Potential FDA Regulation of Biofeedback
Sebastian 'Seb' Striefel, PhD, Logan, UT
Abstract : A number of ethical issues related to instrumentation exist. First, is the importance of buying equipment that is safe and effective. FDA approval is one way of ensuring that the equipment is safe and effective. Second, practitioners are responsible for maintaining the equipment that they buy and being competent in its use. Third, practitioners should be aware of the potential risks associated with allowing clients to take equipment home for home training and with the potential adverse reactions associated with biofeedback in general. Fourth, practitioners doing neurofeedback should be aware of the risks associated with not obtaining a pre and posttreatment QEEG. Last, those practitioners doing telebiofeedback and/or electronic billing should be familiar with relevant laws and procedures for ensuring client confidentiality .
Biofeedback equipment is classified by the Food and Drug Administration (FDA) as a Class II Medical Device (www.fda.gov/cdrh/dsma/dsmaclas.html). Medical devices have varying levels of risks and benefits and the degree of regulation is based on the level of control that the FDA considers being necessary to assure the safety and effectiveness of the device. There are three levels of classification. Class I devices have the lowest level of regulation because they present a minimal level of risk for harm. General controls such as registration, following the Good Manufacturing Practices, and labeling are considered sufficient for ensuring safety and effectiveness. Class II devices are those for which special controls are considered necessary by the FDA for assuring safety and effectiveness and where there are existing methods for providing such assurances. Special controls can include guidance documents, special labeling requirements (e.g. that QEEGs are not to be used to make diagnoses), mandatory performance standards, and post market surveillance (www.fda.gov/cdrh/dsma/dsmaclas.html). Class III devices require the most stringent regulation because insufficient information exists for assuring safety and effectiveness and these devices are generally those that support or sustain human life.
Biofeedback equipment manufactures are generally required to file a 510(k) Premarket Notification so that the FDA can determine if the equipment is 'substantially equivalent' to a legally marketed device that does not require premarket approval. 'Unless exempted from premarket notification requirements, persons may not market a new device, under section 510(K), unless they receive a substantial equivalence order from the FDA or an order reclassifying the device into Class I or Class II (section 513(I) of the Act).' (www.fda.gov/cdrh/modact/frclass2.html, p.2). There is now one exception, and that is for the selling of battery operated biofeedback equipment. The FDA has the regulatory authority to exempt a Class II device from the 510(k) requirement and has done so for battery operated biofeedback equipment, which it evidently considers to be safe and effective (www.fda.gov/cdrh/modact/frclass2.html). Other FDA requirements must still be met. The FDA rules are far too complex to discuss here. As such, readers are referred to the FDA website (www.fda.gov).
MANUFACTURER AND PRACTITIONER RESPONSIBILITY
So what does this mean for the practitioner? It means that if a piece of equipment has received Premarket Approval from the FDA, it is considered to be safe and effective for the purposes filed with the FDA. Biofeedback equipment (other than EEG equipment) has been approved for relaxation training and muscle reeducation. Biofeedback equipment that is to be marketed for other purposes requires a separate Premarket Notification approval and a prescription legend (www.fda.gov/cdrh/ode/143.html). Not all manufactures of equipment have gone through the FDA approval process which is relatively simply for devices that are generic in design and purpose to devices already approved. All biofeedback equipment manufactures are strongly encouraged to go through the FDA Premarket Notification/approval process and thus reduce their own legal risk, help ensure that the public is protected, and to help practitioners in ensuring that the equipment they purchase is in fact considered to be safe and effective by the FDA. Legally, the selling of a device classified by the FDA as a Class II Medical Device without meeting their requirements is a violation of law. Violation of existing law is also an ethical violation. At minimum, manufacturers should inform those who are considering purchasing a piece of equipment, of its FDA status, i.e., Premarket Notification filed and accepted (approved), Premarket Notification filed and pending, Substantial Equivalence Received, or Premarket Notification not yet filed.
Buy Approved Equipment
Practitioners are encouraged to ensure that the equipment that they buy has gone through the appropriate FDA process and has been approved by the FDA for the intended use(s). Why take an unnecessary risk when good equipment by vendors who have gone through the process is available? If a practitioner purchases a piece of biofeedback equipment that has not been approved by the FDA or exempted by them and it turns out that the equipment is defective in some way and not safe, e.g. a client gets an electrical shock because of faulty optical isolation, the practitioner shares in liability and could readily be considered negligent because he or she is ethically and legally responsible for what he or she does or fails to do.
