Highlights
- Different regulatory paths for neuromodulatory devices in the United States and Europe have resulted in different availability of these devices around the world. (Marjenin, Tariah)
- The regulatory requirements for non-invasive neuromodulatory devices will vary depending on the perceived level of risk, with non-therapeutic neuromodulatory devices facing less stringent regulatory requirements. (Marjenin, Tariah)
- For sponsors seeking regulatory approval of devices in the United States, the FDA has encouraged the use of new tools, including innovative clinical trial designs, such as adaptive trials as well as modeling and simulation. (Connor)
- Many TMS studies use sham controls, but because TMS is intrinsically multisensory, it is difficult to blind the patient and therapist. (Pascual-Leone)
- Important and challenging aspects of clinical trials for neuromodulatory devices include the choice of control or comparison condition, determining dose−effect relationships, and patient heterogeneity. (Hamilton, Lisanby)
NOTE: The points in this list were made by the individual speakers identified above; they are not intended to reflect a consensus among workshop participants.
The FDA in the United States, and the European Commission (EC) in Europe, are charged with regulating neuromodulatory devices. Their different paths to device approval have resulted in varying availability in different countries. For example, while the Neuronetics Neuro-Star system for the treatment of depression received FDA approval in 2008, CE Mark approval1 was not obtained until 2012 (PR Newswire, 2012). Meanwhile, the Brainsway TMS system for the treatment of depression received CE Mark approval several years prior to receiving FDA clearance in 2013 (Brainsway, 2013). Also in 2013, the FDA approved the Cerena TMS system for the treatment of migraine headaches that are preceded by an aura. This device was approved through the de novo pathway as a Class II device, indicating that it is a low- to moderate-risk device not substantially equivalent to an already marketed device (FDA, 2013). The different classes of devices are described below. In 2014, the FDA also cleared eNeura’s SpringTMS device for migraine (PR Newswire, 2014a).
At the FDA, responsibility for regulating most non-invasive neurostimulatory devices lies primarily with the Neurostimulation Devices Branch in the Division of Neurological and Physical Medicine Devices at the Office of Device Evaluation. By contrast, EC has 28 member states, represented by “Competent Authorities,” equivalent to the FDA but for individual countries. Competent Authorities have enforcement powers and designate “Notified Bodies” to assess applications for CE marking, which indicates that a product complies with European Union (EU) regulations and may be marketed in that country.
The FDA defines devices as things that are “intended for use in the diagnosis of disease or other conditions, or in the cure, mitigation, treatment, or prevention of disease in man . . . intended to affect the structure or any function of the body of man and which does not achieve any of its primary intended purposes through chemical action” (FDA, 2015b). What that means, according to Tim Marjenin, chief of the Neurostimulation Devices Branch at the FDA, is that even if a product is non-therapeutic and non-diagnostic, it may still be regulated as a medical device. However, he emphasized the word “may,” noting that the
__________________
1See http://www.emergogroup.com/resources/europe-process-chart (accessed June 1, 2015).
decision on whether or not to regulate a device depends on many factors. Sponsors who wish to know whether their product is considered a “medical device” are advised to contact the Device Determination Group in the FDA Office of Compliance.
According to Ibim Tariah, technical director at British Standards Institution (BSI), the EC definition of a medical device is similar to that of the FDA, specifying that it is intended to be used in human beings for diagnosing, preventing, monitoring, or treating a disease and does not achieve its action by pharmacologic, immunologic, or metabolic means. Non-medical devices are not subject to EC regulations; however, Hannah Maslen noted that the European Commission has proposed a revision of the Medical Devices Directive that would include a new annex covering implantable and invasive devices with a non-medical purpose.
Regulators’ primary concern is ensuring that products reaching the market demonstrate both efficacy and safety. According to Marjenin, most neurostimulatory products are thus classified into one of three classes of devices based on the level of regulatory control needed to “provide a reasonable assurance of safety and effectiveness.”
- Class I: These devices are subject only to general controls. Excluded from this class are devices that are life supporting or life sustaining, or that are needed to prevent impairment of human health or that present a potential unreasonable risk of illness or injury.
