Demonstrating the value of RWE

A new cross-industry effort hopes to research the benefits of RWE and promote acceptance of alternatives to traditional trials.

While the buzz surrounding real world evidence (RWE) may make it seem like it is already widely accepted, this is not necessarily the case.

“A big question is whether real world outcomes can be trusted,” explains Dr Nancy Dreyer, chief scientific officer and senior vice president at the IQVIA Real World Solutions Center for Advanced Evidence Generation.

“Without established guidance like that used elsewhere in drug development, those who want to use RWE are left guessing as to whether their design, data and processes will be acceptable. The appeal of rich, diverse real world data (RWD) is tempered by a lack of scientific, operational and regulatory know-how.”

Another concern has to do with channeling treatments.

“Random treatment assignment in randomised clinical trials (RCTs) is used to balance known and unknown confounders,” says Dr Dreyer. “In real world settings, treatments are not prescribed at random but are systematically given to those considered most likely to benefit or are prescribed according to an algorithm of treatment sequencing devised by a given pharmacy benefit program. This makes comparison of real-world effectiveness more challenging, since some of the factors needed to understand why particular treatments have been prescribed are not recorded or accessible.

“In pharmacoepidemiology, we have a variety of design and analytic tools we use to address these concerns, but some of them rely on black-box implementation, which is difficult to explain and not conducted using a standard approach, which in turn makes it difficult to evaluate whether the right tools were used. Analyses become even more complex as we use machine learning tools to assist in identifying comparable groups.”

She adds that another factor holding back adoption of RWE is the lack of fit-for-purpose data quality standards.

“We need to make a common set of key descriptors available for every RWD source, including (1) the provenance, (2) data structure, (3) how the data are integrated and (4) what curation is conducted to ensure that the data are appropriately coded and linked. This approach will help users understand what data are available and what data may be systematically missing.”

This can be frustrating when there are already many well-understood ways in which RWE can be used to enhance clinical research.

Speaking about RWE’s potential in oncology, Dr Jennifer Christian – vice president clinical evidence within the IQVIA Real World Solutions Center – says that one transformational change would be to bridge the divide between clinical research and clinical care, so that capturing and using data from routine care becomes fundamental to the evaluation of clinical effectiveness and safety of new treatments.

“This would require a systemwide reorientation by embedding interventional studies directly into routine care and, as a result, would improve the quality of data that is routinely captured, reduce the cost of traditional clinical trials and produce meaningful results for clinical decision-making,” she says.

In the absence of this framework, there are more immediate ways RWE can be embraced in clinical research.

“We can use RWD to inform the design and implementation of trials by optimising protocols, identifying patients who qualify for studies and enrolling sites that have eligible patients,” says Dr Christian. “We are already beginning to see significant savings by using RWD this way, including reductions in protocol amendments and faster recruitment times.

“Clinical development strategies are beginning to include RWE for measuring expected safety and effectiveness background rates and providing context for interpreting the results of single-arm trials.

“Greater adoption is needed in extension studies, where RWD is used to follow patients after a clinical trial ends to assess longer-term benefits and safety of treatment regimens. There is also a need to link RWD sources to capture the complete picture of clinical care that patients receive from various providers across institutions, which incorporates labs, imaging, outpatient prescriptions, genetic testing and mortality. A more robust and rigorous evaluation of cancer care will enable greater precision in treating each and every patient.”

Can RWE compare to tradition research?

In an attempt to help address challenges in adoption and harness the full potential of RWE, IQVIA has teamed up with Friends of Cancer Research to learn about the similarities and differences between RWE and randomised controlled trials, and encourage wider adoption.

As part of this project, a study involving six research centers in the US followed a common protocol to assess real world endpoints among cancer patients, including overall survival, time to next treatment, time to treatment discontinuation, time to progression and progression-free survival.

Results from the collaborative project, An Exploratory Analysis of Real-World End Points for Assessing Outcomes Among Immunotherapy-Treated Patients With Advanced Non–Small-Cell Lung Cancer, were recently published in JCO Clinical Cancer Informatics.

