Proposed Changes to the Device Intended Use RegulationApril 2, 2017
Not long ago, we recommended that the Food and Drug Administration (FDA) begin a new rulemaking to update the intended use regulation. We promised a blog post to get the ball rolling with suggestions for improvement. The time has arrived.
In the discussion below, the first section describes the shared role of the FDA and the medical profession in ensuring the safe and effective use of medical devices. The second section outlines the chief problems with the device intended use regulation. The final section explains the actual proposal and marks up the text of the device intended use regulation to show how the language would change. (A similar approach could be applied to the drug regulations, albeit there are technical differences.)
We welcome comments on this blog post by email to email@example.com. If a critical mass of comments is received, we will publish a round‑up, including our responses and potential modifications to the proposal based upon the comments. (To protect privacy and to encourage free discussion, all comments are guaranteed to be summarized or quoted anonymously without reference to name or place of employment.) We hope that, with the help of our readers, the result will be a well‑vetted proposal that harmonizes FDA’s regulations with the First Amendment.
Such a proposal will be well‑timed, because FDA put a hold on the recent final rule amending the intended use regulation until March 19, 2018, to allow time for more consideration (see our previous post here).
The Statute: FDA’s Role
FDA’s authority to conduct premarket review of medical devices comes from the Federal Food, Drug, and Cosmetic Act of 1938, as amended (FDCA). The FDCA empowers FDA to exclude dangerous devices from the market to prevent tragedies from occurring, rather than removing dangerous devices after injuries occur. Additionally, FDA is empowered to bar ineffective or sham devices from the market.
Accordingly, the sponsoring firm has the burden to convince FDA that a device is labeled for safe use and has demonstrated effectiveness. The marketing of a device without FDA’s grant of 510(k) clearance or premarket approval causes it to be adulterated or misbranded. (Some low risk devices are 510(k)‑exempt if the sponsoring firm self-determines that their device is within the exempt category.) It is unlawful to introduce an adulterated or misbranded device into interstate commerce.
To conduct premarket review, FDA has a corps of professionals who perform careful multi‑disciplinary scientific evaluations of device safety and effectiveness:
FDA assigns review teams and primary reviewers who specialize in that scientific discipline to review . . . the application [for clearance or approval] and to generate a written evaluation. FDA then integrates the conclusions from these separate review activities to determine the appropriate outcome for the application. [Public Health Interests and First Amendment Considerations Related to Manufacturer Communications Regarding Unapproved Uses of Approved or Cleared Medical Products, pp. 8‑9 (Jan. 2017).
FDA generally approves a device with labeling for a specific intended use. Yet, many devices have more than one potential use in diagnosing or treating patients. The FDCA might have made FDA the sole gatekeeper for these additional uses. In this world, the burden would be on the sponsoring firm to return to FDA and obtain a separate approval of each additional use. Medical professionals, in turn, would be barred from deploying a device for any use other than those in the FDA‑authorized labeling.
The FDCA does not go down this road. The medical profession may lawfully utilize the same device for unlabeled additional uses in the practice of medicine. In fact, the FDCA expressly provides that it is not to be construed to limit or interfere with ability of the medical practitioner to use or prescribe a device off‑label in the practice of medicine (Section 1006).
The medical profession has developed an ecosystem for the evaluation of device performance. There are academic studies, journal articles, disease‑focused societies, symposia, conferences, case studies, clinical guidelines, talks by key opinion leaders, and a host of other mechanisms for weighing clinical evidence and reaching conclusions about the appropriate use of device technology. In addition, the medical professional has the benefit of clinical experience and knowledge of the individual patients they are treating. State licensure and malpractice exposure impose constraints on the use of devices as well.
The FDCA still gives FDA an important role to play after a device is legally marketed. FDA continues to provide premarket review of significant modifications to the device technology or its labeling. If a firm wishes to bring an additional use on label, it generally must seek clearance or approval.
In its postmarket role, FDA’s mandatory and voluntary reporting systems for malfunctions and adverse events apply to both on label and off‑label uses. Thus, FDA has good visibility into postmarket safety and performance. In fact, the Center for Devices and Radiological Health has an entire division, the Division of Postmarket Surveillance and Biometrics, which is responsible for ensuring the continued safety and effectiveness of medical devices after they have reached the marketplace.
