Research shows that real-world data is not yet ready to replace post-approval confirmatory testing

For drugs and biological products that have received accelerated approval, can information extracted from real-world data sources replace traditional post-approval clinical trials to confirm the benefits of these therapeutic products? A recent study pointed out that real-world data (RWD) are not enough to confirm the benefits of accelerated approval based on surrogate endpoints
.

Joshua Wallach and co-authors of the Department of Environmental Health Sciences of the Yale School of Public Health stated in a health policy research letter [1] published on JAMA Network Open on November 9, 2021, “The results of this horizontal study show that in 2009 Among the 50 FDA-required post-approval confirmatory trials for therapeutic drugs that received accelerated approval between 2018 and 2018, none of them can use currently available claims and/or structured electronic health record data
.

"

The FDA has issued guidance on how to use the evidence extracted from RWD to meet the requirements of post-approval research
.

However, Wallach and his co-authors pointed out that using RWD-based research to replace post-approved clinical trials has many limitations, including difficulties in extracting data from electronic health records (EHR), billing, and claims data

Nonetheless, the potential advantages of constructing RWD-based post-approval studies are still tempting: “A better understanding of the feasibility of simulating the post-approval studies required by the FDA to confirm the clinical benefit is critical, as these studies often face recruitment challenges and continue Focus on surrogate markers as an end point, and confirmatory studies were postponed for several years after approval
.

"

The study followed the Enhanced Epidemiology Observational Research Report (STROBE) guidelines and reviewed all 50 post-approval confirmatory trials of 41 new molecular entity drugs and biological products that received accelerated approval between 2009 and 2018
.

For each drug or biological product that received accelerated approval, the researcher reviewed the information provided on ClinicalTrials.
gov along with the abstract and study description to determine whether the RWD source can provide data similar to the data obtained during the post-approval trial
.

Specifically, Wallach and co-investigators measured the number of trials for clinical indications, controls, and primary endpoints that can be "routinely determined from RWD

Of the 50 post-approval confirmation trials reviewed by the research team, only 12 (ie 24% of the trials) had clinical indications that could be more easily collected from claims or structured EHR data
.

Of the 46 trials with inclusion and exclusion criteria found on ClinicalTrials.

The researchers did not find that any one trial could collect all the necessary data from RWD
.

They wrote, “If researchers find it difficult to use RWD sources to develop a computable phenotype, then these characteristics are considered unlikely to be routinely determined from observational data

[1] Wallach JD, Zhang AD, Skydel JJ, et al.

[1] Wallach JD, Zhang AD, Skydel JJ, et al.
Feasibility of Using Real-world Data to Emulate Postapproval Confirmatory Clinical Trials of Therapeutic Agents Granted US Food and Drug Administration Accelerated Approval.
JAMA Netw Open.
2021;4(11) :e2133667.
doi:10.
1001/jamanetworkopen.
2021.
33667