Real-World Evidence: Advice, Principles And Examples Emerge From FDA

Biopharma sponsors are eagerly awaiting guidance on how to apply 'real-world evidence' to drug development and regulatory decisions. Recent workshops provide some clues about what is likely to be included in the first drafts.

US FDA officials appear most comfortable applying “real-world evidence” to the challenges of patient selection for clinical trials – rather than jumping ahead to using those data sources as a basis for a drug approval decision.

During a week of workshops focused on the theme of real-world evidence in mid-September, FDA officials repeatedly made the point that the agency can, does and will use real-world evidence to support regulatory decision making – but not that it will tie drug approval decisions directly to applications built on real-world data.

Instead, FDA officials suggested ideas like using data from health systems electronic records and claims to help identify populations for more effective Alzheimer’s Disease clinical trials. That specific suggestion came from Center for Drug Evaluation & Research Deputy Director Robert Temple, speaking from the audience during a Sept. 13 public workshop on creating a Regulatory Framework for the Use of Real World Evidence put on by Duke-Margolis under a cooperative agreement with FDA.

Temple suggested that real world data might provide a way to define an at-risk population before they “are overtly sick” and thus permit sponsors to work with endpoints to show impact on progression in a short-term (two- or three-year) period.

“The endpoints,” Temple said, could “be an early diagnosis of mental dysfunction or next stage” of disease. That seems “sort of promising,” Temple said, “because those are things that are picked up in the system” of real-world data.

Temple’s comments at meetings like the Duke-Margolis workshop may appear as ruminations or digressions from the main topic. However, given the ongoing work to draft guidance on applications of real-world evidence (RWE), it is hard not to interpret them as a guidepost to the agency’s thinking at this point.

In this case, using the multiple-failure clinical trial area around Alzheimer’s as enticement, Temple suggests that an appropriate use of real-world evidence is to improve the design and potential for success from clinical trials.

That’s a step back from the firm commitment that many in industry have been seeking that FDA will accept real-world data as evidence for regulatory decisions on certain types of submissions, such as supplemental and post-marketing applications.

However, Temple is defining one of the safe functional uses for real-world data at this point, and presumably identifying one idea that may be incorporated into the guidance that FDA is working on as part of its commitments under the PDUFA VI agreement and in response to the directives of the 21st Century Cures Act.

A statement on the value of RWE to support endpoint development and tailor populations for shorter clinical trial trials would be akin to the general advice sections of some of FDA’s drug guidances like the first part of the recent January guidance on “Multiple Endpoints in Clinical Trials.” (See sidebar.)

The RWE effort has the clear support of FDA Commissioner Scott Gottlieb. He told one of the recent workshops that “We’re now working on policies to support the use of RWE in the approval of new indications for already marketed drugs. This may be especially relevant in settings like rare diseases or other unmet medical needs, where it can be hard to enroll patients in clinical trials.”

Two of the three September workshops were held by the Duke-Margolis Center and one by the National Academies of Science, Engineering and Medicine. All three were in conjunction with FDA and are part of the ongoing effort to move FDA and the regulated industry to a better understanding and regulatory framework for more widely incorporating real world data and evidence into regulatory applications.

Gottlieb declared the agency’s current view that it can accept real-world evidence at its discretion in a keynote address to the Sept. 19 NASEM meeting. “Make no mistake,” Gottlieb said, “there’s nothing in our statute or regulations that prevent FDA from using a broad range of informative sources of evidence."

FDA Principal Deputy Commissioner Rachel Sherman struck that same theme later at the workshop. She pointed out at the that neither the drug or device standards for approval “talk about where or how one collects the information, data, or evidence.”

Center for Drug Evaluation & Research Director Janet Woodcock was more specific about the types of projects where FDA has already accepted real world-evidence. Like Temple’s advice on patient selection, these situations are likely to be reflected in the upcoming guidance.

From the current hints from FDA, the guidance looks like it will incorporate advice, principles (randomization and pre-specification of endpoints and analysis add reliability) and examples.

Here are two examples offered by Woodcock:

How to incorporate registry data into support for a new indication: Woodcock’s example of recent FDA activity in this area is the May 17 extended indication (from 10 cystic fibrosis mutations to 33) for the Vertex product, Kalydeco (ivacaftor), based on registry data with mechanistic information from lab studies and trial data. (Also see "Kalydeco Expands Indication Without Clinical Data; Keytruda Is Latest Bladder Cancer Approval" - Pink Sheet, 21 May, 2017.)

Woodcock described her active involvement in the agency’s decision to use a combination of registry data, mechanistic lab studies and trial data.

“I asked the Cystic Fibrosis Foundation to look” at their “very good” registry for data on FEV1 readings before ivacaftor treatment and after, Woodcock recounted. “They found that they had a sustained increase in FEV1 after switching over to ivacaftor. Subsequently we were able to use mechanistic data along with some trial data to add those mutations to the indication.”

The CDER director projected this experience to other diseases. “You could see in the future that you could use this [approach] to indications to different diseases especially as we’re approving sort of mutation-by-mutation in these targeted therapies.” She specifically cited the relevance to oncology.

Creating an environment for merging clinical trials with treatment (“platform trials”): Woodcock continues to push for more sponsors to participate in multi-product trials run under master protocols – “continuous ongoing trials rather than start-and-stop trials and they can study many things” such as biomarkers, devices, and outcomes in a disease. (Also see "Master Protocols Are Both Welcome And Inevitable – US FDA's Woodcock" - Pink Sheet, 6 Jul, 2017.)

Woodcock described these trials, which have to this point “been run by consortia that have experienced trialists as principal investigators,” as a source for evidence generation from which FDA would feel more comfortable accepting a range of data. Woodcock explains how these trial environments could effectively merge treatment (aka real-world) data with clinical trials.

“I think the most important thing here is that because they are continuous rather than start-and-stop, they could add elements to incorporate data collection and trial procedures into the care process,” Woodcock said. That would give data generated from these trials “the characteristics of real-world evidence but continue randomization and certain other things that are necessary … for studying new products.”

From the editors of the RPM Report