Because the agency is not the gatekeeper for unlabeled uses, the medical profession and payors play a role in determining whether FDA’s approval of new uses in the labeling is worthwhile. In some situations, medical professionals and/or payors may demand that an unlabeled use be incorporated in the labeling before they will adopt it. In other cases, they may not. Firms will respond accordingly. There is a benefit to having this market‑based influence on whether a firm brings a use on label. FDA’s system of multi‑disciplinary review is excellent for ensuring robust, science-based labeling. Nevertheless, it tends to be relatively slow, cumbersome and resource‑intensive. Not all clinical uses of device technology require it.
There is a certain logic and balance to this system. On the one hand, FDA’s expert review provides a sophisticated screen against unsafe or wholly ineffective devices. On the other hand, FDA’s control over device labeling cannot possibly serve as a basis for directing all potential uses of medical devices across fifty states and 320 million people. It would require prodigious resources even to try to do so. The effort would be doomed to failure anyway, because crucial knowledge will always remain dispersed. It would not be possible for FDA to oversee the development of labeling for devices suited to the nuances inherent in all possible clinical uses and patient populations. Instead, the use of devices would be rigidly limited to fit the procrustean bed of FDA‑authorized labeling emanating from Washington, D.C. No one should desire this outcome.
To achieve the full potential of device technology, then, it is undeniable that the medical profession, including those in the front line treating patients, must play a significant role in determining the uses of medical devices, and they do. The reality of shared responsibilities between FDA and the medical profession for device safety and effectiveness is entirely consonant with the text, structure and history of the FDCA.
Problems with the Intended Use Regulation
The intended use of a new medical device is initially determined during FDA’s premarket review. FDA typically rests this determination upon information within the four corners of a 510(k) submission or premarket approval application, especially the draft labeling.
Once the device is legally marketed, FDA is authorized by the “intended use regulation” to base a determination of intended use on a much wider data set. The regulation states:
The intent is determined by such persons’ expressions or may be shown by the circumstances surrounding the distribution of the article. This objective intent may, for example, be shown by labeling claims, advertising matter, or oral or written statements by such persons or their representatives. It may be shown by the circumstances that the article is, with the knowledge of such persons or their representatives, offered and used for a purpose for which it is neither labeled nor advertised. [21 C.F.R. 801.4]
If FDA determines that the “circumstances surrounding the distribution” create a new intended use for the device, then FDA imputes the new intended use to all shipped units, even if their labeling complies with the existing clearance or approval. According to FDA, the device is now in the same legal position as if it had shipped without any clearance or approval at all. The device thus defined by FDA is adulterated and/or misbranded under the statute. The firm has now, according to FDA, committed a crime and/or incurred civil liability.
Last year, two criminal cases illustrated the extraordinary sweep of the intended use regulation. In United States v. Facteau, the prosecution invoked the intended use regulation in closing argument:
You’re going to hear from the judge, and you can look at all of the circumstances surrounding the distribution of the device to figure out what would be the intended use of the device. You can look at how the device was designed, you can look at how it was priced, you can look at how it was sold and you can look at what they say internally and what they say externally. And this is critical. [Tr., p. 26-78, emphasis added.]
FDA had cleared the device design in multiple 510(k) submissions between 2006 and 2011. Yet, in 2016, the jury was urged to consider the same design as evidence of an illegal new intended use. The jury also was urged to consider unpublished internal firm discussions of an unlabeled use as evidence of the creation of an illegal new intended use.
In another case, United States v. Vascular Solutions, in a pretrial motion, ruling the court interpreted the intended use regulation to allow FDA to infer “objective intent” from statements or claims not made public:
[N]owhere does the [intended use] regulation state that such statements or claims cannot be used to show objective intent unless they were published to the marketplace. To see the absurdity of defendants’ argument, consider a hypothetical in which a medical device manufacturer sells device D, which is approved for use A but frequently prescribed by doctors for off-label use B. If the firm creates a bumper sticker with the words `I intend D to be used for B: Prescribe D for B Today,’ by defendants’ logic that poster is inadmissible evidence of subjective intent so long as it sits in his briefcase, but admissible evidence of objective intent once he sticks it on his car. The Court is not persuaded that there is a legally relevant distinction here; in either scenario, the defendant has manifested into the physical world `oral or written statements’ that may be weighed as evidence of objective intent. [United States v. Vascular Solutions, Inc., 181 F. Supp. 3d 342, 347 (W.D. Tex. 2016).]
Following the court’s logic, a new intended use for a marketed device could be inferred from a secret diary, even if no one else in the world has seen it, because it also falls in the category of “oral or written statements” that have been “manifested into the physical world.” Id. at 347. Thus, objective evidence of subjective intent apparently may play a role in determining objective intent. This turns “objective intent,” an oxymoronic concept to begin with, into a hall of mirrors.
