Challenges With Feasibility Studies

ACT: What are some of the biggest mistakes stakeholders tend to make with their feasibility studies? How can those challenges be addressed?

Nambisan: So I think historically, feasibility studies have been managed through surveys that go out to investigators that ask them how many participants of a particular type there are. And anecdotally, when you talk to folks who manage feasibility teams, typically the approach comes back like, “Okay, if this investigator says they have 10 participants, we'll just dock that by 50% and expect them to have five.” Right? So it's not a precise science, in some senses. What I believe at this point is it's more of an assessment on motivation of that investigator to participate in a trial, rather than highly quantitative numbers as to what you can expect in terms of the participant populations, etc. I think the other aspect is, oftentimes, we discount the fact that investigators that may have more diverse populations of patients that are maybe more representative of particular disease states may not be good investigators, because they haven't had much experience as investigators in the past, for historic biases, etc.

So I think in relation to the first point I made, I think it's absolutely critical to supplement some of that assessment of motivation on the investigator side by an understanding of participant densities, looking at multiple data sets, some of which I've described real world datasets primarily, at an eligibility criteria basis, understanding, no, we're not just looking for diabetes type two patients, we're looking for these diabetes type two patients with lots of other different characteristics. How do they cluster? Where did they get care? How many of them are there in specific geographies? Right? And that gives a much better assessment of what might be feasible as compared to a survey that comes back. And to be clear, I'm not discounting the fact that investigator motivation is very important and feasibility, it is, it is one factor that's important. But historically, we've used that as the input for the quantitative model rather than an assessment of motivation.

Now on to the second point that I made about maybe disenfranchising investigators that may actually have diverse populations. I think it's fair to say that historically, the representation in clinical trials has been massively dominated by Caucasians in the US at least, to be clear. And on top of that, male Caucasians, primarily. And so I think that we've seen some challenges, many different challenges, not just access and perception of, let's say, vaccines during COVID, for example, but we've seen serious safety events occur in the past when specific therapies were not tested on those populations that best represent the disease state. And so I think it's important, especially given FDORA (Food and Drug Omnibus Reform Act of 2022) coming about in the new legislations about diversity in the feasibility period, to start looking at new datasets, not just your preferred provider networks and provide preferred site networks that you've historically drawn from in the past, as a CRO or sponsor, you’ve got to look outside that, look at community healthcare, you’ve got to look at different healthcare systems that may not have traditionally been academic medical centers that have been in say, 50% of the trials for oncology. And in order to do that, you're going to have to bring in different datasets that you haven't brought in historically. And that brings up the point of, again, with these different approaches to feasibility, it's not just about sending a survey out to the last 10 Investigators you've worked with in the last 10 trials, but looking at the real world data understanding where these participants clustered, identifying the physicians that treat those patients that may also be more representative of the disease states, and then finding a way to create collaborations with those investigators, those institutions as well to truly represent the condition under investigation.