Real-World Data, Digital Patient Profiles Are Revolutionizing Trial Design, Patient Recruitment: Gen Li, PhD, MBA

Innovative digital patient profiles and comprehensive real-world data are transforming oncology clinical trials by driving smarter study design, improving patient recruitment accuracy, and lowering the risk of trial failure, explains Gen Li, PhD, MBA, president and founder of Phesi.

The American Journal of Managed Care^® (AJMC^®): Phesi's analysis shows a massive volume of real-world data for cancers like colorectal, breast, and lung. How is the trial accelerator platform helping sponsors translate this data into a smarter trial design and more precise patient recruitment?

Li: In a typical situation, we are usually needed to be guided by some sort of document from the clients. That can be a very simple outline of a protocol or various versions of the protocol at different development stages. Guided by those documents, we start to "recruit" patients according to the design from that 106 million patient pool, using the same kind of inclusion/exclusion criteria—just like you do in a clinical trial—except we’re not in the hospital but working from a database.

With that kind of patient data—for example, nearly 6 million of those are colorectal cancer patients—we then start looking at the alignment, or misalignment, between the patients we are seeing in reality and the design of the trial itself. Not surprisingly, that kind of process basically lends us a view of where the design of the treatment needs to be improved. That can be the inclusion/exclusion criteria, the outcome measures, or the treatment duration. Basically, 360 degrees of the trial design can be examined against real patient data.

From there on, in that kind of stage—which we call a digital patient profile—we’ve constructed it to improve the design of the trial and to assess the feasibility of operationalizing the trial. Usually, as a result of that process, we’re able to reduce the protocol amendments that typically happened in the past.

As we gain an understanding of the body of the patient profile, we also start to learn who the players are—the care providers and the physicians working around those patients. That way, we’re able to improve precision in identifying the investigator sites that are going to be much better at helping us recruit patients wherever they are around the world.

Product design improvement is one direction; operationalization precision is another direction our capabilities are going. Then we can go further. As soon as we get the patient to be "perfectly aligned" with and finalized in the protocol, the digital patient profile is transformed into a digital twin of the patient. If you’re going out to recruit this patient, it's at the baseline—so it becomes a digital version of the patient at the baseline level of the real clinical trial.

That also allows us to look at the control arm. It can be an active control arm, a standard of care, or a placebo. We can use the same treatment duration according to the design of the trial to project the safety and efficacy outcome—without actually doing that kind of trial.

That becomes a complete set of digital data. And that digital set not only can be used for replacing the control arm, but it also has many other benefits. For example, if we have a small clinical trial with very few patients, we can use that digital set to better interpret the results with a higher level of precision—especially for smaller clinical trials—to avoid attrition when you're going from phase 1 of an oncology trial to phase 2, or from phase 2 to phase 3.

We have a much better idea of what those amplified, perennial trials would look like—without conducting that traditional trial. I can give you many other examples, but the value has been quite substantial. It’s not an exaggeration to say that this platform, given the time and opportunity, will transform what clinical development will look like.

AJMC: How do you see tools like these reshaping the early stages of clinical trial planning, and what impact could they have on reducing trial failure rates in later phases?

Li: What we announced with the 106 million patients—it’s a snapshot in a long process. That process is still ongoing. A few years ago, we started with 30 to 60 million patients, and that has gradually grown to 106 million—and now 167 million at this point. And it’s still growing.

What we are inspired to do is basically collect every single possible footprint left by human beings in our ability to collect and document patient data. Regardless of where they’re coming from, who they are, who collected the data, where the data was collected, or under what kind of study design the data was collected, we try to include it in our database.

By doing that, we are obviously increasing the statistical confidence—the level of confidence—in those data but also giving ourselves the opportunity to remove or minimize the bias in how the data is being taken and analyzed, because we are trying to look at the totality of the data.

That means not only are we able to look at some of the very large diseases that have been extensively studied, but we're also able to look at much more rare situations. And when I say rare situations, it could be a particular biomarker presented in a cancer that appears in a very small percentage of patients, or it could be rare diseases in general.

One recent example is a case mentioned in various literature. There are about 500 patients around the world, and we were able to actually construct a digital patient profile including over 200 of those patients—only male patients. That’s the type of work we’re doing and the type of freedom we now have to help people design better trials, realistically operationalize those trials, and put controls in place without exposing patients to high-risk situations.