FDA Releases Two Medical Device Draft GuidancesAugust 16, 2011
On August 15, 2011, FDA released two Draft Guidances: “Factors to Consider when Making Benefit-Risk Determinations in Medical Device Premarket Review” and “Design Considerations for Pivotal Clinical Investigations for Medical Devices.” Each is discussed in turn below.
Factors to Consider when Making Benefit-Risk Determinations in Medical Device Premarket Review
This Draft Guidance appears to be the first time FDA has articulated, in a formal manner, the factors that, theoretically, FDA considers in balancing the risks and benefits of a medical device, and how, ideally, that balance should affect the clearance or approval of a device.
The guidance states that it applies to devices subject to premarket approval (PMA), and, “in limited cases,” devices subject to premarket notification (510(k)) requirements. As noted in the Draft Guidance, under section 513(a) of the Federal Food, Drug and Cosmetic Act (FDC Act), FDA determines whether PMA applications provide a “reasonable assurance of safety and effectiveness” by “weighing any probable benefit to health from the use of the device against any probable risk of injury or illness from such use.” FDA may also utilize a similar analysis in the review of 510(k) devices when there are differences between the proposed device and the predicate device that may affect the safety and/or effectiveness of the proposed device.
The Draft Guidance splits the factors FDA may consider into three categories: effectiveness, safety, and other. For effectiveness, the consideration is the extent of the probable benefit(s), measured by taking into account the following factors:
- Type of benefit
- Magnitude of the benefit
- Probability of the patient experiencing a benefit
- Duration of the effect(s)
The safety considerations that help FDA determine the extent of the probable risk(s) and/or harm(s) include:
- Number, severity, and types of harmful events associated with use of the device: Device-related serious adverse events, Device-related non-serious adverse events, and Procedure-related or indirect harms
- Probability of a harmful event
- Duration of harmful events
- Risk from false-positive or false-negative for diagnostics
- Number of different types of harmful events that can potentially result from using the device and the severity of their aggregated effect.
The additional factors FDA may consider when weighing probable benefits and risks include:
- Uncertainty as to the benefits and risks. This may be due to a poorly designed clinical trial, or the generalizability of the trial results to the intended treatment and user population.
- Characterization of the disease. How does the disease affect patients? What is its natural history and progression?
- Patient tolerance for risk. In determining patient tolerance for risk, FDA may consider disease severity, disease chronicity, and the availability of alternative treatment/diagnostic options.
- Availability of alternative treatments or diagnostics. The Draft Guidance indicates that in considering alternative treatment options, FDA may even consider off-label uses of marketed products if there are no approved or cleared products for the intended condition and patient population. The Draft Guidance states: “[I]f a new device has a very small benefit and there is significant uncertainty about that benefit, we may still approve the product if there are no available alternative treatments or diagnostics and the risk profile is acceptable.”
- Risk mitigation, such as the inclusion of warnings in the labeling or restricting the indication to a more limited use.
- Novelty of technology. FDA recognizes that new technologies carry with them more uncertainty than established technologies, but also may offer previously unavailable advantages. As a result, FDA states that it “may approve a device with less benefit or more risk than would be generally tolerated for more established technologies, particularly where providers and patients have limited alternatives available, to facilitate patient access and encourage innovation.”
The factors put forth in this Draft Guidance are likely familiar to sponsors of premarket device submissions, although presented in a more systematic fashion. As always, of course, the question is less about what the Draft Guidance states, and more about how it will be implemented, e.g., how, exactly, the competing factors are weighed. In light of the ongoing debate over FDA’s impact on innovation, it is interesting that FDA explicitly considers novelty as a factor in favor of approval, in part to “encourage innovation.” While this Draft Guidance may helps sponsors better understand what information they may need to balance, for instance, uncertainty generated by a small clinical trial, it will perhaps never be possible to determine precisely what FDA will consider sufficient to outweigh the possible risks of a device. Hopefully this Draft Guidance is a reasonable starting point in analyzing the factors that should be assessed.
Design Considerations for Pivotal Clinical Investigations for Medical Devices
The most important lesson from this Draft Guidance does not appear to be the descriptions of the different types of clinical study designs, or even FDA’s assessment of the value of those designs, but rather a conclusion that is reached from reading between the lines: the concept of least burdensome is no more, and if a sponsor decides not to conduct a randomized, double-masked (blinded), controlled, parallel group clinical study, it better have a good reason for not doing so. Not only must it have a good reason, it must provide that reason to FDA: the Draft Guidance states that an IDE application “should include the details of the proposed study design and a rationale for the study design chosen, including an explanation of the alternate study designs considered and why those study designs were dismissed as inappropriate, impractical, or not possible.” This is at odds with the Guidance FDA issued in 2002 regarding the “least burdensome provisions,” in which FDA specifically stated: “If clinical data are needed, FDA and industry should consider alternatives to randomized, controlled clinical trials when potential bias associated with alternative controls can be addressed.” It is also at odds with the Draft Guidance on Benefit-Risk Determinations discussed above, which states that clinical testing may include not only randomized clinical trials, but “partially controlled studies, studies without matched controls, well-documented case histories conducted by qualified experts, reports of significant human experience, and testing on clinically derived human specimens.”
For those readers not familiar with the concept of the least burdensome provisions, here is a little background. There are two provisions in the FDC Act known as the “least burdensome provisions.” One of these provisions, section 513(a)(3)(D)(ii), generally applies to PMAs, and states that “[a]ny clinical data, including one or more well-controlled investigations, specified in writing by [FDA] for demonstrating a reasonable assurance of device effectiveness shall be specified as a result of a determination by [FDA] that such data are necessary to establish device effectiveness.” The second least burdensome provision, generally applicable to 510(k) submissions, states that, in requesting information to demonstrate that devices with different technological characteristics are substantially equivalent, FDA “shall only request information that is necessary to making substantial equivalent determinations. In making such a request, [FDA] shall consider the least burdensome means of demonstrating substantial equivalence and request information accordingly.”
