FDA Finalizes Guidance on Leveraging Real-World Data for Medical Device Submissions

The proliferation of electronic health information has at once presented exciting opportunities for innovation in health care as well as regulatory and business challenges regarding the proper use of such data. One such opportunity centers on the leveraging of data that were originally collected for purposes of routine clinical care or billing as evidence to support U.S. Food and Drug Administration (“FDA”) decisionmaking for FDA-regulated products. Because data are required to establish the safety and efficacy of certain products, the pharmaceutical, biotechnology, and medical device industry have closely monitored for indications of FDA’s position on the proper methodological approach, collection, and use of certain categories of data to support its regulatory decisionmaking. In that regard, FDA Commissioner Scott Gottlieb has made it an agency priority to provide more clarification on the proper use of “Real-World Data” (“RWD”) and its progeny “Real-World Evidence” (“RWE”) to support (and possibly expedite) FDA regulatory activities, including review of premarket submissions and new product functions. The effort to encourage the use of RWD and RWE also aligns with FDA’s Digital Health Innovation Action Plan and may incentivize market entry for new products or indications with safety and efficacy supported by less conventional data sources.

RWD, as a topic of conversation, is not necessarily novel, but it has gained traction in recent years. The 21st Century Cures Act (the “Cures Act”), enacted Dec. 13, 2016, has renewed the push to leverage such data. Specifically, the Cures Act requires FDA to evaluate the potential use of RWE and develop a framework for the use of RWE to support the authorization of new drug products and to support or satisfy post-approval study requirements. Since the Cures Act was signed into law, FDA has issued guidance and hosted public workshops on the use of RWE in medical device regulatory decision-making and the development of a framework to assist with “regulatory acceptability” of the data. However, the FDA Center for Drug Evaluation and Research’s “plan” to implement the Cures Act requirements is not statutorily required until Dec. 13, 2018.

On Aug. 31, 2017, FDA issued the “Use of Real-World Evidence to Support Regulatory Decision-Making for Medical Devices” final guidance (the “Guidance”). The prior draft guidance was issued on July 27, 2016 (the “Draft Guidance”). The Guidance describes FDA’s position that data derived from real-world sources may be used to support FDA regulatory decisions and clarifies the criteria by which FDA evaluates RWD to determine whether the data are sufficient for generating the types of RWE that can be used in regulatory decision-making. The Guidance is broad and applicable to all medical devices (including software regulated as medical devices) that meet the definition of a device under the Federal Food, Drug and Cosmetic Act (“FDCA”). Notably, the Guidance expressly maintains the “reasonable assurance” evidentiary threshold that FDA-regulated devices must meet, while also allowing stakeholders discretion in meeting this threshold—as long as the characteristics (outlined below) are met.

The Guidance also includes recommendations for determining when the collection of RWD would be subject to investigational device exemption (“IDE”) requirements. In that regard, and likewise in furtherance of directives in the Cures Act, the Guidance signifies FDA’s effort to address longstanding ambiguities in how the collection and use of RWD must comply with FDA human subject protection requirements and to harmonize FDA’s approach toward secondary use of data with that under the Health Insurance Portability and Accountability Act (“HIPAA”) and the federal human subjects research regulations known as the “Common Rule.” Medical device manufacturers as well as health care providers and other sources of RWD should consider these parameters as they assess how to leverage such data for purposes of FDA submissions.

I. The Guidance

a. Overview and summary

A premarket submission—a premarket notification (“510(k)”) or premarket application (“PMA”)—is submitted to FDA before a manufacturer may market a medical device, subject to certain exemptions. The evidence used to support the clearance of a medical device does not always require clinical data (e.g., in low-risk products). Instead, manufacturers provide other types of performance data to FDA to support the product’s clearance or approval, including non-clinical bench, animal, and laboratory study information. When data from a clinical trial are used to demonstrate substantial equivalence in a 510(k) or are otherwise included in a PMA, the clinical trial is required to comply with IDE, institutional review board (“IRB”), and human subject protection regulations. The data must constitute “valid scientific evidence” that demonstrates the benefit-risk profile of the product—providing reasonable assurance that the device is safe and effective.

