Innovation Unleashed: How AI Technology is Reshaping the Future of Clinical Trials

Key takeaways:

• Clinical trials are rapidly evolving with artificial intelligence, real-world evidence and decentralized trials – all playing important roles in the future of clinical research.
• These capabilities hold immense promise in supporting innovation in clinical trial recruitment and operations to ensure all populations have access to breakthrough medical technology faster than ever before.
• In an on-demand webinar, experts from PINC AI™ Applied Sciences are joined by healthcare and research professionals to discuss the implications of technology in clinical trials and the impact these solutions are having on accelerating real-world evidence and the development of novel therapies.

The world of U.S. clinical trials is advancing at lightning speed – from artificial intelligence (AI)-powered machine learning (ML) that processes large datasets and recognizes patterns, to natural language processing (NLP) that can read and interpret more than 2 million records per hour, to the potential of decentralized clinical trials (DCTs) and real-world evidence (RWE) in increasing access to more diverse patient populations, improving study design and enhancing patient outcomes.

These innovations and more were discussed during a recent webinar on the future of clinical trials hosted by PINC AI™ Applied Sciences (PAS), the research and analytics division of Premier. During the session, Myla Maloney, Chief Growth Officer at PAS, spoke to Dr. Christina Brennan, Senior Vice President of Clinical Research at the Feinstein Institute (part of Northwell Health); MarieElena Cordisco, Associate Vice President for Clinical Trials, Research and Innovation at Nuvance Health; and Dr. Yosef Khan, Executive Director Clinical Trials and Real-World Evidence at PAS.

What follows are excerpts of their conversation; the full webinar is available for viewing on demand.

Maloney: How have your clinical trial capabilities evolved over the last few years? Have you created anything new?

Cordisco: One of the things I’ve noticed is an increase in the complexity of clinical trials, which is being driven by difficult inclusion/exclusion criteria, and that makes it hard to recruit from your general population of patients and meet enrollment targets.

To overcome these challenges and identify the right patients for studies, we’re working with our IT department to develop unique patient identification solutions as well as working with third-party companies that give us access to real-world data sets to track patients across many sites and see how we can fit our studies into where those patients are located.

Dr. Khan: One of the biggest challenges, or changes I should say, has been the emphasis on identifying and including more diverse patient populations into studies and leveraging real-world evidence (RWE) from electronic health records (EHRs) or claims data to provide insights into patient outcomes and treatment patterns in real-world settings.

As a result of this increased focus, the PAS team has honed in on identifying, leveraging and understanding actionable real-world data (RWD), which enhances our research capabilities. We’re looking to leverage the best data source that’s out there – linking it, tokenizing it, building it out and leveraging it along with AI technologies.

Dr. Brennan: I would agree with the other two panelists and add to the comments on the ability to utilize the EHR; I don’t think we use it enough today. We also need to use AI capabilities more. We must move past using just ICD-10 codes and diagnosis codes. NLP can help us parse through the EHR’s unstructured data and get into progress notes and other areas to help identify patients.

Cordisco and Dr. Brennan further the discussion around building a strong partnership between IT and research departments for effective data utilization:

Maloney: How are DCTs different than traditional clinical trials? If we really believe DCTs move us in the right direction, what capabilities do we need to enable them and even accelerate them in the market?

Dr. Brennan: You’ll hear a lot of terminology around what a DCT is, but essentially, it’s a clinical trial that’s decentralized when one or more parts are conducted using technology. And we saw DCTs come into the light more during the COVID-19 pandemic when we had to do some things remotely, even something as simple as a phone call to monitor our patients.

When I think of DCTs, I think of “access.” I think of patient populations that we may not have been able to reach before, because perhaps patients couldn’t physically come to the study site for one reason or another. At Northwell, our experience with DCTs has been very positive.

Dr. Brennan shares an example of a successful DCT and lessons learned:

Dr. Khan: DCTs are a novel spin on trial design and trial conduct. I think by reducing the need for in-person visits and streamlining the process, we can certainly improve the efficiency and accuracy of trials while at the same time reducing the time and cost associated with the research itself.

DCTs allow us to enroll patients from a more diverse population and, as mentioned, really help us to focus on the patient – not just meeting them where they are but understanding them as individuals living with a certain condition and increasing their access to critical healthcare. I think there are challenges with DCTs, but as we move along and get more examples out there, I think they really are the way to go.

Cordisco: We have a lot of trials running where a component is decentralized. It might be a wearable or patients might have to put data into a device that’s given to them. It’s important that the instructions for those devices are extremely clear, written in a way that patients can understand and be in their native language. We can’t enroll patients in clinical trials if we don’t have the right information.

It’s also important if we’re going to use devices or deploy devices to a site that we know in advance what problems there may have been with them, so we can mitigate those at the start. If we know that a button always sticks or there’s something “funny” you need to do, we know that in advance and can help. Making us a part of the conversation early on is so important to achieve success.

Cordisco and Dr. Khan expand on the importance of having all stakeholders at the table early to design trial protocols:

Maloney: How are you integrating RWE into clinical trials?

Dr. Khan: RWE can be integrated into clinical trials in multiple ways to help improve the design and to conduct analyses of the studies. PAS uses RWE to provide insights on patient characteristics, treatment patterns and outcomes – all of which help to inform the design of the trial itself.

