Food and Drug Administration Advisory Committees (1992)

The conduct of an advisory committee meeting involves many elements. This chapter considers the following: setting the advisory committee agenda, scheduling committee meetings, meeting preparation, the conduct of a meeting, and meeting follow-up.

Although some explicit policies guide FDA advisory committee operations, relatively few current policies are documented. One question that confronted the IOM committee, therefore, was to determine how much written policy guidance was needed to ensure effective performance. Although such guidance provides the basis for uniform practices, it may also introduce unnecessary and unwanted inflexibilities.

Just as there are differences among FDA's centers in the recruitment of members and the assignment of functions to committees, committee operations currently reveal substantial variation among and often within centers. This variation originates from differences in their statutory missions, histories, administrative cultures, the scientific and clinical field in question, and the personal habits of the relevant FDA officials.

Some variation among and within centers is justified, and the IOM committee wishes to avoid recommending inappropriate standardization in such cases. In general, however, substantial standardization of policies and procedures in advisory committee operations is both desirable and feasible, the benefits of which will accrue to the agency, the sponsors, and the general public.

The IOM committee deliberated at length about recommending that FDA adopt a policy of scheduling advisory committee meetings as long as one or even two years in advance. Meetings scheduled in this way would require the following associated deadlines: agency (and presumably sponsor) agreement to review an application at a scheduled meeting; timely sponsor submission of all data to go to the advisory committee; timely completion by the agency of its review; and on-schedule distribution of material to the advisory committee members.

The proffered referent in this case is the submission of a research grant proposal to the National Institutes of Health (NIH) by a specified date to ensure its review at a particular time. Included in this scheme is a decision rule that late applications are not reviewed until the next cycle. Although the committee recognizes that the NIH experience provides an imperfect comparison for the submission and evaluation of an application to the FDA, it believes that there is great merit in introducing some comparable discipline in the FDA review process.

At least three reservations about such a proposal have been expressed. First, scheduling conflicts with major professional society meetings could occur. This is the least serious problem and could be handled similarly to NIH procedures. (NIH schedules grant proposal cycles and study section meetings one or two years in advance and consults with the major professional societies in particular fields before doing so.)

Second, in informal discussions industry representatives responded favorably to advance scheduling provided the particular advisory committee met at least three times per year and on a regular basis. The "cost" of three or four months delay if a meeting deadline were missed under such circumstances was seen as tolerable. Indeed, working to deadlines elicited a generally favorable industry response. However, for those advisory committees that met only twice a year, failure to meet the associated deadlines would result in a slippage of six months; the industry representatives did not find this period acceptable.

The most serious reservation was voiced by FDA representatives, who expressed the view that establishing a certain date some 6 to 12 months in advance for the end of an FDA review would be very difficult. Furthermore, meeting such advance deadlines would impose a demand on scarce agency resources of medical reviewers, which would make it difficult for the agency to comply easily.

The IOM committee saw the benefits of advance scheduling as twofold: imposing greater discipline on the internal product evaluation process, and making it easier to schedule the time, and thus ensure the participation, of

busy advisory committee members. Currently, the scheduling of advisory committee meetings is affected by the availability of committee members, the length of FDA review times, deference to a sponsor's desire to submit data up to the last minute, and long-standing agency practice of scheduling meetings on an ad hoc basis.

Modification of current practice would require that the agency issue an explicit policy on advance scheduling, plan for an appropriate transition period, and carefully monitor the implementation of scheduling in the transition period. The IOM committee believes that advance scheduling would be justified as a means for making better use of advisory committees.

The criteria for determining the matters to be brought to an advisory committee vary from center to center. CDRH, for example, was obligated by law to bring all PMAs to an advisory committee until the Safe Medical Devices Act of 1990 gave it some limited discretion. CBER brings both product-related biologics development and intramural research issues to its advisory committees.

Only one center has a written policy. CDER recently clarified its general criteria in a September 1991 document,¹ that identified a range of matters that the agency might bring to an advisory committee. These include advice on the approvability of specific drugs, general drug development issues, issues pertaining to marketed drugs, and management of the new drug evaluation (NDE) program. We examine the components of this document below.