These court decisions did not break new ground. FDA has always taken the position that the regulation provides it the freedom to draw inferences based upon a virtually unlimited variety of evidence. Just last year, FDA provided this gloss:
In determining a product’s intended use, the Agency may look to “any . . . relevant source,” including but not limited to the product’ s labeling, promotional claims, and advertising . . . . For example, FDA may take into account any claim or statement made by or on behalf of a manufacturer that explicitly or implicitly promotes a product for a particular use . . . .
To establish a product’ s intended use, FDA is not bound by the manufacturer or distributor’s subjective claims of intent, but rather can consider objective evidence, which may include a variety of direct and circumstantial evidence. Thus, FDA may also take into account any circumstances surrounding the distribution of the product or the context in which it is sold . . . . In the context of medical products, generally, circumstantial evidence often ensures that FDA is able to hold accountable firms that attempt to evade FDA medical product regulation by avoiding making express claims about their products. [80 Fed. Reg. 57756, 57757 (Sept. 25, 2015).]
The last sentence in this quotation addresses articles marketed without overt medical claims in order to evade FDA device regulation. It makes sense in such cases to look at all the circumstances surrounding distribution to ascertain the intended use.
It does not make sense in cases when a device is already legally marketed with FDA‑authorized labeling. There is a legal question about when additional uses for such devices should be treated as “intended uses” for regulatory purposes. Nevertheless, that implicates a different set of concerns compared to ferreting out hidden medical purposes for articles being marketed to evade regulation altogether. FDA frequently conflates the two cases to its own advantage when discussing the intended use regulation.
In so doing, FDA takes the position that promotion of an additional, unlabeled use for a cleared device creates a “new” intended use. FDA treats the new intended use as if it were the sole intended use while the cleared or approved intended use simply evaporates. Thus, FDA takes the position that if a firm engages in off‑label promotion of any kind, a violation springs into being even for devices shipped with FDA‑authorized labeling. FDA does not even concede that it must show that a purchaser saw the off‑label promotion prior to the sale.
The correct statutory analysis is that a device cleared for intended use X, and shipped with labeling matching that intended use, may be employed by a medical professional for additional use Y. It is true that use Y is not in the labeling and has not been evaluated by FDA. Yet, under the scheme of the FDCA, it is the medical professionals, not FDA, who are the gatekeepers as to this additional use. There is no illegality under the FDCA if a firm disseminates truthful and non‑misleading information to the medical profession about use Y outside of the labeling. The statute, in fact, is agnostic about whether a firm returns to FDA for an evaluation of use Y. There is no statutory requirement to do so. If a firm voluntarily seeks to add use Y to the labeling, FDA is authorized to grant the request if the statutory requirements are satisfied.
Nevertheless, FDA deploys the intended use regulation to threaten firms with criminal and civil liability for disseminating information about unlabeled uses. (There are certain narrow exceptions, such as responding to unsolicited requests and disseminating peer reviewed journal articles.) FDA justifies doing so as an “incentive” for firms to obtain bring additional uses on label.
This use of the intended use regulation appears to be unlawful under the Administrative Procedure Act, 5 U.S.C. § 706(2)(C) (agency action may be set aside if “in excess of statutory jurisdiction, authority, or limitations”). The statutory incentive for firms to seek supplemental clearance or approval is that it allows them to augment the device labeling with an additional use. FDA’s speech restrictions go beyond this quid pro quo to apply criminal and civil coercion not authorized by the FDCA. This additional measure of coercion is rooted in the intended use regulation. It is not rooted in the statute.
FDA is also interfering with the practice of medicine. It is incontrovertible that some off‑label uses are beneficial and some even represent the standard of care. As discussed, the medical profession has developed an ecosystem for the evaluation of device performance, both on label and off‑label, in order to make appropriate use of the technology. FDA’s evaluation (or lack thereof) certainly plays an important role in the decision‑making of the medical profession, but it is not always and everywhere determinative.
In the meantime, sponsoring firms often have significant knowledge about unlabeled uses and an economic incentive to disseminate the information. By suppressing this knowledge, FDA materially distorts the quality and quantity of information reaching the medical profession. For FDA, less information is better. But, under the statutory scheme, it is the medical profession that decides whether to employ legally marketed devices for unlabeled uses. FDA’s interference with the free flow of truthful and non‑misleading information pertinent to the lawful practice of medicine amounts to an interference with it.