Notwithstanding a reliance on certain of these distinct and conventional types of performance data, there exist troves of real-world performance, experiential, and outcomes data that may contain valuable insights regarding the safety and efficacy of medical devices. Mobile medical apps, wireless-enabled medical devices, and other tools that facilitate remote care delivery and the ongoing collection of robust health-related data are further indications of the promise that such information presents for purposes of supporting regulatory decisionmaking.

b. RWD: Defined and characterized

Under the Guidance, FDA defines RWD “as data relating to patient health status and/or the delivery of health care routinely collected from a variety of sources.” RWD may include data derived from electronic health records (“EHRs”), billing and claims information, product registries (e.g., the National Evaluation System for health Technology (“NEST”)), and mobile devices and other digital health-care tools. The general distinguishing feature of RWD is that they are collected in a real-world setting. RWE, in turn, is derived from the use and analysis of RWD and is defined as “clinical evidence regarding the usage, and potential benefits or risks, of a medical product.”

FDA does not endorse any specific type of RWD under the Guidance. Rather, it indicates that it will determine whether the quality of the RWD is sufficient to support a particular regulatory purpose. Purposes may include, but are not necessarily limited to: (1) clearance, approval, or reclassification; (2) broadened indications for use; (3) contribution to or supplementation of other types of pre- and post-market evidence; (4) post-market surveillance studies; (5) post-approval studies as a condition of device approval; and (6) alternatives to ethical dilemmas and other clinical trial challenges. The necessary degree of quality of RWD may vary based on the regulatory purpose and current stage of the Total Product Life Cycle (“TPLC”). For example, registry data may be insufficient to demonstrate substantial equivalence, but appropriate for post-market surveillance purposes under the criteria set forth in the Guidance.

The Guidance describes characteristics of RWD that FDA encourages sponsors to consider in selecting RWD sources and how the suitability of RWD may vary based on the regulatory purpose for which they will be used. Stakeholders are encouraged to consult with FDA during the pre-submission process to address and mitigate any concerns about the proposed use of RWD for a specific purpose or the development of a new RWD source.

The characteristics described in the Guidance include:

▸ Accuracy. RWD and sources should be accurate when compared to verifiable source documentation. To demonstrate accuracy, stakeholders may choose to compare treatment-specific database information with performance data that exist within the device—e.g., logs of the treatments delivered.
▸ Relevance. RWD and sources should be capable of addressing the pertinent regulatory need. FDA will review and assess the adequacy of RWD for the particular regulatory purpose based on a number of factors outlined in the Guidance.
▸ Reliability. RWD, sources, and resultant analyses should be reliable. FDA assesses reliability based on data accrual and quality control.

• Data Accrual. RWD sources should have an operational manual or other formal documentation that specifies the parameters of data collection (i.e., data elements to be collected, methods for data aggregation and documentation, and the relevant time windows for data element collection). To assess data accrual, FDA will consider a number of factors which are outlined in the Guidance. Note that EHRs and other RWD sources may demonstrate sufficient reliability even in circumstances where all of the accrual factors are not met.
• Quality Control. It is important to confirm that there are processes in place to assure data quality and integrity. FDA encourages stakeholders to follow published recommendations concerning registries, such as the International Medical Device Regulators Forum (“IMDRF”) Registry Working Group’s “Registry Essential Principles.” However, FDA also recognizes that certain types of databases—including EHRs—may not have the data quality control processes necessary to comply with these recommendations. To assess quality control, FDA will consider a number of factors that are outlined in the Guidance.

c. Determining when the collection and use of RWD requires an IDE

Under the Guidance, FDA also clarifies when the collection of RWD for a legally-marketed device requires an IDE. Generally, an approved IDE is necessary to lawfully distribute or ship a product that would otherwise be required to comply with all applicable FDCA requirements prior to commercial distribution for the purposes of conducting investigations of the device. An approved IDE exempts a device from the requirements of certain sections of the FDCA. IRB oversight and compliance with abbreviated IDE requirements apply to clinical investigations that involve the use or testing of an FDA-regulated, non-significant-risk device on human subjects, while FDA approval and compliance with a broader set of IDE requirements apply to clinical investigations involving significant-risk devices.