We’re also using RWE to “flip the funnel” for site recruitment. If you think about the traditional recruitment model, a sponsor contract research organization (CRO) will typically reach out to familiar sites and then leave it up to the site to find the patients for the studies. By leveraging RWE, we flip the funnel and find sites with high volumes of patients and then engage with those sites to enroll the trials, which results in an increased success factor because we know the patients are there for the study.

Cordisco: I agree with Dr. Khan, but from where I sit, I don’t see RWE coming down to a level where a research coordinator can use it unless we engage a third-party vendor. I know that our EHR has access to the RWE data sets, but it’s in Python [a computer programming language] and so that makes it hard for us to use it. There are ways around that of course, but I think the data needs to be more fingertip accessible to us.

I do think we can end up with a lot more information, publications and good research questions just by having access to the data.

Dr. Brennan: I agree that RWE isn’t trickling down to the site level and we’re not privy to some of that data, which would be very helpful. I want to turn it back to Dr. Khan and ask, especially in his new role, how do you obtain your RWE or how does the industry as a whole?

Dr. Khan’s response:

Maloney: We’ve talked about the importance of diversity in clinical trials. How is everyone addressing this issue?

Cordisco: To address racial and gender bias and increase diversity, you have to address it from an organizational level, not just my department or just the research department. I want to give a shout out to my organization in recently hiring a Health Equity and Diversity Initiative Officer. We’ve started to look at specific task forces that address racial and gender biases, specifically in our lab values and the way we treat hypertension patients and the lab values that go with chronic kidney disease.

We’re going to be using that data to recommend changes to our practice. The good thing about this is the Officer is fairly new to the organization and, right off the bat, has already included the research department early in the process. What this does is allow us to be there as they start to develop research questions and be able to look at the data that they’re looking at and say, “hey, we can help identify a study based on the data.”

Dr. Brennan: We do have a Chief Diversity and Inclusion Officer, and that was our inaugural officer back 10 years ago. We have the Office of Equity of Care, business employee research groups and a community outreach group, which we’ve tapped into for research. Essentially, what we’ve done throughout the years is reach the patient in their communities and where they feel the safest – in areas of worship and faith, barber shops and salons, community events, and even at Little League games.

Dr. Brennan and Dr. Khan further discuss the importance of language and communication in increasing diversity of clinical trials:

Maloney: What’s the best opportunity to deploy AI and machine learning (ML) in your organization, especially in clinical trials?

Dr. Brennan: I think the ability to use AI to help us match patients to trials would be very helpful. When you look at the data that’s out there for marketing and recruiting for your trial, the data shows that 80 percent of your recruitment will come from patients known to your organization or site.

When we know that 80 percent is going to come from within, how can we utilize AI to help find those patients, especially when we look at health systems and we may not know what patients are at what hospitals. And over 70 percent of hospitals today are part of health systems. That’s the way research is operating and evolving so I see AI playing a big role in recruitment.

Cordisco: I have to agree with that, there’s not much more I can add because where I see the benefit of AI at this point is being able to identify patients in a more exact manner. When we’re getting lists of potential patients from our IT department, it’s not good enough because we spend an enormous amount of time going through every single patient chart matching inclusion exclusion criteria.

For an AI system to do that more efficiently, to find patients that truly are a match, that’s really where the sweet spot is.

Dr. Khan: I think the power of AI really is the algorithms that can help you identify the patients and even provide insights on informed decisions and monitor the progress of the trial.

Maloney: We’ve covered quite a lot of ground today. I’d love to finalize by sharing what you think is the biggest opportunity around innovation in clinical trials.

Dr. Khan: I believe that the clinical research world at this point in time is in a great spot. The biggest opportunity we have is improving patient outcomes and experiences. New technologies, including what we’ve been discussing – AI and RWE – can help providers diagnose and treat patients more effectively and give greater access to medical care.

How do we bring these opportunities to life? I think it’s about keep moving, keep innovating. When I look back over the last decade and where we’ve come, we’ve come a long way. Where we’re more focused now is on patient centeredness, and as mentioned throughout the discussion, there are certainly a lot of challenges out there. But I think collectively, as researchers in the healthcare community, we’ve got this.

Cordisco: I think the opportunity and the challenge at the same time is connecting trial candidates with a health system that provides research. So, whether it’s DCTs or AI that we use to connect with them, I think being able to reach those patients and offer things they aren’t able to otherwise get is really where we are.

If they have an illness and you can make them feel better, then their quality of life is going to improve. So again, it’s outcomes, but I also think it’s outreach as well.

Dr. Brennan: We must keep up with technology (which we’re always behind in healthcare and clinical research) to have more opportunities enabled for our research participants but, at the same time, not lose sight of choices they’re familiar with. For some patients, the person they trust most in healthcare is their provider and may prefer face-to-face visits for that reason.

In closing:

For more on this topic:

• Tune in to the full webinar – Clinical Trial Innovation: How Healthcare Technology is Evolving.
• Deep dive into five key AI trends shaping the healthcare industry, including clinical research and discovery. Download our new eBook – The Pulse of AI: The Future of Technology in Healthcare.
• See how PINC AI™ Applied Sciences’ prospective research and clinical trial capabilities can benefit your organization.

The insights you need to stay ahead in healthcare: Subscribe to Premier’s Power Rankings newsletter and get our experts’ original content delivered to your inbox once a month.