- Class II: General controls alone are insufficient to provide a reasonable assurance of safety and effectiveness and there is sufficient information available to establish special controls. Devices in this class typically require submission of premarket notification 510(k) if the technology is similar to something already classified, or a new device regulation if a device is not comparable to something on the market. Clinical data are generally not needed for a 510(k) submission, but are typically needed for a de novo application.
- Class III: General controls are not sufficient, and not enough information is available to establish special controls. Devices considered Class III generally require premarket approval
- (PMA). Clinical data are always needed for a PMA (FDA, 2014).
Marjenin said non-invasive devices are generally not Class III, but they may be assigned to this class if there is insufficient evidence to establish special controls. Marjenin said that special controls can mean a number of things, such as postmarket surveillance or performance standards. In Europe, classification of devices is based on four risk classifications. Devices with higher risk classifications receive more scrutiny by the appropriate authorities, said Tariah. At the two highest levels, both notified bodies and competent authorities may be involved in regulatory approval.
In 2012, the FDA issued a Guidance on Factors to consider when making benefit-risk determinations in medical device premarket approval and de novo classifications (FDA, 2012). This Guidance describes the FDA’s approach to making benefit−risk determinations, including an evaluation of the possible and probable risks; the type, magnitude, and duration of benefit; the level of uncertainty; patients’ tolerance for risk and perception of benefit; and the availability of alternative treatments.
With regard to off-label use, neither the FDA nor the notified bodies in the EC regulate the practice of medicine, including off-label use of products, according to Marjenin and Tariah. Victor Krauthamer concurred, noting that his division would not necessarily investigate an off-label use unless it had some relevance to public health as a general question, such as if a safety concern popped up.
Marjenin also discussed the regulatory perspective on products that present a low risk to user safety and are intended to be used to maintain or encourage a general state of health or healthy activity. In January, 2015, the FDA issued a General Wellness Draft Guidance (FDA, 2015a) that outlines their proposed approach to evaluating such general wellness products. Importantly, said Marjenin, a product’s inclusion under this guidance does not mean it has been shown to be safe and effective, and not misbranded for its intended use.
Maslen noted that the regulatory approach to assessing risks and benefits for non-therapeutic devices, including enhancement devices, diverges from the approach used for medical devices. She suggested that more thought needs to go into how best to evaluate the risk−benefit profile of enhancement devices. Both objective and subjective benefits should be considered, along with the value of the enhancement effect on
the individual. For example, the objective benefit of an improvement in working memory might be very valuable to a mathematician, but of less value to an athlete; the subjective benefit of feeling sharper or more alert might have more value for an elderly person who is starting to worry about cognitive decline than for a healthy 20-year-old who does not have such concerns. Maslen suggested that effects labeled as treatment, because they are more fundamental and more universally valued, are more amenable to cost−benefit or risk−benefit analyses, compared to enhancement benefits, which are harder to weigh against risks. For this reason, she proposed that more consumer freedom is appropriate for enhancement devices.
Alvaro Pascual-Leone added that the neuroscience community may not currently have the expertise to address outcomes for specific indications, which is why it is especially important to listen to patients and clinicians. Marjenin concurred, noting that regulators typically ask sponsors to demonstrate clinical meaningfulness in particular patient populations. In addition, patients’ perception of risk and their perspective on risk−benefit trade-offs is an important part of risk−benefit considerations, said Marjenin.
In 2004, the FDA published the Critical Path Initiative (CPI), which recognized that the “tools used today to predict and evaluate product safety and efficacy are badly outdated.” Clinical trial design was one area they cited as needing reform to make possible smaller and smarter trials (FDA, 2004). Publication of the CPI led to the founding in 2005 of the Critical Path Institute, a public−private partnership comprising eight consortia established to develop tools in precompetitive space to accelerate drug development.