Most, if not all, FDA approved, Class II medical devices are supposed to be sold only to licensed health care professionals. They are not supposed to be sold to the general public. To market biofeedback equipment to someone other than professionals require the filing of other paperwork with the FDA and their authorization (approval) that it is safe to do so.
Responsibilities after Purchase
Once purchased, practitioners are ethically responsible for ensuring that the equipment is properly maintained (e.g. no frayed electrical cords, periodic recalibration as appropriate, etc.), used in accordance existing ethical principles and practice guidelines and standards (e.g. AAPB, 1995; Striefel et al. 1999), and used in accordance with relevant law.
Practitioners are also expected to be competent in all aspects of the equipment that they use. This means at minimum, reading the manual and practicing with the equipment until it can be used competently before using the equipment with clients. Some equipment is either complex to use or the data obtained is complex to understand (e.g. that used for brain mapping, QEEG). As such, practitioners should obtain additional training, supervision, or consultation, as needed. If fact, initially it may be a good idea to have ongoing supervision or even have someone else complete complex tasks. For example, removing artifacts and interpreting QEEGs correctly requires considerable skill. Some groups offer such services for a fee. One useful training task is to have a competent external source conduct the EEG artifacting and interpretation of a protocol while one also does it oneself and then compare the results to see if they are the same or very different. The more similar the results, the more likely it is that the task has been completed correctly.
Should biofeedback equipment be sent home with clients so they can practice at home and thus speed up the treatment process? The answer to this question depends on a number of factors related to client risk. Fortunately, adverse reactions to biofeedback training are overall rare, and when they occur they are relatively transient or readily dealt with by competent practitioners (Hammond, 2001; Schwartz & Schwartz, 1995). How can a practitioner deal with an adverse reaction if the practitioner is not available? Relaxation induced anxiety is reported to occur in up to 40% of individuals getting some form of relaxation training and the anxiety may be a sign of underlying pathology (Smith, 1985). Theta training can reportedly induce both seizures and extremely traumatic material (Stockdale & Hoffman, 2001), or even depression (Nash, 2001). Theta training can occur as a result of direct training using EEG biofeedback or it can occur as a result of deep relaxation training via some other mode of biofeedback training.
Clearly, before home training is considered, a practitioner should have a good clinical history; should have ruled out a history of seizures, post-traumatic stress disorder, identity disorders, or other conditions that might result in predictable, adverse side effects; and should have conducted enough office-based sessions to minimize the risk of adverse reactions during home training. In addition, informed consent should have been obtained that indicates that adverse reactions do rarely occur and that if they occur, clients are to stop the home training and consult with the practitioner to determine the next step or steps in treatment. Sometimes, reassurance is all that is necessary, and other times, intervention requires that a practitioner have a good background in psychotherapy. Nash (2001) recommends that practitioners unlicensed in a mental health discipline, not provide treatment for DSM IV diagnostic conditions unless supervised by someone who is so licensed and thus competent to intervene if and when adverse reactions occur.
To Map or Not to Map
One special area related to instrumentation is concerned with whether or not a practitioner doing neurofeedback training should first obtain a brain map or QEEG. Several practitioners have reported a variety of adverse reactions that may not be detectable if one does not have a QEEG (Ayers, 2001; Hammond, 2001; Nash, 2001; Stockdale & Hoffman, 2001). With a pre and posttreatment QEEG, a practitioner may be able to ensure that the appropriate treatment protocol is selected, adverse reactions are detected, appropriate corrective action is taken (Ayers, 2001; Hammond, 2001; Nash, 2001; Stockdale & Hoffman, 2001). One might also be able to detect potential problems through good observation of clients and by obtaining regular feedback from the client. Based the limited current adverse reaction information, most of which is anecdotal, it is important for practitioners to take special care to do all that is possible to avoid harm to clients. At minimum, it seems important to obtain a client's informed consent concerning the potential risks of doing neurofeedback and to systematically collect data concerned with detecting any negative side effects. Whether or not to do a QEEG is at present a decision that each clinician must make based on the specific issues associated with the presenting client's problems and the belief system of that clinician. The more serious the client's presenting problem, the more important it is to take extra precautions and to consider the possibility that a QEEG or more careful and regular feedback from the client might be helpful.
