me insight into the research and scientific process, as well as the policy development process that happens in the FDA, although I was not medically involved in those aspects, I was more on the technology and scientific data analysis space.

Harsha Rajasimha:

So that got me into the private side, where I realized commercially reliable and scalable products, which is what really the patients eventually need access to. I spent the last about 7, 8 years in the private sector as well, developing technology products, launching diagnostic tests, in teams at Strand, or developing genome sequencing data analysis cloud-based software analysis platforms at Genome International Corporation. And currently, with my own technology startup focused on overcoming travel burden in clinical trials, where most of the clinical trial participants have to visit in person to participate in a trial at a site.

Harsha Rajasimha:

I’ve also been a technology consultant with the MTD data services, consulting with the big biopharma companies in the R&D side, looking at how AI and big data technologies can help speed up or make R&D more efficient in the big pharma.

Harsha Rajasimha:

Right now however, I’m focused on Jeeva, so as you can see, this perspective is very important for the rest of the talk, as this ties in with various existing amazing work that has been done by various organizations, like the Global Commission to End Diagnostic Odyssey for Children with a Rare Disease, or International Rare Diseases Research Consortium, the Undiagnosed Diseases Network International, and Rare Disease International, that’s based on the EU side, but globally focused.

Harsha Rajasimha:

So being privileged to be part of this community, and not gathered much of the advocacy side of the rare disease development as much as science and technology aspects.

Harsha Rajasimha:

So with that perspective and background, I’ll jump right into today’s topic of innovation in AI and digital health, and its relevance to rare disease. I love this quote from Sachin, and maybe others have said the same, but we all kind of live in an innovation bubble while there’s a booming innovation industry, a new startup being created every day, a new app is being launched every minute, the actual experience of delivering or receiving care is changing scarcely. I hope all of you can relate too, as that’s not changing very much.

Harsha Rajasimha:

What I have done as an entrepreneur is went out and conducted customer discovery interviews. Yes, there are various market reports, but listening firsthand from over 300 customer stakeholders involved in clinical trials, what we see is that the orphan drug development takes much longer compared to non-orphan drug development, and some of the key problems we heard, according Pfizer, and this was a very general comment theme across the sponsors, who said, “Every trial feels like the first ever trial ever undertaken by mankind.”

Harsha Rajasimha:

What that means is we are not learning from past clinical trials to make the next trial more efficient. And that to me is a problem that is perfectly suited for AI, only if we have good quality data.

Harsha Rajasimha:

The second major problem I heard is, “Recruitment is always the biggest problem in clinical trials,” not just for rare disease, but for any disease.

Harsha Rajasimha:

And lastly, the patient advocacy groups opine that, “We have all the tools, yet the burden remains so high for us to participate in clinical trials.”

Harsha Rajasimha:

So what became evident from this discovery exercise is that travel is a addressable major problem in clinical research, which when addressed, could solve about a quarter of the patients who are not willing to enroll in clinical trials after reviewing the informed consent form, citing travel as the major problem, or location of the site, or number of visits that’s associated with participating with this trial.

Harsha Rajasimha:

So we set out on a journey to address this travel burden problem. If you look at the way clinical researcher is conducted historically over the last 70 years or so, almost all of the informed consents are conducted by the principal investigator in that trial meeting with the patient or the family and providing information and answering questions about the risk and benefit of the trial and getting the consent signed in a brick-and-mortar hospital or a clinic site where the physician and patient interact in person.

Harsha Rajasimha:

The paper questionnaire clinical assessments are conducted at the site, follow-up visits are done at the site, patients have to go back, of course severe adverse event management, collection of any samples and vital signs, as well as delivering of the investigational medicinal product, all are done in person.

Harsha Rajasimha:

Recently, however, there has been a significant move towards doing some of these where, in my opinion, there is significant opportunity to do more and more of these informed consents electronically, although it’s currently at about 2% of all informed consents are done electronically. I think the opportunity to do more and more, there may still be reasons to do in person, but probably more than half of the consents can be done electronically.

Harsha Rajasimha:

Patient-reported outcomes can be collected by apps instead of having paper-based questionnaires. That’s also being done in a piecemeal or, one of ePRO tool that’s being deployed already in certain clinical trials, but still much smaller percentages.

Harsha Rajasimha:

Videoconferencing as a way to assess clinical outcomes, as well as teleconference, chat, and SMS-based follow-up visits as well as adverse event monitoring and collection.