Each of these provisions implies that a requirement to provide clinical data to demonstrate a reasonable assurance of safety and effectiveness, or substantial equivalence, must be based on a determination that such data are necessary to make the showing. In other words, if a showing of reasonable assurance or substantial equivalence may be made without clinical data, FDA should—or perhaps, must—permit the sponsor to provide other than clinical data to support its applications.
While there may be room for debate as to whether a showing of substantial equivalence between two products with different technological characteristics always requires clinical data, requiring clinical data for PMAs is not new or even particularly controversial. The type of study, however, has generally been determined based on the type of device and the endpoints desired. This Draft Guidance, while describing many different types of clinical study designs, doesn’t explicitly reject any particular design, but makes very clear that well-controlled, randomized, masked studies are the clear favorite, and plants a seed of doubt as to whether other studies will really ever be acceptable to FDA. The text of the document belies the assurance in the Draft Guidance stating that “the principles of study design discussed in this guidance are consistent with the principles discussed in the Least Burdensome Guidance, but expand upon them by discussing the considerations that may affect the level of evidence necessary to meet the standard for premarket approval or clearance.”
The specific clinical design study factors discussed in the Draft Guidance are not, in and of themselves, anything particularly surprising. The Draft Guidance breaks clinical studies into two broad categories: clinical outcome studies, for therapeutic and aesthetic devices, and diagnostic clinical performance studies, for diagnostic devices. For both types of studies, the Draft Guidance addresses some general principles in designing a clinical study: avoiding bias and variability in device performance; clear statement of study objectives; subject selection that adequately reflects the target population for the device; and selection of enrollment sites appropriate for the intended use of the device.
The Draft Guidance identifies the below elements in designing a clinical outcome study:
- Selection of appropriate endpoints. These should be pre-specified at the design stage of the pivotal clinical study, and should provide sufficient evidence to fully characterize the clinical effect of the device for the desired intended use.
- Randomization. The Draft Guidance recommends randomization of subjects so the groups are comparable at baseline prior to the intervention, and states that the inability to randomize may subject the study to “bias of unknown size and direction” which can in turn adversely impact the level of evidence provided by the study. At the same time, the Draft Guidance recognizes that there are situations in which randomization is impossible, difficult, or potentially inappropriate. In such a situation, the Draft Guidance recommends sponsors contact FDA to discuss concerns with randomization and determine an appropriate study design.
- Masking (blinding). The Draft Guidance notes that masking is important to reduce bias, because knowledge of treatment may affect the behavior of subjects or interpretation of clinical outcomes by investigators. The Draft Guidance suggests certain steps a sponsor may take to minimize bias where masking is not possible, such as masking subjects until after the procedure and drafting a script for clinical staff to use to standardize the follow-up questions asked of study participants.
- Controls in comparative clinical trials. The Draft Guidance discusses active intervention control, placebo control, no intervention control, subject as own control, and subject-level data on a parallel group (historical control), making clear that historical control is the least desirable, which has long been FDA’s policy.
- Placebo effect and other phenomenon. Because placebo devices may appear to demonstrate effectiveness, the Draft Guidance recommends use of a placebo to compare the investigational device to a therapy that is ineffective. If superiority to the placebo can be demonstrated, it can be inferred that that investigational device is effective.
- Non-comparative clinical outcome studies. The Draft Guidance states that study designs that do not use concurrent (or historical) controls are not well-controlled studies. These studies may include single-group study with objective performance criterion; single-group study with performance goals; observational studies or registries; meta-analysis; and literature summaries.
The Draft Guidance addressed the following with regard to diagnostic clinical performance studies:
- Consideration of intended use. The pivotal diagnostic clinical performance study must support the intended use of the diagnostic device, by evaluating what the device measures or detects; what the device reports; cell, tissue, organ, part, or system examined; specimen source(s), specimen type(s), and specimen matrix(-ces); how the device is used; when the device is used; by whom the device is used; for what; and on whom device is used.
- True status of the target condition. There generally must be an assessment of the true status of the target condition. If no gold standard exists to make this determination, an alternative type of assessment may be used.
- Study population for evaluation of diagnostic performance. Sites selected for investigational use should be representative of the types of sites where the device is intended to be used, and study subjects should represent the target condition spectrum.
- Study planning, subject selection, and specimen collection. For a prospective study, specify inclusion/exclusion criteria, method of subject recruitment and selection, testing protocol, and analysis methods to be used. Retrospective selection of previously archived specimens may introduce issues of bias, such as non-representation of the target population. Sponsors should consult with FDA to determine if available specimens or subject data are appropriate to support a diagnostic device’s intended use.
- Diagnostic clinical performance comparison studies. Comparative studies comparing the investigational device with an established device is only possible when a clinical reference standard is used, and sponsors designing such studies should consult with the appropriate review division at the design stage.
- Masking in diagnostic performance studies. Diagnostic device clinical studies may involve multiple evaluations and users/readers. The user of the investigational diagnostic device should not be aware of the result from the clinical reference standard or other diagnostic evaluation, and vice versa.
- Skill and behavior of persons interacting with the device. Protocols should account for variability in the performance of persons interacting with the device, because use of a diagnostic device may require certain levels or types of skills or knowledge.
This Draft Guidance may in fact help sponsors design their clinical studies because they now have additional insight into what FDA likes and dislikes, and how FDA suggests adjusting for bias or other complicating factors. Nevertheless, the unspoken conclusion of the Draft Guidance is clear—FDA will be requesting, or perhaps requiring, increasingly demanding study designs.