Telebiofeedback and Electronic Data
A number of groups are now doing biofeedback training at great distances using telephone lines and/or satellite system connections (Striefel, 2000a). Such practitioners must ensure that they are in compliance with both the laws of the state in which they operate and in the state(s) in which the client who is receiving services is located (Striefel, 2000b). In addition, they must take extra precautions to ensure that client confidentiality is not compromised by inadvertently allowing unauthorized persons to access the session or clients' data from the session. Carefully documented informed consent is also important when doing such training, e.g. if one is hundreds of miles from the client and he or she has an adverse reaction, how will it be dealt within an ethical and legal manner? See Striefel (2000a & b) for more info. By the way, Medicare will now pay for some telehealth psychotherapy services (Ballie, 2001). It requires two-way, audio and video, real-time interactive communication between the patient and practitioner and the patient must be receiving the service in a health care setting, such as a clinic, hospital, or doctors office.
Those who use computerized storage of data or electronic billing must also take precautions to ensure that unauthorized access to confidential information does not occur and that one does not send the information to the wrong email address. Encrypting files, passwords, and storing data on discs that can be taken out of the computer and locked up in storage files are all part of the process for ensuring confidentiality of client information.
AAPB (1995). Ethical principles of applied psychophysiology and biofeedback . Wheat Ridge, CO: Author
Ayers, M. E. (2001). Question: Have you seen any negative effects associated with EEG neurofeedback ? Journal of Neurofeedback, 4 (4), 63-65.
Ballie, R. (2001. Medicare will now cover some telehealth psychotherapy services. Monitor on Psychology, 32 (10), 84-65.
Hammond, D. C. (2001). Adverse reactions and potential iatrogenic effects of neurofeedback training . Journal of Neurofeedback, 4 (4), 57-62.
Nash, J. (2001). Stockdale, S. & Hoffman, D. A. (2001). Question: Have you seen any negative effects associated with EEG neurofeedback ? Journal of Neurofeedback, 4 (4), 65-69.
Smith, J. C. (1985). Relaxation dynamics: Nine approaches to self-relaxation . Champaign, IL: Research Press.
Schwartz, M. S. & Schwartz, N. M. (1995). Problems with relaxation and biofeedback: Assisted relaxation and guidelines for management. In M. S. Schwartz (Ed.), Biofeedback: A practitioner's guide (pp. 288-300). New York: Guilford Press.
Stockdale, S. & Hoffman, D. A. (2001). Question: Have you seen any negative effects associated with EEG neurofeedback? Journal of Neurofeedback, 4 (4), 62-63.
Striefel, S. (2000a, Summer). Telehealth uses in biofeedback and applied psychophysiology. Biofeedback, 28 (2), 8-10.
Striefel, S. (2000b, Summer). Ethical and legal issues in telehealth. Biofeedback, 28 (2), 4-5, 10.
Striefel, S. Butler, F. Coxe, J. A. McKee, M. G. & Sherman, R. A. (1999). Practice guidelines and standards for providers of biofeedback and applied psychophysiological services . Wheat Ridge, CO: AAPB.
by Richard A. Sherman, PhD* Chair, Instrumentation Committee
This article summarizes my (a) understanding of the statutory relationships between biofeedback devices and the Food and Drug Administration's (FDA) requirements and (b) opinion of how this could potentially impact on the Association for Applied Psychophysiology and Biofeedback's (AAPB) membership. The information presented comes from my contacts with the FDA and from reading the regulations and literature they supplied.
It is critical that readers realize that this is a 'worst case' scenario and that the FDA is not likely to enforce every aspect of the law as they interpret it because they are already overwhelmed with work and the people I spoke with do not feel that biofeedback is highly likely to harm anyone. Thus, it remains a very low priority unless someone does something so far out of line so publicly and blatantly that they can't ignore it. Many of our members feel that these regulations are wrong while others feel that they do not apply to some or all aspects of the use of biofeedback devices. For example, when biofeedback devices are used for non-medical uses, such as education not involving clinical conditions, they probably do not fall under the FDA's regulations.
Our biofeedback device providers and manufacturers have seen this information and have had a chance to comment on it. Comments from this vital part of our 'biofeedback' community will follow this article.
The sale and use of biofeedback devices is governed by FDA regulations. The most relevant is 21 CFR 882.5050. The two main issues covered by the regulation which I feel we should be most concerned with relate to (a) what applications can biofeedback devices be advertised (labeled) as being used for and (b) who can buy and use biofeedback devices.