Harsha Rajasimha:

There are also opportunities, of course, as our host today, Sanguine does home-based sample collection for patients with rare disease and other diseases like Lupus. That’s been happening as well again, not at scale. So I think if we look at this holistically in individual pieces, some of these are already being done at some minimal adoption levels, but I think if we bring these pieces together in a way that can reduce the burden on patients, replacing the travel visits without having to do travel. That’s when the value of these technologies will start making sense.

Harsha Rajasimha:

So that’s very consistent with what industry leaders are saying, such as Vas from Novartis saying, “The first thing we have learned is the importance of having outstanding data,” right? And this is heavily underestimated, how little clean data there is out there, particularly in the clinical trial space. The entire clinicaltrials.gov portal and database is downloadable, publicly accessible, but we do not have the level of subject-level data that’s necessary to train our AI and machine learning models.

Harsha Rajasimha:

We need to start with having access to, and generating, and maintaining and sharing good quality data in clinical trials for really taking advantage of AI and ML.

Harsha Rajasimha:

If you look at the industry, though, there are a large number of… It’s quite clouded, the number of companies that are in the space for AI, and this is a wonderful slide that I borrowed from Andrea Coravos from 2018, and there’s been several more new entrants into this space as well, and some of these logos that you see here are probably potentially in multiple boxes as well, and the technology startup that we are focused on, Jeeva, is in bottom of this slide here, and of the decentralized trials, which kind of cuts across the patient recruitment retention compliance data aggregation and analysis, sort of in a new paradigm of decentralized clinical trials options.

Harsha Rajasimha:

So when you look at the real, practical application of AI, it is still in the early stages in proof of concept and data collection, whereas the industry already has the tools, just that we don’t have good quality data.

Harsha Rajasimha:

What we are doing to address one of our customer problems is the continuous learning and continuous improvement to optimize clinical trials. We have attrition, so if we need 100 patients in a clinical study, you need to reach out to maybe 1,000 subjects, and this is typical, not specific to rare disease, but out of those, how many are screened are determined to be eligible, and how many of them actually review the informed consent form and how many of them actually consent to being part of the clinical study, and then they get either randomized, or depending on the trial protocol design, then the administration of the drug itself, collecting data via surveys and visits, either in person or telehealth visits, and then monitoring.

Harsha Rajasimha:

Imagine conducting this type of study, which is not 9 months, or 3 years, but 5 years, or 10 years, or 15 years. Every step of this workflow, there is attrition happening. Patients are dropping out, they are dropping out very early in this process of recruitment, or after they are, about 30% of them drop out during the course of a trial, even with an average trial duration of 2 years or so. But in the longer term trials, that problem can be significant.

Harsha Rajasimha:

And particularly in rare disease with very small populations, the problem of patients dropping out can mean significant loss. 1 out of 10 patients dropping means 10% of the data is missing now.

Harsha Rajasimha:

So what we believe, is that having AI to manage this process, which can monitor what’s happening through all this life cycle and continuously improving our process can significantly reduce the attrition rate, as well as improve the patient recruitment speed indirectly as well.

Harsha Rajasimha:

So what we are doing at Jeeva, is a decentralized clinical trials technology platform that brings together the end-to-end of the clinical trial process from prescreening the participants, identifying those who are eligible, performing electronic consent, patient reported outcomes, data collection via mobile apps, and electronic wizards, outcome assessments, all of that via mobile app and a web-based app that the clinical trial research team and the patient can engage through. They get reminders, alerts, notifications, and the monitoring is done instead of in the traditional process where a sponsor engaging with a contract research organization or otherwise has recruited a number of brick-and-mortar sites and those sites are recruiting patients in person, that paradigm is now entirely streamlined using software as a service, and using technology on a cloud, with centralized data capture and monitoring technology, which is also driven by AI to some degree where it makes sense.

Harsha Rajasimha:

AI in this context could mean anything from natural language processing, voice recognition, to learning from past clinical trials data sets, that we are conducting ourselves to help optimize the trial in real time as the trial is ongoing to ensure the timelines are accelerated, the attrition of patients during the trial is reduced, and the overall user experience of being part of a clinical study is enhanced significantly.

Harsha Rajasimha:

Of course, anything in clinical trials has significant regulatory compliance requirements. A couple of them include HIPAA and FDA 21 CFR Part 11, but of course, there is also GDPR and a few other state-level requirements and other regulatory requirements as well.

Harsha Rajasimha:

So, what AI is not going to replace is the drug developers or the clinical investigators; however, drug developers and investigators who use AI will soon replace drug developers who don’t. So AI is a means to an end, and not the end in itself, and there are ways in which AI can be integrated practically today, in clinical trials; however, it’s still in proof of concept and pilot stages and the models are being trained and it’s going to be a while before it’s all fully implemented across the clinical trial process.