Jason Connor, an adaptive clinical trial designer with Berry Consultants, addressed issues related to clinical trials for neurostimulatory devices. Connor serves on the FDA’s Neurologic Devices Advisory Panel. He emphasized the importance of communicating with key stakeholders—including the FDA, clinicians, patients, and statisticians—early in the trial design process. The FDA, and particularly CDRH, has a wealth of experience in clinical trial design and can often give key clinical insights as well has help navigate the regulatory landscape, he said. Moreover, he said, the FDA is far more
innovative than most people realize, and he urged investigators to be creative and seek the help of statisticians to design a trial that will answer relevant clinical questions.
One of the innovations supported by the CPI was the use of modeling and simulation. Connor noted that these tools allow investigators to simulate different population groups, different levels of effectiveness of the device, different dose−response curves, and different study designs enabling trials to be tailored to a specific patient population where a benefit is most likely to be seen. Simulations also allow key stakeholders to imagine why the trial might fail and what could have been done differently, for example, by changing dose, selecting a different study population, etc. For example, many investigators overestimate treatment effects or underestimate variability and thus fail to reach their primary endpoints.
Adaptive designs may maximize the potential for success in a trial, Connor said. Adaptations can be made for a variety of parameters, including sample size, randomization, dose, etc. When you adapt according to prespecified rules, he said, you can overcome some obstacles in the trial and better understand the error rate. Adaptive randomization may be especially valuable in combination therapy and for testing treatments in a variety of subpopulations. Many questions arose regarding the regulatory requirements for combination therapies. Marjenin said the FDA has an Office of Combination Products that will work with the drug and device divisions at the FDA to advise sponsors on the appropriate design of combination trials.
Another important issue with regard to trial design for neurostimulation is the choice of control or comparison condition. The gold standard for pharmacologic studies is the placebo-controlled, double-blind study; yet neurostimulatory devices present challenges with regard to the control. Many studies use sham controls, and according to Sarah H. Lisanby, much work has been done to develop shams that accurately as possible simulate the effects of active TMS. A more powerful approach, she said, is to use an active control, where subjects receive active stimulation to a region of the brain not expected to exert the desired effect but are blinded to the expected outcome. However, with regard to sham controls, Pascual-Leone noted the difficulty of mimicking or masking the clicks and taps that accompany TMS. Some devices switch current direction, but for the most part do not completely blind the patient or therapist. He suggested that alternative trial designs may be needed with active as well as task controls.
Other challenges that sponsors must address when conducting clinical trials for neurostimulatory devices mirror those encountered for drug trials, but require different tools. For example, determining the appropriate dose−effect relationships and addressing recruitment and eligibility concerns, blinding issues, and patient heterogeneity are particularly challenging in trials of neurostimulation devices, said Roy Hamilton. Another issue is the washout period, which according to several participants differs with stimulation modalities and parameters (e.g., duration of stimulation), although very few studies have been done to address this question.
The question was raised about whether independent assessments are needed for different devices with similar mechanism, or if data can be pooled from studies of devices with a similar mechanism. John Reppas, director of public policy for the Neurotechnology Industry Organization, said there is no good way to establish dose equivalence among devices, making it difficult to pool results.
REGULATORY OVERSIGHT FOR OVER-THE-COUNTER USE
OTC products typically need human factors and usability testing to ensure that users can determine themselves whether they are appropriate candidates to use the product and that they can use the device properly. Marjenin emphasized that not all devices are appropriate for OTC use.
The Foc.us gamer headset is a prime example of a consumer device that is not regulated by the FDA. Marjenin said his office is typically aware of devices marketed to consumers and may be discussing it internally; however, they are bound by confidentiality not to discuss their deliberations publicly. More generally, he said that for any marketed product that seems, for all intents and purposes, as if it should be considered a medical device, there may be an ongoing action or correspondence, but this also would be confidential. Krauthamer added that the regulatory science group does not typically look at specific devices, but only general endpoint; examples include running animal and cell tissue experiments at the field strengths covered by a range of devices, but not specifically that device.
Hank Greely asked whether there is any way for consumers to find out if a product that is on the market has been approved or cleared by the FDA and considered to be a “wellness product” not requiring regulation,
or if it is still “in limbo.” Marjenin said that unless the FDA has weighed in on it, consumers would not be able to find out.
