Originally published August, 1996
Halyna Breslawec and Nancy Teague
FDA regulates all aspects of clinical trials for medical devices through enforcement of a variety of regulations. The investigational device exemption (IDE) regulation imposes responsibilities on all those involved in clinical investigations, including sponsors, monitors, clinical investigators, and institutional review boards (IRBs).1 Additional IRB responsibilities are outlined in the agency's institutional review board and informed consent regulations.2,3
In addition to the regulations themselves, a variety of FDA guidances also contain essential information about the agency's requirements and expectations with regard to clinical trials.4,5 Many of these requirements--particularly those related to designing, conducting, and monitoring a good clinical trial--have already been discussed in this series.6,7 The specific responsibilities of sponsors, investigators, and study monitors have also been spelled out previously.8 This article addresses how FDA enforces adherence to these requirements and responsibilities and how study sponsors can avoid becoming the subjects of agency enforcement actions.
COMPLYING WITH REQUIREMENTS
The need to comply with the study requirements should factor in to every study decision a sponsor makes. For example, when selecting a clinical investigator, the sponsor should consider a potential investigator's capabilities in the areas of recordkeeping and reporting. Furthermore, a sponsor should be prepared to dismiss an investigator when he or she is not fulfilling the responsibilities previously agreed upon.
Correspondence and Communications. Obtaining and maintaining FDA approval of a clinical trial under an IDE is an interactive process. The sponsor of the study interacts with reviewers in the Center for Devices and Radiological Health's (CDRH) Office of Device Evaluation (ODE), which is responsible for prompt review and action on IDE applications. Although FDA may communicate orally with study sponsors about an IDE, serious questions or concerns will always be documented in FDA correspondence to the sponsor. Such correspondence may include questions about supplemental applications proposing to change aspects of the study, comments and questions about progress or adverse reaction reports, or questions about other aspects of the study. Often, FDA asks a series of rather specific questions but does not candidly state its underlying concern. Care needs to be taken to read between the lines of such letters. FDA provides sponsors with a time frame for response and a description of what actions it may take if an acceptable response is not forthcoming.
Sponsors should respond promptly to such correspondence. In general, the less response time provided by FDA, the more serious its concern. The intensity of proposed action if no response is forthcoming is also a barometer of the agency's concern. Responses should address each item outlined by FDA, even if some appear duplicative. When a sponsor detects an underlying but unexpressed concern, it should be dealt with directly in addition to responding to the specific concerns expressed in the letter. Sponsors should not hesitate to call for clarification, particularly when they truly do not understand the nature of the agency's concerns.
Reporting Requirements. The IRB and IDE regulations impose reporting requirements on sponsors, investigators, and IRBs involved in clinical trials.9 Some reports, such as progress reports, are routine; others are required as a result of an event, such as an unanticipated adverse device effect, or a deviation from the investigational plan. Complying with reporting requirements forms the basis of good records, which FDA will review during an inspection.
Documenting Compliance. The IDE regulation outlines recordkeeping requirements for all those involved in clinical trials.10 Records must be maintained for at least two years following the end of the study or the date the study data are no longer needed for preparing a premarket notification (510(k)) or premarket approval (PMA) application. During inspections, FDA will compare records maintained by sponsors, monitors, investigators, and IRBs to evaluate their accuracy and authenticity. Maintaining complete and accurate records is the best preventive measure to avoid FDA enforcement action.
Prohibitions. In addition to outlining clinical trial responsibilities, the IDE regulation prohibits participants in clinical trials from performing certain actions. Commonly referred to as prohibitions, these actions are described below.11
Promotion or Test Marketing. Sponsors and clinical investigators may not promote the use of or test-market investigational devices during their investigations. The prohibition does not preclude advertising the availability of the investigational device for the purpose of obtaining clinical investigators or subjects for participation in the study, provided that certain conditions are met. FDA's rather dated guidance describes the appropriate content of such notices or presentations, and discusses certain acceptable and unacceptable activities.12
FDA permits sponsors to announce the availability of a device for testing only in publications or arenas that reach individuals who would possess the qualifications necessary to be an investigator, and the agency considers undirected mass mailings inappropriate. Advertising is permitted only until an appropriate number of investigators or subjects are enrolled. Any notice of availability must clearly state its purpose, with its content limited essentially to the name and address of the sponsor, how to apply to be an investigator, and a list of the investigator's responsibilities. Comparative or subjective comments about the device or its performance may not be included, and no claims of safety or effectiveness may be made. Furthermore, an investigational device statement that reads as follows must be prominently included: "Caution--Investigational device, limited by federal law to investigational use." Inclusion of pricing information is also considered inappropriate.