Harsha Rajasimha:

In conclusion, and I will open up for discussion and Q&A, what we covered is that AI is a means to an end and we have come full circle with the FDA, now recently has been driving the patient-focused development paradigm, in many attempts where technology and scientific advances have been trying to push the drug discovery and development process without fully taking the patient needs into account. So we are now starting with the patient first approach, starting with what really matters to patients. What symptoms and clinical endpoints that really matter. And then ask the question, “Where can AI digital health really help?”.

Harsha Rajasimha:

And to end data silos, we need to generate complete and high-quality data. So what I mean by “generate complete data” is that having been part of interdisciplinary clinical research projects, more often, the data generated is very focused, for good reasons, given the budgets and grants and other restrictions on what type of data is generated.

Harsha Rajasimha:

If I’m a genomics data scientist or a clinical genomics scientist, I won’t only look at the genomic data, and over a period of time now, there has been recognition that we also need the phenotype data more, along with the genome and that has been improved quite a bit, but even the extent of phenotypic data is very minimal still. Environmental factors, social determinants of health, weather and other factors, all play a role, and capturing those types of, and patient reported outcomes.

Harsha Rajasimha:

So there is good recognition, yet there is still lack of how data is generated with short sight, or a very focused approach to generating a specific type of data. And often, drug development is much slower compared to non-orphan drugs.

Harsha Rajasimha:

Global silos in geographies have to be bridged to accelerate clinical development.

Harsha Rajasimha:

Geographic and mobility challenges can be addressed using mobile digital health technologies, where once we have the approach onto which investigators, patients, regulators, and sponsors all are aligned and creating clinical trial designs that can be conducted without requiring in-person interactions between physicians and patients. That’s the situation where we want to be, and I’m hoping that with the current temporary relaxations that the FDA has provided to the telehealth and state barriers within the US will continue on a more long-term and permanent basis, like the FDA commissioner and the CMS head have vowed to see that they continue for the greater good.

Harsha Rajasimha:

So with that, I will conclude on my journey here, and happy to take any questions.

Harsha Rajasimha:

“Do you think with COVID AI and digital health will be used more? What has the stall on our trials shown us? Do we need new tools?”

Harsha Rajasimha:

So, great question. With COVID-19 as what we saw is that a lot of the clinical trials lost their trial integrity, in that there’s been missing data or incomplete data as a result of COVID-19, especially when the lock downs began. A lot of the in-person visits had to be brought to a halt abruptly. What that meant is a number of the clinical trial protocols were done in a very rigid manner, in terms of… There is some flexibility, but at the same time, certain data points, sample, have to be collected at certain intervals, and if that’s not happening…

Harsha Rajasimha:

Going back and starting the trial all over again means lost time, lost money, and the patients have to probably go through some sort of a wash-out period that they typically do in certain types of trials, et cetera. So it becomes a very challenging situation where a lot of loss and irreparable damage occurred and trials lost their integrity.

Harsha Rajasimha:

To be able to maintain trial integrity in similar pandemic situations, yes, AI and digital health tools can certainly be of significant value, if the trials are designed with that type of a option in mind, where patients could continue to participate from the convenience of their home, either through home-based sample collection or through remote participation.

Harsha Rajasimha:

And we do need new tools because the kind of tools necessary to support these types of clinical trials have not been well tested, validated at scale, still, although there are a number of tools already out there. We do need good, new tools, or good validation of these existing tools, but also the kind of tools that are scalable and flexible.

Harsha Rajasimha:

You can’t have a tool that works today for one disease that cannot at the same time work for the next clinical trial, even for the same disease, let alone for other diseases. These tools require significant flexibility and customizations. So, absolutely we need new tools.

Harsha Rajasimha:

Let’s look at the next question.

Harsha Rajasimha:

“Do you see a shift where we can or will see more treatments for rare diseases?”

Harsha Rajasimha:

Absolutely yes. I’m very optimistic given all the recent good news and progress, at the same time cautious about it. On the one hand, we have NF1 or n-Lorem foundations that are now starting to look at these oligonucleotide-based medicines or gene therapies that are being quickly set up with a very innovative regulatory process, and that FDA has shown the ability to support a trial for even one patient.

Harsha Rajasimha:

At the same time, there are still many rare diseases, more than 90% without any treatment, but I see a lot of good science, that more treatments for rare disease will be available with all the collective efforts of these organizations globally.