Finally, as several courts have found, FDA’s application of the intended use regulation to stifle speech may conflict with the First Amendment by suppressing the free flow of truthful and non-misleading information to highly trained experts lawfully using devices off‑label. Moreover, it does so by targeting a class of “disfavored” speakers while other speakers may provide exactly the same information without sanction. The courts accept such content- and speaker-based restrictions only when narrowly tailored to directly advance a legitimate governmental interest. The courts have generally concluded that FDA’s use of the intended use regulation is not narrowly tailored and does not directly advance the FDA’s putative governmental interest in bringing additional uses on label.
One court has expressly found in a First Amendment context that FDA has a legitimate governmental interest in bringing all additional uses on label. Washington Legal Foundation v. Friedman, 13 F. Supp. 2d 51 (D.D.C. 1998). This conclusion was based upon misreading the FDCA to “require that manufacturers get all uses approved by the FDA.” Id. at 71. In fact, as discussed above, the FDCA is agnostic whether additional uses are brought on label, and it does not grant FDA authority to require it. It is difficult to see why FDA may claim a legitimate governmental interest in coercing an action that the statute makes voluntary. That is especially so in cases when the speech used for this coercion is not labeling, which is the locus of FDA’s statutory authority. FDA has never pointed to any provision in the FDCA that authorizes regulation of speech outside of labeling and restricted device advertising. (Restricted devices are Class III devices, hearing aids and analyte specific reagents.)
Proposal to Restore the Statutory Scheme
This section suggests specific changes to FDA’s device intended use regulation (21 C.F.R. § 801.4) as it stands after FDA’s recent amendment, although the amendment has been delayed for at least a year.
The easiest solution to the problems discussed above would be to restrict FDA to determining marketed intended use based upon a device’s labeling, just as it primarily does during premarket review. That way, FDA would simply make sure that the firm’s labeling in the marketplace matches the labeling that was cleared or approved. Advertising would not be part of the picture, because FDA does not have jurisdiction over most device advertising. Other speech would be excluded as completely outside the purview of the FDCA.
Unfortunately, FDA’s definition of labeling is so broad that it could easily encompass speech deserving of First Amendment protection. It is also so nebulous that it would still be too difficult for the regulated community to know in advance what communications will be the basis for determining intended use. Arguably, neither the breadth nor the nebulousness of FDA’s definition of labeling is faithful to the plain statutory language. Nonetheless, the dysfunction is deeply entrenched due to long‑ago court decisions and decades of administrative action. It grew up chiefly during the period from the 1930s to the 1960s, in the heat of a 30-year turf war between FDA and the FTC as to who would have jurisdiction over drug advertising. Origins of the Prohibition Against Off‑Label Promotion, 69 Food Drug L.J. 161 (2014).
Ideally, Congress would revisit the regulation of device (and drug) labeling and advertising and update the FDCA to make it a better fit for the age of electronic communications. That is too heavy a lift within the confines of this proposal. Our aim is simply to mitigate the worst excesses of FDA’s device intended use regulation. Two objectives underlie this proposal:
- Firms should have certainty about the labeling that will serve as the basis for an intended use determination. Providing this certainty will enable firms to resume their original statutory role as masters of their intended use. As the First Amendment requires, these firms can disseminate truthful and non‑misleading information about off‑label uses without fear of creating an unapproved new intended use.
- There is potential for FDA to apply an excessively broad definition of labeling when determining intended use, thereby chilling protected speech. Narrowing the subset of labeling that is the basis for intended use will mitigate this problem.
In light of these considerations, the specific proposal is as follows:
The preamble in the first sentence of the device intended use regulation is unchanged.
Subsection (a): This part of the intended use regulation continues to authorize FDA to determine intended use based upon a full consideration of the circumstances surrounding distribution of an article. This provision applies to any article not listed with FDA, i.e., the sponsoring firm has not conceded that device regulation applies. If FDA finds that the article actually has an intended use falling with the definition of a medical device, it may impose device regulation on that basis. Likewise, the existing approach to intended use determinations will continue for devices marketed over-the-counter (OTC) to lay users. The only change here is to remove from this approach the class of devices described in subsection (b).
Subsection (b): This new subsection governs the determination of intended use for articles meeting two requirements: (i) The article must be already listed with FDA as a legally marketed medical device; and (ii) the listed medical device must be sold for prescription use only by licensed medical practitioners under 21 C.F.R. § 801.109.
If requirements (i) and (ii) are both met, then FDA shall determine objective intended use exclusively from the labeling included per § 801.109(c) on or within the package from which the device is dispensed (or an electronic substitute as authorized under section 502(f) of the FDCA). FDA may not consider other labeling or advertising in making an intended use determination.