Commercialization. FDA considers commercialization to be the charging of a price for the device that is greater than necessary to recover costs of manufacture, research, development, and handling. The agency has not defined these recoverable costs.
The initial objective of this provision was to recognize that cost recovery for investigational medical devices was permissible and even necessary to encourage development of new devices. When the IDE regulation was issued, the idea of cost recovery for investigational products, especially drugs, was novel. This is no longer the case.
FDA enforces this provision by requiring sponsors to reveal the amount they intend to charge for the device. If the amount appears excessive, FDA will ask for a justification. This is usually the extent of the agency's enforcement in this area because FDA acknowledges that it has little interest and even less expertise in performing economic analyses of manufacturers' costs. However, it will take enforcement action when it becomes aware of blatant violations of this provision, such as the sale of regional distributorships to clinical
Perhaps the most confounding aspect of this provision is that to actually recover the costs of manufacture, research, development, and handling during the course of clinical trials (and not during presumed postmarketing sales), a manufacturer would need to charge an amount for the device that would greatly exceed the price of a comparable approved device. If a sponsor proposed a price that represented a true recovery cost, FDA would likely challenge that amount and ask for its justification.
The recent Medicare decision to cover the costs of certain investigational devices further illustrates the difficulty of setting prices for them.13,14 Medicare reimbursement for investigational devices designated as "nonexperimental, Category B" is based on and may not exceed the cost of comparable approved devices. While the price of an approved device is likely to include a profit, the total amount
is still likely to be less than an actual recovery cost.
FDA and Health Care Financing Administration (HCFA) cost decisions are made independently. The amount reimbursed by Medicare is established by HCFA without routine input from FDA and is intended to be the amount necessary for cost recovery but no more than the cost of approved devices. FDA evaluation of the price charged for an investigational device is done as part of the IDE review and does not include any consideration of reimbursement cost. As a result, the cost of an investigational device to a patient may be different than the amount reimbursed by Medicare, as well as from the true recovery cost. These factors contribute to the difficulty in justifying--and evaluating--a price to charge for such devices. Sponsors who select an amount within the general range of the cost of a comparable marketed device will probably not be challenged.
Unduly Prolonging an Investigation. Manufacturers may not prolong an investigation to avoid making a premarket submission (i.e., a 510(k) or PMA application) or to delay marketing approval for a product. A sponsor can be tempted to do this when studies have shown that the device is ineffective, when the device has been superseded by a new technology or has become obsolete, or when the cost of preparing an FDA submission would not be recovered through postapproval sales.
This provision has occasionally been interpreted as prohibiting continuation of an investigation once the data needed for product approval have been collected. However, FDA often finds itself in a position of wishing to permit, or even encourage, continued patient access to devices during the review period. This is especially the case when there has been public criticism or pressure to approve a product more rapidly. FDA has been creative in developing reasons that permit firms to continue investigations even when all necessary data have long been collected. The agency will take enforcement action, however, when a premarket application or responses to questions about the application are not forthcoming.
Representations of Safety or Effectiveness. FDA prohibits manufacturers from representing an investigational device as safe or effective for the purposes for which it is being tested, and prominent labeling of the investigational status of a device is required whenever the device is displayed. FDA has frequently enforced this requirement; fortunately, compliance is simple. More difficult to interpret and enforce is the applicability of this prohibition to discussions or published materials dealing with the safety or effectiveness of an investigational product. FDA recently developed new guidance regarding the dissemination of information about unapproved uses of approved products.15 While the guidance reinforces the agency's unwillingness to sanction the distribution of information about unapproved or investigational uses of a device, actual enforcement remains difficult except in blatant violative cases.