Historically, there has been significant ambiguity regarding the scope of the regulatory definition of “clinical investigation” and whether it would encompass any use of RWD—even the retrospective analysis of RWD generated in the normal course of treatment. As a result, the notion of leveraging RWD to support FDA submissions has generated substantial uncertainty among sponsors, IRBs, and entities such as health care providers and data aggregators that serve as potential RWD sources. Absent clarification from FDA, such stakeholders struggled with whether their collection and use of RWD would constitute a clinical investigation subject to IDE and good clinical practice (“GCP”) requirements.

Compounding this challenge is the fact that, until recently, FDA did not allow for IRB waiver of informed consent except under very limited circumstances, such that sponsors were effectively required to obtain informed consent for the retrospective use of RWD even if it would have been impracticable to obtain informed consent from the subjects of the data. For example, if a sponsor sought to use de-identified RWD to evaluate the safety and efficacy of a drug for an off label use, and the results of the analysis were intended for inclusion in a submission to FDA, the use of the de-identified RWD would not be clearly exempt from the informed consent and other human subject protection requirements under FDA’s prior regulatory paradigm. Further, IRB waiver of FDA’s informed consent requirement would not have been available even if it was impracticable—or even impossible—to obtain the informed consent of the individuals who were the subject of the data. This approach deviates from HIPAA and the Common Rule, which generally allow IRBs to waive HIPAA authorization and informed consent, respectively, for certain minimal-risk research that could not practicably be conducted without a waiver.

This regulatory hurdle is now likewise being addressed by FDA, which has begun the process of changing its requirements for informed consent due to directives under the Cures Act. Shortly before issuing the Guidance, FDA issued a guidance titled “IRB Waiver or Alteration of Informed Consent for Clinical Investigations Involving No More Than Minimal Risk to Human Subjects” (the “Waiver Guidance”). In the Waiver Guidance, FDA indicated that it intends to revise its regulations to allow for waiver of informed consent for certain minimal-risk clinical investigations such that FDA’s approach would more closely align with that under the Common Rule. In the meantime, FDA states in the Waiver Guidance that it does “not intend to object to an IRB approving a consent procedure” involving waiver or alteration of informed consent if the IRB finds the clinical investigation meets certain criteria, including that it “involves no more than minimal risk” and “could not practicably be carried out without the waiver or alteration.” Absent from the Waiver Guidance is any express indication of the significance of de-identification for purposes of satisfying such criteria and whether the use of de-identified data and biospecimens would constitute a clinical investigation if other elements of the definition of a “clinical investigation” are satisfied.

FDA’s efforts to modernize its approach toward IRB waiver and alteration of informed consent and the use of RWD have been closely followed by other offices and entities within the U.S. Department of Health and Human Services. The Secretary’s Advisory Committee on Human Research Protections (“SACHRP”) in particular has urged FDA to provide more clarity in this area and encouraged harmonization of FDA’s requirements with the Common Rule and HIPAA. SACHRP provided comments on the Draft Guidance, some of which are reflected in the Guidance.

On the heels of the Waiver Guidance and the added flexibility it provides in conducting clinical investigations, the Guidance seeks to articulate clearer standards for determining whether the collection of RWD requires an IDE (which is often required before initiating a clinical investigation). Under the Guidance, determining the need for an IDE is context- and fact-specific. Notably, the Guidance provides that if a device is used in the normal course of medical practice or routine clinical care under the authority of a health-care practitioner—even if the use is related to administration of a legally marketed device for an uncleared or unapproved use—an IDE is likely not required. Thus, generally, retrospective analyses of existing RWD involving the use of a device for uncleared or unapproved indications would not be subject to IDE regulations. If, however, the objective is to generate data on the safety and efficacy of the device and the process for gathering the data influences treatment decisions (e.g., prescribing one device over another or administering a device in a particular way), an IDE may be required. If a registry is created to determine the safety and effectiveness of a new indication of an otherwise approved device, IDE requirements would presumably apply when physicians are advised to treat only certain patients or when follow-up activities (e.g., additional diagnostic studies) are performed for the purposes of research.