Labeling of Biofeedback Devices
In order to get approval to sell biofeedback devices designed after May 28, 1976, (the date the law regulating medical devices went into effect), a manufacturer has to demonstrate to the FDA that the device is essentially equivalent to one already approved or grandfathered in by having been marketed prior to May 28,1976. This is a pre-market notification performed under section 510K of the above regulation and has come to be referred to as a '510K.' Please note that the application includes an 'intended use' statement or literature which describes what the device is used for. The 'labeling' of a device includes all newsletters from the manufacturer, advertising, written material sent with a device, etc. The only intended use statement or 'label' for open sale devices I have been able to find with the help of several people from the FDA is one that states that biofeedback devices are approved for 'relaxation' (of the stress control type). Several manufacturers have had 'biofeedback,' 'muscular rehabilitation,' or 'treatment of fecal incontinence' accepted as the intended uses. No other claims have been approved and, thus, any other 'labeled' claims are illegal. Open sale biofeedback devices can not be legally advertised to 'cure' hyperactivity, urinary incontinence, etc. unless the manufacturer has filed the claim with the FDA.
A manufacturer does not get '510K' approval as an entity or for an entire line of products. In other words, because one device manufactured by a company has 510K approval, that does not mean others do. It also does not mean that because a manufacturer was in business selling biofeedback equipment prior to May 1976, that it has approval to continue selling all equipment. Some particular, individual items manufactured prior to May 1976 may be grandfathered in but 'first cousins' introduced afterwards are not. The definitions of equivalent and how much you can change an older device and still have it fit the grandfathering provision are up to the FDA. They can be picky but are subject to pressure from ombudsmen and potential law suits.
If a manufacturer wishes to have a different intended use for a product which either (a) already has 510K approval or (b) for which such approval is being applied for, the manufacturer must supply clinical efficacy studies which can convince an FDA scientific review panel of their validity. If they are convinced, a 'hybrid' 510K can be issued. Please note that because biofeedback devices are not used to support or sustain life, are not implants, and do not put the subject at 'unusual' risk, the studies do not have to have been pre-approved by the FDA prior to being performed to be considered by the panel. However, they do have to have been approved by a legally constituted Institutional Review Board (IRB). This is a time consuming and relatively complex action to attempt but is much easier than the following alternative. I feel that we clearly have the data which would support altering the intended use statements for several devices. This is especially true of devices marketed for EMG biofeedback for amelioration of tension headaches.
If a device is not essentially equivalent to an approved or grandfathered device, then 'pre-market approval' (PMA), as opposed to the pre-market notification discussed above, has to be applied for under section 515 of the Act. This is much tougher to get. Many biofeedback devices may not be the equivalent of devices marketed prior to 1976 and probably require this kind of approval. Much of our software could potentially fit into this category.
The starting point for approval of new devices is usually getting an 'Investigational Device Exemption' (IDE) from pre-market notification and pre-market approval so experimental devices can be supplied to researchers for clinical trials on safety and effectiveness (Section 520(g) of Title 21, FDA 90-4159). These are relatively easy to get but the device's use is very strictly regulated and clinical trials must frequently be pre-approved by the FDA before their data can be used to support an application for efficacy. Specific safeguards for humans who are subjects of investigations, maintenance of sound ethical standards, and procedures to assure development of reliable scientific data are required of each user. Studies must be approved by an Institutional Review Committee! Although manufacturers frequently fill out the paperwork for an IDE, the individual investigator is named in it and is responsible for it.
There is an exception to the requirement for getting an I DE. If, and only if, a legally constituted IRB (operating under the guidelines of part 56 of 520(g) and, usually, registered with the U.S. government) states that the device and study present no significant risk to the subjects, no IDE submission to the FDA is required. However, the definition of 'significant risk' includes the concepts of welfare and important treatment rather than being limited to physical damage. The relevant portion of the regulation is: 'Épresents a potential for serious risk to the health, safety, or welfare of subjects, is for a use of substantial importance in diagnosing, curing, mitigating, or treating disease or otherwise preventing impairment of human healthÉ' This provision should hold even though many biofeedback devices meet other requirements for exemption as most are noninvasive, do not by design or intention introduce energy into a subject (except GSR devices), and are not used as a diagnostic procedure without confirmation by another medically established diagnostic procedure. Thus, I believe that all biofeedback devices fall within the requirement for an IDE.
The IDE and IRB approval cannot be used as a cover to sell unapproved devices for supposed 'research.' I do not believe that any biofeedback device can be sold labeled as being 'for research only' (the required wording is 'CAUTION – Investigational Device. Limited by U.S. Law to Investigational Use') and meet the requirements of the law unless each purchaser has an IDE or, if a non-significant risk device, IRB approval priorto sale. Prolongation of the study and commercialization and promotion of an investigational device are strictly prohibited. In addition, 'the manufacturer must not charge the investigator more that the cost necessary to recover the costs of manufacturing, research, development, and handling.' In other words, no profits!