The first of these matters, advice on the approvability of specific drugs, usually pertains to a new chemical entity (NCE) but may also include a new indication for a marketed drug. Advice may be sought on the following aspects of a given application: adequacy of the clinical trial design and the conduct of studies to provide substantial evidence of effectiveness; adequacy

of the data supporting safety; adequacy of the data about dosing and scheduling; consideration of surrogate endpoints, as appropriate to the compound; the need for postmarketing surveillance or additional studies; the need for limiting indications to specific populations; the overall risk-benefit of a new agent; special labeling concerns; and switches of prescription drugs to over-the-counter (OTC) status. On occasion, the center may ask an advisory committee member to conduct a primary review of selected portions of a new drug application (NDA).

Issues of drug development that go beyond the evaluation of specific products on which advice may be sought include development guidelines for classes of drugs, discussion of clinical study design issues, and specific safety issues for particular drugs. The Cardio-Renal Drugs Advisory Committee, for example, held a two-day advisory committee meeting this past year, the first of which dealt with the question of dose-response measurement of angiotensin-converting enzyme inhibitors. In addition, the center may seek advisory committee counsel on marketed drugs when adverse drug reaction data emerge from surveillance, animal studies, or new clinical trials.

Product evaluation issues, broadly construed, receive the greatest attention in the CDER document on advisory committee agenda items. However, the document also suggests several subjects for advisory committee agendas that relate to the management of the NDE program. The first deals with the periodic review (usually annually) by an advisory committee of the pending NDAs and the major new indications of other drugs in the CDER pipeline. The center's emphasis is on those drugs that may have an important public health impact, whose development is unusually complex, or that are subject to great public scrutiny.

Second, the periodic review of "important products under development" involves using committees earlier in the product development process than the licensing stage. The Lasagna Committee report called for this kind of early involvement, especially for cancer and AIDS drugs. The FDA argues—correctly, in the judgment of the IOM committee—that defining early involvement as participation in the review of investigational new drugs (INDs) is a practical impossibility. The inventory of active INDs is quite large, a substantial number of new INDs are received each year, and the agency is required by law to assess the safety of a planned clinical study within 30 days of the IND's receipt (lack of response by the agency allows the sponsor to initiate the clinical trial). Thus, it is not feasible to routinely involve advisory committees in initial IND reviews.

Nevertheless, the impact of AIDS on drug development and evaluation has been to involve CDER more deeply in clinical trial protocols than has been true historically. A similar early involvement is taking place in oncological drug development.

The CDER statement explicitly contemplates that reviewing divisions will periodically review the IND ''portfolio'' with their advisory committees. The presumed benefits of such a review are guidance to the agency and a clearer sense of participation by advisory committee members. The factors limiting the pursuit of this policy include the resource costs to agency personnel and to advisory committee members, as well as the disclosure of early-stage proprietary information to an increased number of individuals.^*

The third innovation suggested in the document is the most far-reaching. It is that advisory committees consider the periodic analysis of priorities and resource allocation for management of IND applications, NDAs, abbreviated new drug applications (ANDAs), and NDA supplements.

Setting the agenda of an advisory committee meeting involves two stages: (1) formally scheduling a meeting and publishing an announcement of that meeting in the Federal Register, and (2) a few days before the meeting, sending committee members a detailed agenda with specific questions on which advice is sought.

The first stage of this process is initiated by the Federal Register announcement, which must be published at least 15 days before a meeting is held. Publication lead time requires that an announcement be submitted by the center about six weeks before a meeting. The announcement sometimes includes a general description of the agenda, for example, the

specific NDA of a given sponsor (identified by number) and the general topics of the meeting.^*

Members of the IOM committee who serve on FDA advisory committees noted that they seldom see the Federal Register announcement.

The general questions that the FDA must consider in assessing the safety of drugs and biologics are whether the risks of a compound are outweighed by its benefits and whether there is "substantial evidence" from well-controlled trials to support the claims of effectiveness. It would help the review process if advisory committee members were regularly reminded of these decision criteria as they review a sponsor's data.