True enforcement of the requirements of a clinical trial necessitates the integration of all aspects of the study, the interaction of those involved in it, and their acknowledgment and acceptance of their
responsibilities. FDA's role, in effect, is to verify that study requirements were enforced during the course of the study. The agency's enforcement options are broad and include administrative actions such as withdrawal of IDE approval, as well as traditional enforcement
actions such as issuance of warning letters and legal action. Most of the agency's enforcement activities related to clinical trials occur after trial completion.
Withdrawal of IDE Approval. One of FDA's most effective enforcement tools is its ability to withdraw approval for an ongoing study. While such an action rarely occurs, the proposal to do so is invoked more frequently, particularly when a sponsor fails to respond to FDA correspondence or when the agency becomes concerned about safety aspects of a study. In correspondence, sponsors are asked to respond to concerns about ongoing trials by a certain date; if no response is made, the agency states that it may take action to withdraw approval of the IDE application. Grounds for withdrawing IDE approval include:
* The sponsor has not complied with the requirements of the IDE regulation, or any other applicable requirements.
* The application or any IDE report contains untrue statements or omits required material or information.
* The sponsor does not respond to a request for additional information within the time frame prescribed by FDA (this includes progress reports).
* There is reason to believe that the risks to subjects are not outweighed by benefits, that informed consent is inadequate, that the study is not scientifically sound, that the device used is ineffective, or that any part of the IDE or its conduct is inadequate.16
Although the grounds to withdraw approval are broad, FDA must follow certain procedures that offer the sponsor an opportunity to correct what the agency identifies as problems.
FDA must notify a sponsor in writing that it is proposing to withdraw approval of a study and must include a complete statement of the reasons for withdrawal. The agency must tell the sponsor that it has a right to request an informal hearing with FDA. If granted, the meeting--chaired by a hearing officer (an FDA employee)--gives the sponsor a chance to respond to the agency's concerns. The hearing officer then decides whether the IDE should be withdrawn. This decision constitutes final administrative action. The study can continue until this final action is taken.
If FDA determines that the continuation of testing under the IDE would result in an unreasonable risk to public health, however, it may immediately order withdrawal of IDE approval before any hearing. FDA rarely does this. When it does, it offers the sponsor the opportunity for a hearing on the reasons the IDE approval was withdrawn. The hearing occurs, however, after approval for the IDE has been withdrawn, after the sponsor has had to stop the study, and after IRBs have been notified of the study cessation.
FDA is usually willing to enter discussions with the sponsor regarding problems with the IDE outside of the hearing process, particularly when it believes the sponsor is motivated to take corrective action. The hearing process is time and resource intensive for FDA, and the agency's objective is usually to correct the problems with the IDE. In such a case, it is usually in the sponsor's interest to address the problems outside of the hearing process.
Nonsignificant-Risk (NSR) Studies. One area in which FDA enforcement tends to overstep its bounds is NSR device studies. Such studies are subject to abbreviated IDE requirements that include lesser recordkeeping and reporting requirements and approval procedures.17 An NSR study is considered to have an approved IDE when an IRB determines that the study presents no significant risk and then approves it. FDA approval for such an IDE is not required; in fact, the agency does not need to be notified of the approved IDE. FDA has published lists of device studies it considers to be significant-risk (SR) and NSR.18
Ways in which FDA may become aware of an ongoing NSR study include notification through a competitor, an IRB inspection, or an IRB report of an unanticipated adverse event or study noncompliance. FDA may have concerns about the study, may disagree with the IRB's determination of NSR, or may simply wish to know more about the study. In such cases, FDA sometimes oversteps its bounds by writing to the sponsor of the NSR study to demand that it be stopped and that an IDE application be submitted for review. FDA may not summarily demand that an NSR study be stopped unless it believes that continuation would result in an unreasonable risk to public health. If this is the case, the agency may immediately withdraw IDE approval. If it isn't, FDA must follow the same procedures as necessary to withdraw approval of an IDE for an SR device study; that is, it must propose withdrawal of IDE approval and offer the sponsor an opportunity for a hearing. When this occurs, the sponsor of the NSR study should remember that it has an approved IDE for the study, assuming that the NSR IDE approval was appropriately obtained and that FDA has not previously designated the study as an SR.