Sale and Use of Biofeedback Devices
There are two ways biofeedback devices can be sold. According to FDA representatives, they can be 'class two non-prescription devices' only if their single stated ('labeled') use is for relaxation (as in stress/anxiety reduction-not muscular relaxation for control of abnormal muscle tension).
They are 'class two prescription devices' if their stated ('labeled') uses are for or include any other purpose. If there is another stated purpose, they can not be used unless they are 'prescribed' by someone competent to do so. This can include licensed clinical psychologists, depending on state regulations. A device is classified as 'prescription' 'if adequate directions could not be prepared for the lay person' (21 CFR) to use the device. Thus, logically, they should not be sold to the general public. For example, Biosearch's fecal incontinence pressure biofeedback devices were described in the 510K as being for 'biofeedback training of anal sphincters for people with incontinence sold to licensed physicians and psychologists.
It is clear from the above information that most biofeedback devices logically fall into the 'prescription' category. Once the FDA takes a careful look at what we use biofeedback devices for, I fully expect the FDA to change the classifications of all but a very few devices to prescription.
The FDA has the power to stop the sale of improperly advertised devices. It can seize them from manufacturers and practitioners and/or fine the manufacturers if it wishes. As increased attention is drawn to biofeedback, the FDA is very likely to take a closer look. When they do, they could require major changes immediately.
FDA Regulation of Individual Practitioner's Use of Biofeedback Devices
Numerous members of the AAPB have asked what the ramifications of the above information are for individual therapists using biofeedback devices beyond their legal labels. The FDA definitely has the right to stop any practitioner from using any device it deems dangerous. When a representative of the FDA was asked to comment about this issue, he said that 'the FDA does not regulate the practice of medicine' but they do protect patients. I asked him the following questions and received the associated answers:
a. Can licensed practitioners use biofeedback devices for non FDA approved uses? This constitutes adulteration and is essentially a 'new intended use.' If approval has not been granted from an Institutional Review Board (IRB) and, frequently, an investigational device exemption (IDE) has not been granted the provider is at great risk of being sued by the patient if something goes wrong. Using the device outside of its label is similar to having a patient use a prescription drug for non-approved purposes. The patient should at least be informed that the device is being used outside of its label. The manufacturer should file either a new or amended/ hybrid 510K to have a substantiated use added to its label.
b. Can the FDA seize devices from practitioners using them for nonapproved uses/therapies and fine them for using them in this way? YES. Commissioner Kessler has just 'revisited' this issue for other types of devices. This is essentially an 'illegal use of a device' and the FDA could theoretically intervene. They do intervene if patients are in danger. The representative said that since biofeedback devices are unlikely to cause harm it is unlikely that they will intervene.
It is clear from these responses that most biofeedback practitioners are putting themselves at some risk by using devices beyond their legal labels. It is also clear that the FDA is not likely to intervene unless they feel that patients' safety is at risk. This does not mean that research can not continue or that workshops on specific uses can not be given. It does mean that practitioners need to use their devices as labeled if they wish to remain within the letter of the law.
I feel that the vast majority of our members are not knowledgeable enough of laws governing approval of devices for sale or intended use statements to make informed decisions about whether they are considering using a displayed or advertised product for legal use or even if the product can legally be sold for its intended use. Thus, it is my opinion that our association is ethically bound to require suppliers and manufacturers intending to display products at meetings sponsored in any way by the AAPB or to advertise in AAPB published materials of any kind to assure the AAPB that the individual device (a) is approved for sale by the FDA and (b) that all claims are included in the 510K 'intended use' statement or literature supplied with the application. Devices which do not have such approval but have an IDE or for which an adequate 510K has been applied for should be permitted to be displayed, but not sold, as long that they are clearly designated as such and do not make any efficacy claims. If a manufacturer can not meet this guideline, they should not be exhibiting or advertising the individual product or claim. The AAPB would make no attempt to police the assurances. I recommend that this policy be put into force beginning with the March 1996 meeting. This should give all of our manufacturers time to come into compliance with the policy. This policy should protect our members and help our manufacturers recognize the need to comply with the FDA's regulations.
There appears to be little need for practitioners to change what they are doing other than to inform their patients in appropriate circumstances that the device is not yet approved by the FDA for the intervention they are getting. Since most practitioners should already being giving their patients a document explaining their treatment, this statement could be added to it.
* The opinions and assertions contained in this manuscript are the private views of the author and are not to be construed as official or as reflecting the views of the United States Departments of Army or Defense.