As the collection and use of RWD is an evolving area, FDA encourages sponsors, IRBs, and other stakeholders to contact FDA for additional clarification and guidance. Further, although the Guidance provides some clarity on the parameters for IDE requirements, areas of uncertainty remain. For instance, in its comments to the Draft Guidance, SACHRP urged FDA to address certain common sources and types of RWD. According to SACHRP, whether the use of these types of RWD would constitute a “clinical investigation” is open to interpretation. The Guidance, however, does not explicitly address all of the areas of ambiguity that SACHRP raised, including for the following scenarios. In the absence of clarification from FDA, such scenarios should be analyzed according to the factors described above.
• De-identified versus identifiable data. The Guidance does not discuss whether the de-identified nature of an RWD source would be relevant to the IDE determination. Although the above-referenced Waiver Guidance may make it easier to use de-identified RWD in connection with a clinical investigation under an IDE if the clinical investigation is considered minimal risk under the criteria referenced in the Waiver Guidance, FDA does not address whether and how the de-identified nature of the RWD would be relevant for purposes of determining whether the activity would be regulated as a clinical investigation in the first instance.
• Commercial purpose. The Guidance does not reference whether the commercial intent or any potential for revenue based on the use of RWD should impact the analysis of whether the activity constitutes a clinical investigation. SACHRP notes that FDA has issued guidance indicating that the applicability of its investigational new drug application (“IND”) regulations do not turn on commercial intent, and recommends that FDA issue guidance that echoes this position for purposes of its IDE requirements.
• “All comers” registries versus product-specific registries. The Guidance does not explicitly exempt from IDE requirements the use of certain types of “all comers” registries (e.g., a registry that contains information about all individuals treated for a certain disease, irrespective of which device or other clinical intervention is used), versus a registry about an individual product.
• Prospective versus retrospective studies. Although it states that retrospective analyses of existing RWD “would generally not be subject to the IDE regulations,” the Guidance does not categorically exempt such activities from regulation as clinical investigations. In its comments on the Draft Guidance, SACHRP wrote that prospective studies are more likely to constitute clinical investigations and recommends that FDA consider aligning its interpretation with the Office for Human Research Protections’ (“OHRP”) interpretation of the retrospective exemption under 45 CFR §46.101(b)(4), which applies to research involving the collection or study of existing data, documents, records, pathological specimens, or diagnostic specimens, if these sources are publicly available or deidentified.
• On vs. off label. The Guidance does not directly distinguish between the on and off label use of a device for purposes of determining whether the activity constitutes a clinical investigation. However, existing FDA regulations and guidance specify that on label, prospective studies are generally exempt from IDE requirements. SACHRP noted that determining whether the use of a device is on or off label is an important consideration when analyzing whether a particular activity, such as retrospective or prospective record review, or an observational study, is a clinical investigation subject to IDE and other requirements.
Notwithstanding these questions, the Guidance provides welcome insight into factors and potential guardrails that stakeholders should consider for purposes of determining when the use of RWD would be subject to an IDE and human subject protection requirements. As such, the Guidance is helpful not only for the biotech industry, but also health care providers, data aggregators, and other organizations with large amounts of available data that may be used for such purposes, and who have long sought clarity on the representations and disclaimers that they can make regarding how such data may be used.

II. Considerations for stakeholders

With the issuance of the Guidance and expanding universe of available data that potentially holds valuable insights into medical products, stakeholders will need to grapple with questions that remain regarding the suitability of certain RWD for specific contemplated purposes, as well as various operational questions surrounding the collection, use, and sharing of RWD. A few high-level recommendations are presented below to assist readers to—at the very least—begin conversations about how to leverage RWD to drive business decisions and priorities.

III. Implications and the future of RWE

The Guidance featured in this article brings the industry some clarity surrounding the use of RWE, especially for medical devices. However, stakeholders await additional guidance, including the framework mandated by the Cures Act, to inform the use of RWE to support regulatory activities for drug products. On Sept. 19, 2017, Commissioner Gottlieb remarked that FDA intends to publish a guidance document that includes “consensus definitions” and “a detailed description of RWE and its potential applications for satisfying aspects of FDA’s pre- and post-market requirements.” Although the considerations within the Guidance may be generalizable to other FDA-regulated products, the additional guidance promised by the Commissioner is essential to harnessing the ever-growing amount of RWD across the drug life-cycle continuum. And, we may not need to wait long. FDA has issued several guidances on various other topics in the last few months.