FDA may require an application for an NSR study, even while it is ongoing.19 FDA will, of course, review the submitted IDE application and may approve the study, with or without conditions. Although the regulations do not address this issue explicitly, disapproval of such an IDE is not an option for FDA, since such a disapproval would have practically the same effect as withdrawing approval of an approved IDE. Sponsors may, of course, voluntarily offer to suspend enrollment or otherwise curtail the study until FDA's concerns have been resolved.
Bioresearch Monitoring. FDA not only has various enforcement tools available to it, it has the discretion to apply them selectively. The agency bases its enforcement decisions on information obtained from inspections of sponsors, investigators, monitors, IRBs, and testing laboratories, and from review of information submitted to FDA as part of the IDE.
Inspections. Inspections related to clinical trials are conducted as part of FDA's bioresearch monitoring program. Although FDA has the authority to conduct unannounced inspections of anyone involved in a clinical trial, inspections are usually scheduled as a courtesy and in recognition of the impact caused by an unannounced inspection. This courtesy should not be misinterpreted as FDA's lack of interest in conducting thorough, sometimes lengthy, inspections and in inspecting all necessary records.
Unlike GMP and other routine FDA inspections, those related to clinical trials are almost always done for a specific reason. Most inspections are directed inspections, conducted to ensure that data submitted in PMA applications or 510(k)s are accurate and authentic. If the inspection shows that the study was conducted in compliance with the IDE, IRB, and informed consent regulations and that the data submitted to FDA are accurate and authentic, review of the application is permitted to continue. If violations of the IDE regulations or data integrity issues are raised, however, application review may be delayed. And in addition to having an impact on the review of the pending submission, enforcement actions may be forthcoming.
Bioresearch monitoring inspections may be assigned to FDA district offices by the Division of Bioresearch Monitoring (BiMo), a component of CDRH's Office of Compliance. This group assigns most inspections in order to verify data submitted in pending PMA applications under review by ODE. In fact, all original PMA applications must pass a BiMo inspection prior to their approval. Bioresearch monitoring inspections are conducted in the same way as other FDA inspections, with most inspections conducted by investigators from district offices. During inspections, FDA investigators review records to determine compliance with IDE, IRB, and informed consent regulations and to determine whether violations have occurred. Investigators also collect copies of documents to support observed compliance or noncompliance with the IDE regulations.
Upon leaving the inspection site, investigators may issue a form FDA-483--a list of inspectional observations. This form will include observations of activities the investigator believes deviate from FDA regulations. Generally, if no FDA-483 is issued at the completion of an inspection, no violations were observed.
Following the inspection, an establishment inspection report (EIR) is prepared by the investigator. This report contains the FDA-483, any records obtained during the inspection, and a recommendation for action. The EIR is sent to BiMo for review. BiMo staff investigates alleged IDE, IRB, and informed consent violations by reviewing inspection reports of the sponsor, monitor, investigators, and IRBs affiliated with a study. The BiMo group also coordinates involvement of ODE staff in supporting enforcement action. Enforcement decisions resulting from BiMo inspections are made by BiMo, often in consultation with ODE.
A few bioresearch monitoring inspections are "for cause" inspections to examine a specific concern or complaint about the study. Sponsors of clinical studies without pending PMAs should be concerned if anyone affiliated with the study is targeted for inspection.
In 1995, 315 bioresearch monitoring inspections involving medical device clinical investigations were conducted. Most of these inspections were directed data audit inspections of clinical investigators (62%), IRBs (25%), sponsors or monitors (10%), or laboratories (3%) involved in studies of products subject to pending PMAs. The rest of the inspections were "for cause" inspections resulting from FDA concerns regarding ongoing investigations or submitted data. In 1995, these inspections resulted in 40 warning letters issued by FDA.
FDA Enforcement Actions. For the clinical trial itself, the consequences of failing a bioresearch monitoring inspection are limited simply by virtue of timing; the trial is usually completed long before the inspection occurs. The consequences of inspection failure on those involved in the clinical trial are quite significant and can affect much more than their involvement in it.
Delay in Product Approval. If the inspection shows that the study was conducted in compliance with IDE, IRB, and informed consent regulations and that the data submitted to FDA are valid, application review will continue. If violations of these regulations are found or data integrity issues are raised, application review may be delayed.