Researchers used non-identified patient data from assets such as administrative claims and electronic health records to assess real world endpoints and found that they were generally consistent with each other and with outcomes observed in randomised clinical trials.

The study’s findings help confirm that clinical benefits seen for therapies in clinical trials were consistent with the benefits of those treatments within real world settings.

Dr Christian notes that there were notable differences in the patient characteristics across data partners, such as mean age of receiving frontline therapy, histology and staging, that may explain variation in survival rates observed.

“A few of the organisations are working together now on restricting the general real-world population to match more closely with a trial population.”

She adds that IQVIA is beginning to learn that less than 60% of the total population will be included once they apply the I/E criteria and that survival rates are worse among the general real-world population compared to patients more similar to the trial population.

“These preliminary findings suggest that the average results from trials are not generalisable to everyone with advanced non-small cell lung cancer and, therefore, it is important to generate meaningful endpoints from RWD as well.”

Meanwhile, Dr Dreyer says that it’s key to remember that there are still meaningful differences between RWE and traditional data.

“Consider, for example, benchmarks of disease progression for single arm clinical trials, known as external comparators. In oncology, disease progression is measured through systematic examination of biopsies, a practice not followed in typical care settings. The real world proxy used here is “time to next treatment” (rwTTNT), which is based on the presumption that, generally, an effective treatment wouldn’t be stopped.

“We see that rwTTNT is always longer compared to disease progression measured in RCTs. The importance here is that comparisons between RCT and RW on disease progression can systematically make the study drug look worse, regardless of its effectiveness.

“Once understood, however, these differences can be addressed in the analysis.”

RWE in regulatory settings

The researchers also hope these results will substantiate the validity of using real world data to support decision-making by regulatory agencies and healthcare payers – something that Dr Dreyer says already has significant momentum across the world.

“While regulatory use of RWE has traditionally been focused on evaluating safety and effectiveness of risk mitigation strategies after product launch, we are seeing a dramatic increase in global use of RWE to support drug approvals and label expansion,” she says.

“The newest use is to provide context for single arm trials of rare diseases and other conditions that are difficult to study using traditional randomised, double-blinded controlled trials.”

The FDA has already been using RWD as external comparators for some decision support and is aiming to produce guidance on use of RWE by 2021.

In Japan, the chief executive of the Pharmaceutics and Medical Devices Agency recently announced that it will produce guidance in March 2020 on its basic position on use of RWD and ‘points to note’ in submissions.

Meanwhile, China’s National Medical Products Administration released draft guidance this year about using RWD for regulatory submissions.

“The Europeans have really dug into the details of RWD, producing an in-depth report on this topic in February 2019,” adds Dr Dreyer. “The report from the Heads of Medical Agencies and the European Medicines Agency noted that RWD is generated under different scenarios and for different purposes, which rarely include medicines regulation, and that data ownership resides with multiple stakeholders, many of whom have no need to engage with the regulatory system. This means that regulators can’t impose the same requirements for data collection and verification as used in traditional RCT. Hence the need to develop new guidance that is appropriate for RWD.”

Dr Christian adds that regulatory policy changes mandating the use of RWD like these have been one of the key drivers for greater adoption and use of RWE.

“Indeed, the more examples that are shared publicly where RWE was used to inform a regulatory decision will lead to greater adoption of RWE by the industry for the evaluation of new treatments.”

Nevertheless, Dr Dreyer says that she doesn’t see RWE as replacing RCT altogether.

“Rather, I see RWE starting to take its rightful place as a complement to RCT. For example, 50% of the new cancer drugs approved using surrogate endpoints showed no benefit in overall survival. This real world endpoint shows the importance of having information about outcomes that are important for patients and clinicians.

“While surrogate markers serve an important role in drug development, we need to balance our desire to have new drugs approved sooner with the need also to establish their real-world profile in terms of benefits and risks. Running complementary RWE and RCT research would meet this need.”