Harsha Rajasimha:

Next question, “Do you think that clinical trials will always need to have a patient go to a clinical site, or will telemedicine become more popular?”.

Harsha Rajasimha:

I am a firm and big believer that the telemedicine or videoconferencing-based approach to conducting clinical visits for patients wherever it is possible. It’s not always possible to do that. There are a lot of good reasons where patients would be wanting to go and see a doctor in person, or the doctor may want to see the patient in person to make certain clinical assessments; however, there is an opportunity for anywhere from 50% to 80% of the time, depending on the trial and the disease, where these visits can be avoided, and telemedicine or videoconferencing could be used in lieu of in-person visit. That’s my belief.I’m very optimistic and bullish on that. Next.

Harsha Rajasimha:

“Can AI help with recruitment?”

Harsha Rajasimha:

Absolutely, and there are a number of products and solutions that are out there that are looking at the patients’ electronic health records. So the challenge is recording patients is not simply a matter of asking three questions and getting the patient onboarded. Patient recruitment requires significant information about the patient’s health, in electronic medical records and health records. And there is no single standard EMR or EHR format across even United States, let alone across different countries.

Harsha Rajasimha:

What that means is we need a way to organize the data and be able to access the electronic medical and health records, including unstructured physician notes, and other types of data to be able to apply and ask the questions, “Which of the patients meet the inclusion/exclusion criteria at scale?,” and that has been a huge challenge and there are a number of companies that are trying to address that problem right now.

Harsha Rajasimha:

So the answer is “yes.” But to some degree, they are applying AI already in using NLP to extract patient health information from these medical records.

Harsha Rajasimha:

Next question, “What ways can AI assist in alleviating the delays in clinical development?”.

Harsha Rajasimha:

I referred to some of this in my presentation, but there are areas where patient recruitment is the biggest problem and a big area where AI is being applied, to some degree already and a huge opportunity exists, and I also showed how AI could help in patient retention in a study, or patient compliance to clinical protocol. Some sort of robotic automation, which is not typically considered AI, but could be, or… It’s not always as fancy as machine-learning mortal or something very advanced, but even the ability to recognize the voice of the patient and be able to determine that it’s really the patient’s voice and then what is the patient asking, and being able to provide a response back with Alexa, for example. That’s also part of the AI as well.

Harsha Rajasimha:

When you look at it in that perspective, there are many ways in which AI is assisting in clinical trials, but it’s not equally distributed. It’s not widespread. So to make it really widespread, it’s still a work in progress at this stage, but there are a small number of areas in recruitment, retention, compliance as well as in routine patient monitoring, where AI is being involved.

Harsha Rajasimha:

“Will telehealth help with patient recruitment? What are the best steps to incorporate telehealth in an upcoming trial, where when COVID lets us get back to work. So telehealth has already been, and is routinely used in patient recruitment in the following sense, right? Where you have the phone number of a patient and subject and a clinical coordinator makes a phone call, and asks maybe three or four questions to assess preliminary eligibility criteria for a clinical trial.

Harsha Rajasimha:

“Are you between the age of 20 and 30?”

Harsha Rajasimha:

“Are you a smoker or a non-smoker?,” et cetera.

Harsha Rajasimha:

And then if that’s the criteria for a clinical trial, that’s when the coordinator asks the patient to come in to the clinic so we can do a full consent, full informed consent and get you into the trial.

Harsha Rajasimha:

So that type of preliminary screening has been done in a number of trials with telehealth, so in that sense it does help in patient recruitment, but there are a number of ways in which that can be incorporated, even automated teleconferencing systems, or even integrating a phone call within the app, like Uber does, for example. That level of integration and widespread deployment at scale hasn’t still happened, and that’s still an opportunity where that option can be driven post-COVID.

Harsha Rajasimha:

So I think I’ve covered most of the questions. There are a couple more questions, which seems like might be redundant, so I’ll give it a moment of pause to see if there are any further comments or questions.

Harsha Rajasimha:

Lisa, John?

Lisa:

No, that’s all the questions we got.

Harsha Rajasimha:

That’s all I had to share today. Thank you for joining. Look forward to hearing from any of you if you had any further questions, feel free to reach out to me at my email address, harsha@jeevatrials.com.

John:

Thank you, Harsha. And thanks to everyone for joining us for today’s S3 webinar. If you want more information about upcoming webinars visit researcher.sanguinebio.com, and if you have a need for patient samples, visit sanguinebio.com for more information.

John:

Thank you, and enjoy the rest of your day.