The role that real-world evidence, and the data that feeds into it, play in the clinical trials arena is changing. For many years, drug developers and regulators have collected real-world data (RWD) on adverse events linked to newly approved medicines. However, this data has not played a significant role in developing or executing pre-approval clinical trials.
The situation is now changing. “There’s always been a need to look at how drugs that are coming out of clinical trials behave in the real world,” Gracy Crane, International Regulatory Policy Lead for RWD at Roche in the U.K., told Inside Precision Medicine.
“A lot of the safety monitoring has actually been done using RWD. But recently, there’s been a step change in that RWD has moved now from the safety space into more of an efficacy and effectiveness space.”
There are several reasons for this change. To begin with, there is simply more RWD available than there has ever been before and there are easier ways of collecting it. For example, advances in digitalization have made it possible to use electronic health record data, as well as claims data from insurance companies.
Other technological advances, such as the development of easily accessible and affordable wearable health devices have also contributed to this trend, and advances in artificial intelligence and machine learning are simplifying the analysis of big data.
“A lot of data is being collected in the real world, but also there are more opportunities to use advanced analytics in order to be able to interrogate that data,” commented Crane. “There’s no use just having tons of data, you have to be able to have a tool to be able to make sense of the data… It’s a perfect storm, all coming together in order to be able to deliver what we need.”
Another reason for the current interest in real-world evidence (RWE), created through the analysis of RWD, is a frustration with the way traditional clinical trials are carried out. While no one denies that randomized controlled trials are the gold standard for showing effectiveness of new therapies, they also have problems, such as high costs and a historical lack of diversity among participants, among other issues.
“Effective clinical trials are very critical for the advancement of health because they enable us to introduce new therapeutics into the market to help address cancer and other diseases. And yet, only 10% of cancer patients participate in a clinical trial and that percentage is even lower for minority and underrepresented groups,” noted Warren Whyte, vice president of scientific partnerships and head of the Engaging Research to Achieve Cancer Care Equity program at ConcertAI, a company using RWD and AI to help improve clinical trials in oncology.
In 2016, the U.S. government passed the 21st Century Cures Act, which aimed to streamline drug and device development and bring treatments to patients faster. Recognizing the potential of RWE to support regulatory decision making, including approval of new indications for approved drugs, Congress added a clause stating that the U.S. FDA should create a framework to evaluate the potential for RWE to support new drug indications, or to satisfy post-approval study requirements.
“RWD has been around for some time, but it’s particularly top of mind right now as the FDA issued a series of guidelines last year that point to its future use for drug approvals,” explained Amy Abernethy, former FDA principal deputy commissioner and current president, clinical studies platforms, at Verily Life Sciences.
“In these, the FDA highlights what they’ll be looking for as they consider the RWD in submissions, including things like data quality, methods for linking between different datasets, and statistically robust methods for analyzing RWD.”
Using Real-World Data to Implement Precision Medicine
Although randomized controlled clinical trials are the gold standard that drug developers aspire to achieve, sometimes it is not possible or even ethical to randomize patients into a trial. This can be especially relevant for pediatric, rare or orphan diseases where patients are hard to find and often critically ill.
Marjorie Zettler is a cancer specialist and executive director of clinical science at Regor Pharmaceuticals. She explained that the use of RWE to support cancer drug approvals, for example, as an external control, is becoming more common.
Expanded access or ‘compassionate use’ is a route for very sick individuals to be treated with investigational drugs that have not yet reached the market. Data from these patients is collected and is increasingly being used to support drug approvals.
For example, in 2016, blinatumomab was approved for acute lymphoblastic leukemia in children, partly supported by data collected from individuals under the age of 18 in a single-arm, open label, expanded access study.
“For childhood cancers, there’s such an unmet need there and the trials can be so difficult to do. In my own research, I found that expanded access data to support efficacy or safety of the product was the predominant type of RWD that’s been used to support pediatric approvals,” said Zettler.
For many rare diseases, recruiting to randomized trials can be a challenge. There are low numbers of people affected and the health status of the patients can be critical and time sensitive, meaning many simply cannot afford to miss out on a potential treatment opportunity.
There is a chance to improve trials in rare diseases by bringing in an external RWD control to what might otherwise be a single-arm trial and this is something Roche has made use of in the past. “Risdiplam [for treatment of spinal muscular atrophy] is a great example where we have utilized RWD to better understand the natural history of the disease and also as an external control for our trials,” said Crane.
A problem with many randomized controlled trials is they have (at least historically) been carried out largely in White, male populations. This can create a distorted view of certain diseases, such as cancer. For example, for many cancers, Black patients are diagnosed later, with more advanced disease, compared with patients of other races or ethnicities, notes Zettler.
In April 2022, the FDA released draft guidance for industry to improve enrollment of participants from underrepresented racial and ethnic populations. While this is only a guidance document, it is at least a step in the right direction.
“We can use real-world data to understand our patient population better so that we can design more inclusive trials,” explained Zettler. “Real-world data can also help us choose our clinical trial sites. There was a really simple but elegant study a few years ago, where the authors used Medicare claims data to map the zip codes of African American and Asian American patients with lung cancer… They put those on a map and when we see them side by side, we can see the obvious conclusion was that there’s a major disconnect between U.S. lung cancer clinical trials, site placement and where diverse patient populations live.”
Whyte says he thinks RWD can help diversify trials in three ways. First, by helping organizations better understand why some people may not be participating in clinical trials. Second, by evaluating old approaches, identifying challenges and using them to find new solutions. Finally, by identifying new sites, neighborhoods and communities where trials have not been conducted in the past.