The second stage-setting the detailed agenda of committee meetings—involves establishing the specific meeting topics and time allocations and preparing the specific questions that the advisory committee will consider. Of these steps, preparing the questions is the most important.

The FDA is primarily responsible for determining these questions. It has the statutory responsibility to review and approve applications, and convenes committees to assist it in that process. As a practical matter, this is the only feasible way to proceed. The division director, in consultation with the office director, usually develops the specific questions. FDA preparation, review,

and approval of questions may take several weeks; they are often sent to committee members just a few days before a meeting.

Three issues have been raised about this process. First, some observers believe that the advisory committee should set its own agenda. (The Lasagna Committee comes close to recommending this.) If advisory committees were adjudicatory bodies responsible for weighing both the sponsor's data and analysis and the FDA's critique, and then rendering a judgment, an argument could be made that they should have control over their own agendas. Because the committees are advisory to the FDA, however, and are convened to assist it in the administrative review of drugs, biologics, and medical devices, it is logical to argue that the agency should develop the committee agenda around the matters on which it wishes advice.

The other two issues are of greater concern to the IOM committee. A recurring criticism of FDA's behavior toward advisory committees is that agency officials—typically the reviewing division—seek to control or influence, or even manipulate, a committee to achieve an outcome that they desire. This charge of "undue influence" is often made about the teleology of the questions posed to a committee; that is, they appear to some observers to be phrased or ordered so as to lead the committee to a conclusion that reflects the preference of the division in the matter. The other complaint, which is closely related, is that advisory committee members seldom have an opportunity to modify the questions prepared by the agency or to add others that they wish to consider.

These issues can be addressed together. It is seldom feasible to involve the committee deeply in advance consultation on questions because of time constraints and the lack of familiarity of members with the specific issues presented by an application. Two actions are possible, however, both of which would improve the process and mitigate the charge of undue influence. First, the questions, which are often developed solely by FDA staff, could be prepared in consultation with the committee chair. Second, the agency could and should inform committee members that they have a right to modify agency questions or add questions of their own.

One issue raised by the Industry Liaison Panel is whether FDA questions should be restricted to an application's scientific and clinical matters or whether they should also extend to the regulatory questions that the FDA must face. The panel's report stated the following: "In cases where a drug or biologic marketing application is under consideration by a committee, the FDA should not ask the committee to advise it on whether or not the application should be approved but, rather, on whether substantial evidence of safety and effectiveness has been provided." The panel recognized that this question was usually but not always asked of drugs and biologics committees, and recommended its uniform use. It also acknowledged that the CDRH interprets its statutory authority as requiring that the advisory panel be specifically asked whether an application should be approved.

The agency decision to approve a drug or biologic is based on two criteria: whether there is "substantial evidence" (consisting of adequate and well-controlled trials) to support the claims of effectiveness; and whether the risks of a product have been shown to be outweighed by its benefits. Given these criteria, Dr. Robert Temple, Director of the CDER's Office of Drug Evaluation I, commented on the panel's point about the distinction between the scientific and regulatory questions. "Once a committee has said there is substantial evidence of effectiveness derived from adequate and well-controlled studies and that the benefits outweigh the risks," he wrote, "it is being unduly coy to suggest that we should not ask whether the committee recommends approval."² The IOM committee concurs with Temple regarding the "distinction without a difference."

A related issue is the charge that the FDA sometimes asks "loaded" or leading questions. It is necessary to distinguish here between the tone and objectivity of questions and the fact that the asking of particular questions will indicate the problems that the FDA has with an application.

A major recurrent complaint from advisory committee members is that the FDA often fails to distribute materials sufficiently in advance of a committee meeting to permit their careful review by members. Delays in the distribution of materials are attributed primarily to limited personnel and administrative resources, the natural tendency of reviews to get done at the last minute, and the tolerance of such practices by the agency. A more sinister charge is that such delays are part of a deliberate effort by agency officials to manipulate the work of advisory committees.