Subsection (b) does not alter FDA’s authority to require that all labeling and restricted device advertising be truthful and non-misleading in accordance with section 502(a) of the FDCA and various implementing regulations (e.g., 21 C.F.R. § 801.6). For the purpose of suppressing false and misleading labeling and advertising, these terms retain their current boundaries, even if overly expansive or nebulous.
Subsection (c): The first part is unchanged from the current regulation. The last sentence of subsection (c) was added by FDA in a final rule issued this year, although it will not be effective until 2018 at the earliest. The language we propose would ensure that the “totality of the evidence” ties back to the two different sources of evidence specified in (a) and (b). (An alternative would be to simply strike the last sentence, given that FDA says it merely reiterates existing requirements and therefore would seem to be unnecessary.)
Subsection (d): This provision would be added to establish that labeling or advertising stating or implying that a medical use has been cleared or approved by FDA (or is exempt), when that is not the case, is misleading and constitutes misbranding. The purpose of this provision is to require sponsors to be clear in labeling and advertising which uses have FDA’s imprimatur and which do not. This clarity may be important to the decision of medical professionals as to whether to employ the device for such use. The new provision is not necessarily a change in law, since the FDCA already prohibits false or misleading labeling and restricted device advertising, but it puts firms on notice and empowers FDA to focus on this particular issue.
Change to Sec. 801.109(c): Under the new provision in 801.4(b), a device’s intended use is determined based upon the labeling in or on the device package. Under this subsection, a device must ship with labeling that has adequate directions for all advertised uses. This provision would force the sponsor to include unapproved uses in the labeling shipped with the device, causing such uses to become “intended uses” under 801.4(b), and subjecting the sponsor to adulteration or misbranding charges due to lack of clearance or approval of the unapproved uses. This unwanted circularity is eliminated by removing this phrase.
Change to Sec. 801.109(d): This section refers to labeling “whether or not” on or within the device package, and so could be read to have the same circular effect as 801.109(c). A parallel change is made, striking the reference to advertised uses.
The following, then, is the mark up of the intended use regulation (new language in red italicized font, deletions in strike‑through):
Sec. 801.4 Meaning of intended uses.
The words intended uses or words of similar import in 801.5, 801.119, and 801.122 refer to the objective intent of the persons legally responsible for the labeling of devices.
(a) Except as provided in subsection (b), Tthe intent is determined by such persons’ expressions or may be shown by the circumstances surrounding the distribution of the article. This objective intent may, for example, be shown by labeling claims, advertising matter, or oral or written statements by such persons or their representatives. It may be shown by the circumstances that the article is, with the knowledge of such persons or their representatives, offered and used for a purpose for which it is neither labeled nor advertised.
(b) If a distributed article is listed as a device pursuant to 807.25 and purports to be sold for prescription use only pursuant to 809.109, then the intended use shall be determined from labeling on or within the package from which the device is to be dispensed (or equivalent labeling provided electronically pursuant to section 502(f) of the act). If there is no such labeling, then intended use shall be determined in accordance with subsection (a).
(c) The intended uses of an article may change after it has been introduced into interstate commerce by its manufacturer. If, for example, a packer, distributor, or seller intends an article for different uses than those intended by the person from whom he received the devices, such packer, distributor, or seller is required to supply adequate labeling in accordance with the new intended uses. And if the totality of the evidence in either (a) or (b), above, as applicable, establishes that a manufacturer objectively intends that a device introduced into interstate commerce by him is to be used for conditions, purposes, or uses other than ones for which it has been approved, cleared, granted marketing authorization, or is exempt from premarket notification requirements (if any), he is required, in accordance with section 502(f) of the Federal Food, Drug, and Cosmetic Act, or, as applicable, duly promulgated regulations exempting the device from the requirements of section 502(f)(1), to provide for such device adequate labeling that accords with such other intended uses.
(d) Any labeling or restricted device advertising that suggests a device has received agency clearance or approval (or is exempt) for a specific intended use, when that is not in fact the case, is misleading under section 201(n) of the Act and constitutes misbranding under section 502(a) or section 502(q)(1) as applicable.
Finally, 801.109(c) would have the following revision:
[A device is exempt from section 502(f) of the Act if, among other conditions:] Labeling on or within the package from which the device is to be dispensed bears information for use . . . under which practitioners licensed by law to administer the device can use the device safely and for the purpose for which it is intended, including all purposes for which it is advertised or represented. . . .
The same change would be made to 801.109(d), which parallels 801.109(c), except that it applies to labeling “whether or not on or within” a device package.
That is the proposal. We look forward to your comments (firstname.lastname@example.org).