Warning Letters (and Other Correspondence). Warning letters are sent by FDA to the inspected party to notify it of observations or conditions the agency believes violate any of the requirements of agency laws or regulations. Warning letters to those involved in clinical trials are always issued by CDRH rather than by local district offices. Warning letters identify the violation and request a prompt written response describing the planned corrective action. If no response to such a letter is forthcoming, further enforcement action can be expected. Warning letters, and company responses, are publicly available and can generate considerable adverse publicity.
As previously noted, FDA issued 40 warning letters based on BiMo inspections in 1995, compared to 48 in 1994 and 12 in 1993. The types of violations noted in the letters included failure to conduct the study according to the approved protocol, failure to maintain complete study records (such as records of protocol deviations), lack of IRB approval for protocol and informed consent changes, and informed consent problems.
FDA also issues what it calls "untitled" letters outlining observations the agency believes need corrections. While a lack of responsiveness to these letters will probably not generate immediate enforcement actions, corrective action should be taken to avoid future problems.
Investigator Disqualifications. FDA has proposed regulations on the disqualification of clinical investigators of medical devices.Similar rules for clinical investigators involved in drugs, biologics, and intraocular lens studies exist; their absence from the IDE regulation was a regulatory quirk. Disqualifying an investigator would prevent that investigator from participating in investigations of any FDA-regulated products and would permit the exclusion of tainted data from FDA consideration.
FDA is now reviewing comments on its proposed regulations and is revising them to be more consistent with the disqualification provisions for clinical investigators involved in FDA-regulated drug studies.21 Under the drug regulation, proceedings to disqualify a clinical investigator may be taken when FDA believes an investigator has repeatedly or deliberately failed to comply with the investigational new drug (IND), IRB, or informed consent regulations.21 The regulation also outlines the procedures for disqualification.
The impact of disqualification on an investigator would be significant: FDA will not permit him or her to participate in any FDA-regulated clinical trial. The impact of disqualification on the sponsor will be equally significant. Data from any study in which the disqualified investigator participated could be rejected for consideration in a PMA application or 510(k). Furthermore, if the investigator is participating in other trials at the time of the disqualification, those studies could be terminated, suspended, or otherwise constrained. Finally, the status of an already-cleared 510(k) or approved PMA based on studies conducted by the disqualified investigator could be affected.
IRB Impact. FDA considers IRBs to be its surrogates in patient protection. As a result, FDA usually deals with problems identified during an IRB inspection in an educational rather than an enforcement manner. While warning letters to IRBs have been issued in cases of serious violations, a letter outlining problems and suggesting solutions is usually sent.
When FDA finds egregious violations in the operations of an IRB, it may apply administrative enforcement actions against the IRB, including disqualifying it from reviewing any FDA-regulated research and preventing the approval of any application involving the violative IRB.22 FDA may also restrict, suspend, or terminate an IRB's use of certain expedited review procedures when necessary to protect the safety of study subjects.23
In 1995, at least two IRBs were suspended from approving new studies until corrective actions were in place.
Applications Integrity Policy. FDA's applications integrity policy permits it to defer review of an application when it suspects that data submitted in a pending application are fraudulent or unreliable.24 The review of any other pending application submitted by the company may also be delayed, pending resolution of FDA's concerns. Resolution of agency concerns is an elaborate process, involving subsequent investigations and audits of data and individuals involved in the study, and intensive FDA involvement. The sponsor has no administrative mechanisms to challenge being subjected to the policy. A number of sponsors have withdrawn their pending applications following their involvement with the applications integrity policy.
FDA has invoked this policy against a substantial number of firms that have had BiMo inspections. In 1995, four were subjected to it, and only one resolved FDA's concerns and consequently was no longer subject to the policy.
The impact of the policy on ongoing investigations is not as significant as on premarket applications. FDA will generally allow continued patient accrual in studies, although no new IDE applications will be reviewed.
Other FDA Action. Sponsors, monitors, clinical investigators, and IRBs are also subject to judicial actions, such as criminal prosecution and imposition of civil money penalties. Such FDA action is rare; however, FDA will move swiftly when egregious IRB violations are found. As always, FDA enforcement action can result in adverse publicity.