There is a virtual consensus that the effective use of advisory committee members requires that they have review materials in their possession for a reasonable period of time before a meeting, preferably, for at least three weeks.

The IOM committee believes that the responsibility for fulfilling this recommendation rests not only with committee executive secretaries, but also with the directors and the application reviewers of the appropriate division. It also believes that advance scheduling of committee meetings should facilitate compliance with this recommendation.

Advisory committee members complain that often the format of materials they receive from a sponsor is not conducive to a careful review of the data. Sometimes this complaint is accompanied by a request that the FDA "repackage" sponsor materials to facilitate review. The agency does not wish to be responsible for the presentation of a sponsor's work, since this could be cited as a means of influencing a committee's analysis. A major deficiency could be remedied, however, through the use of concise (20–25 pages), complete, and integrated summaries of the sponsor's application and the agency's review.

The CDRH assigns primary review responsibility for a particular PMA to one advisory committee member, mainly to obtain a clinical evaluation of the application. The IOM committee believes that this practice also ensures a more thoughtful committee discussion and distributes the work load within the committee. In addition, this practice has great utility in those situations in which the match between committee expertise and a particular agenda item may be weak. (See the "custom tailoring" discussion below.)

Five types of communication before an advisory committee meeting deserve attention: FDA communication to committee members; communication among members; communication between sponsors and members; FDA communication to sponsors; and FDA communication to the public.

Advance communication by FDA officials with advisory committee members before a meeting has generally been limited to one member at a time. The impression gained by the IOM committee during its study, based on discussions with FDA staff, was that agency personnel believed that they could not discuss any substantive issue with more than a single advisory committee member at a time without violating the Federal Advisory Committee Act (FACA). This impression was reinforced by the IOM committee members who were or had been FDA advisory committee members.

However, when asked by the IOM committee about the ground rules for communications from FDA personnel to committee members, the Chief Counsel to the FDA responded in this way:

The practices of some FDA centers, she suggested, may be stricter than necessary to reduce any legal risk.

One caveat suggested by the Chief Counsel was that the Office of Government Ethics (OGE) may hold the opinion that review of materials by an advisory committee member before a committee meeting may constitute participation in "a particular matter" and thus require screening for conflict of interest and, if necessary, the issuance of a waiver before it takes place. "This is not," she wrote, "current practice at FDA." The extension of the OGE view to FDA communication with some committee members might require conflict screening and the issuance of a waiver, even for discussions that are preliminary. Several attorneys who communicated with the IOM study contended that any agency communications with advisory committee members before a meeting should be regarded as ex parte communications and that procedural guidelines should be established to restrict them. The law does not treat agency communications as ex parte; thus, procedural guidelines of the kind envisioned are not required. The need for such guidelines turns on the conceptualization of advisory committees: whether they are adjudicatory bodies hearing the presentations of two contending parties or advisory adjuncts to the administrative process. If the former, detailed guidelines for communication between agency and advisory committee members before a meeting may be appropriate. If the latter, the need for such guidelines becomes less compelling. The IOM committee strongly believes that the latter definition applies to the advisory committees used by the FDA. Thus, the current level of concern about interactions between FDA staff and advisory committee members may be misplaced.

The FACA has also been interpreted by some FDA officials, and in consequence by many advisory committee members, as precluding telephone or face-to-face communication between two or more committee members in the period after receipt of the materials to be reviewed at a forthcoming meeting but before the meeting itself. This presumed limitation has been interpreted to prohibit, for example, consultation by a clinician committee member with a statistician member on matters of clinical trial design, data analysis, or data interpretation. The rationale for such a prohibition stems from a desire to preclude a minority of any advisory committee from establishing a position before a meeting and exerting influence favoring that position in committee discussions.

The presumed limitation unnecessarily restricts discussion among committee members and is not required by the FACA.^* Again, the FDA Chief Counsel has written that "that preliminary discussions among committee members do not violate FACA." In short, staff instructions against such consultations may seem overly cautious.

The issue of FDA communication to advisory committee members and communication among members is poorly understood both within FDA and outside the agency. It is very important that the agency clarify the legal issues governing such communication and provide appropriate written

guidance to FDA staff, all advisory committee members, and other interested parties.