“We understand that there are certain challenges that may prevent people from going to sites where trials are being conducted,” said Whyte. “If we could find ways to identify community centers that have the infrastructure and the resources to conduct clinical trials, and they’re right in these people’s backyards, then we owe it to them to invest in the infrastructure and whatever is necessary to have that clinical trial be conducted in their neighborhoods. That way, it makes it far easier for people to enroll.”
Real-World Data Challenges
The use of RWD and RWE in clinical trials is increasing, but some experts have reservations about how useful it can be.
“Without randomized evidence, one can’t necessarily be certain as to whether a treatment is actually safe and effective or not,” Louise Bowman, a professor in the clinical trials service unit at the University of Oxford, told Inside Precision Medicine. “Our worry is that inappropriate use of real-world data may not be giving us the right answers when it comes to the management of patients.”
Virginia Nido, global head, industry collaborations at Roche, agrees that using RWD effectively can be difficult. “I think the real question is, is it possible to actually change it into evidence that could be used to show whether or not a drug or a device is safe and effective for human use? I think in some cases, it’s possible to do that. But it’s hard.”
Nido advocates instead for the sharing of collected clinical trial data to be used for new trials, either in the design or implementation stage.
“There’s quite a lot of products that companies might have been working on, that we’re just no longer interested in pursuing. That data is sitting on a shelf somewhere… we’re finished with them and we won’t be further analyzing that data. So why not share it,” she emphasized. “RWD is very easy to get and very difficult to use and clinical trial data is very hard to get and much easier to use.”
Reliability of data can be a problem in the RWD and RWE space. If information such as that stored in electronic health records or collected by health insurance companies is not collected to be used in a clinical trial it may not be truly applicable to a research question. Other problems can include missing data, such as tests never ordered in clinical practice and tests ordered but not done, or patients moving from one unconnected medical system to another.
“Many clinical trials evaluate changes in outcome measurements over time, perhaps based on standardized assessments to generate uniform data. If these latter types of outcomes are not used in clinical care or are not recorded consistently, they can be challenging to measure using RWD,” explained John Concato, associate director for real-world evidence analytics in the Office of Medical Policy at the Center for Drug Evaluation and Research, FDA.
Data linkage, where several data sets are combined and then the gaps between sets are filled in, is one way to address missing information in RWD. “This means you have to deal with the data quality challenges that come from combining datasets such as data conflicts from multiple source datasets representing the same variable, which is something we’re working to advance as an industry,” said Abernethy.
Nancy Dreyer, epidemiologist and chief scientific officer emerita at life science analytics company Iqvia, has worked on RWD and RWE for many years. She helped develop the Good Research for Comparative Effectiveness (GRACE) guidelines, now used by researchers and regulators around the world, to help judge whether real world studies assessing comparative effectiveness are rigorous enough to be used for assessing medical treatments and technologies.
She agreed with Abernethy about data linkage. “One of the trends we’re starting to see is the value of data linkage in clinical trials. You can take your very strong clinical information and extrapolate it to a larger population… it allows you to do follow up beyond the end of the trial, and it also allows you to understand a lot more about the path to diagnosis.”
Creating Next Generation Clinical Trials
Randomized controlled trials are still the goal for most therapeutic product developers. While it can be complicated to use, it seems clear that RWD is here to stay. Even those skeptical of RWE use for assessing the efficacy of medicines and other therapeutic products agree that RWD can be invaluable for improving trial design, running clinical trials and assessing post marketing safety.
“We should be using these data to do trials better and more efficiently. I think RWD has a really valuable potential, if used correctly, and if not over interpreted,” said Bowman.
Developments in precision medicine, technology and data analysis; the COVID-19 pandemic; and concerns about the lack of diversity, long duration, and high cost of traditional clinical trials are all driving change in the field.
Precision medicine concepts are increasingly being used to design basket trials, which assess therapies for diseases sharing a common genetic mutation such as different types of cancer, and umbrella trials, evaluating multiple therapies for different molecular subgroups of the same disease or condition. The pandemic has contributed to an increase in decentralized trials, where patients are monitored or treated using telemedicine or medical services close to their home, which seems set to continue.
RWD is being incorporated into new trials as external controls, when appropriate clinical trial data is not available. In November 2022, the FDA announced it would be doing a large trial with the National Cancer Institute, two pharmaceutical companies and an academic cooperative trials group called ‘Project Pragmatica,’ a large pragmatic trial of a combination therapy for advanced non-small cell lung cancer.
Pragmatic trials—where inclusion and exclusion criteria mirror the real world and follow-up is done in clinical practice—are also embracing real-world principles. The widely cited RECOVERY trial, set up in the U.K. to find therapies for COVID-19 is a good example of a successful pragmatic trial.
“With pragmatic trials, there is an opportunity to bring in the randomization part. That’s why this is attractive to health authorities because you’re preserving the randomization. But you also have the potential benefit of having a trial from clinical practice,” explained Crane.
“One of the key challenges is the operationalization of these types of trials, because obviously by loosening up the inclusion/exclusion criteria, there’ll be challenges with assessing the efficacy. We have to be able to understand better how we, by having those loose inclusion exclusion criteria, evaluate efficacy in those scenarios.”
Concato thinks that combinations of the new trial types could be key to improving design and costs of clinical trials. “Real-world data tend to contain data on diverse study populations, providing potential opportunities for addressing health disparities. Specifically, decentralized trials and trials involving real-world data have the potential to reduce costs & increase study diversity while generating robust evidence.”
Source: https://www.insideprecisionmedicine.com/topics/informatics-topic/bioinformatics/broadening-the-scope-of-clinical-trials-the-changing-role-of-real-world-evidence/