As a matter of FDA policy, sponsors are discouraged from communicating directly or indirectly—save through the agency—with advisory committee members before a meeting. The agency informs sponsors and committee members of this stricture. This policy is designed, in general, to protect the independence of the committee from lobbying by sponsors.

The FDA is governed by the Freedom of Information Act, (21 CFR 20), and federal confidentiality laws with regard to the public disclosure of materials it provides to advisory committee members. The agency is not obligated to share with sponsors, or the general public, all of the information provided to advisory committees. "Draft questions, proposed agendas, and FDA staff analyses" are exempt from the public disclosure requirements of 5 USC 552(b)(5).

The FDA takes the view that it is not obligated to share with sponsors its communications to advisory committee members in advance of a meeting. The IOM committee believes, however, that it is appropriate for the FDA to provide sponsors with copies of all information that it sends to advisory committees. This facilitates the preparation by the sponsor of its response to agency questions.

Regarding the public release of the questions prepared for the advisory committee, the general practice of the FDA has been to make them available to the public on the morning of a meeting. The IOM committee agrees with this practice and does not recommend earlier release to the public. The report by Kutak, Rock & Campbell, which dealt with FDA's handling of financially sensitive information, basically concurred that FDA release of the

questions to the public on the morning of a committee meeting was sound practice.⁴

The IOM committee did not examine at any length the questions regarding FDA advisory committees and the effect of their management, including the time of the public release of the questions prepared for committees, on trading in the securities markets. The FDA has before it the Kutak Rock report on financially sensitive information and this IOM report on advisory committees and must address the implications of where these two reports intersect and make the appropriate policy determination.

The primary objective of an advisory committee meeting should be to facilitate the independent, thorough deliberations of committee members. In this context, independence means freedom from influence by the sponsor of the product under consideration, by any other interested parties, and by the agency itself. To provide this independence, a secondary objective should be to minimize the opportunities for the FDA, or other parties, to exert undue influence, or to appear to do so, over committee deliberations. The discussion and recommendations of this section are directed toward achieving these objectives.

There are invariably four principal participants in the typical advisory committee meeting: the committee members, the chair, the FDA professional staff, and the sponsor of an application. Consultants also function in significant ways in many committee deliberations.

The role of advisory committee members is to provide independent, expert scientific advice to the agency by responding to specific questions about study design or methodology, adequacy of data, and assessment and interpretation of risks and effectiveness that have been identified by the FDA professional staff. The ability of committee members to carry out such a role is facilitated by their expertise, the provisions for advance preparation discussed above, and the "rules of the game" for committee meetings, which are discussed here.

The role of the advisory committee chair is critical to the effective performance of a committee. The chair should control agenda time efficiently; protect committee discussion time; ensure, in consultation with agency staff, that the meeting arrangements facilitate committee deliberations; regulate, as necessary, media coverage of meetings; and ensure that

committee deliberations are brought to closure by providing clear advice to the agency on the questions asked of the committee.

FDA staff are involved in the substantive evaluation of an application and in the organization of an advisory committee. Those who review an application have the responsibility to evaluate the completeness, adequacy, and relevance of the data; the analyses of and conclusions drawn from the data; and the completeness of the information presented in the request for approval. When necessary, FDA staff may arrange for a consultant to carry out further analyses of the data or assessments of their sufficiency.

In CDER and CBER, the executive secretaries are primarily responsible for the organization and logistics of the advisory committee meeting. In CDRH, they play a somewhat different role that derives from their dual responsibilities as managers of the substantive review and as administrative support to the committee. In their former capacity, they participate in the deliberations in a manner similar to division directors in CDER and CBER; in their latter capacity, they perform similarly to the executive secretaries in the other two centers.

Product sponsors are the final significant group that participates in an advisory committee meeting. The sponsors usually make a presentation of the data on their product in preparation for an argument that the product should be approved. The sponsors may use consultants to present these data, as well as individual patients or practitioners to testify on their behalf.

Two phenomena are regularly mentioned by advisory committee members as impeding committee deliberations. One is that sponsors' presentations often absorb more time than is initially allocated by the agenda. The possible reasons for this are several: the initial time allocation may not be realistic, which calls for prior consultation between agency staff and the chair; the members may wish to question the sponsor at greater length than anticipated, which suggests the need to poll members in advance for their questions so that the concerns of several members can be aggregated; or the sponsor may consciously present more data than the allocated time allows, which calls for a prior commitment from meeting participants to adhere to the schedule. Similarly, FDA staff presentations may exceed allocated time and thus impede committee deliberations as well.

The principal reason for exercising strict control over the agenda is to protect committee discussion time, which is vulnerable to erosion by the factors listed above. In addition, because committee discussion is often scheduled at the end of a meeting it may lack the participation of all members, especially those from the West Coast, who begin leaving for the airport in the late afternoon.

Television news networks, with cameras, kleig lights, and associated equipment, are a common and often intrusive presence at FDA advisory committee meetings. Guidelines governing ''electronic recording equipment,'' ^** but pertaining mainly to television, have been set forth in (21 CFR 10, Subpart C (200–206)). These regulations vest authority for their administration in the "designated presiding officer," presumably the chair of an advisory committee.

In addition to television, the audience of an advisory committee meeting often includes individuals with other electronic devices, such as recording machines, telephones, and cameras. These are often referred to as nuisance items but pose relatively few problems for advisory committees, which conduct practically all of their business in open sessions.

Variations exist in the use of voting by FDA advisory committees. In CDRH, all committees are asked to vote on the regulatory issue of whether a given medical device should be approved. This practice is based on CDRH's interpretation that a vote is required by the provision of the Medical Device Amendments of 1976 that calls for an advisory panel to submit a "report and recommendation ... with respect to an application" [Food, Drug, and Cosmetic Act, §515(g)(2)(B)]. CDRH advisory panels are not, however, asked to vote on the scientific questions of whether information in an application shows a "reasonable assurance" that a device is safe and effective ''under the conditions of use prescribed, recommended, or suggested in the proposed labelling thereof" [Food, Drug, and Cosmetic Act, §515(d)(2)(A,B].

In CDER and CBER, voting practices follow the discretion of the committee chair or the tradition of the reviewing division. Some division directors take no votes, some ask for votes only on scientific questions, and others request votes on regulatory questions.

The IOM committee discussed at length a proposal by one member⁵ that advisory committee votes on questions before them be scaled (for example, from one to nine), rather than binary (yes, no). This proposal is based on three premises. First, safety, effectiveness, and other factors considered in advisory committee recommendations are continuous variables. Second, given that there are no definitive empirical bases for deciding issues before an advisory committee, the FDA should seek to determine both the range and strength of the experts' opinions. Third, a good deal could be learned by frequent scaled votes about the multiple facets of component questions that come before a committee, including the distribution of views among members on particular issues, as well as any persistent voting patterns or apparent biases.

The elements of the proposal are based on the modified Delphi system used by the RAND Corporation's appropriateness-of-care studies. In the RAND approach to voting, all votes are scaled from one to nine; individual votes are secret, but each member receives the distribution and his or her vote after each round; there are usually two rounds of votes per question,

with committee discussion between each round; summary statistics are generated on each question (e.g., the median after omitting outliers); and the vote distribution and summary statistics, but not individual votes, become a matter of record.

The IOM committee was intrigued by the proposal but found it too novel and formalistic to recommend for general adoption by the FDA. The committee favored binary votes that forced individual members to resolve uncertainty in an up-or-down manner and provide unambiguous advice to the agency. It thought that scaled votes on regulatory decisions to approve or disapprove a product might be confusing to the agency and to interested parties, at least initially. However, the IOM committee commended the proposal to use nonbinary, sophisticated voting procedures and tallying to the FDA for consideration on a pilot or demonstration basis.

Several issues were considered that relate to the neutrality of the agency in its relations with advisory committees. These include agency presentations, agency-committee interactions, and seating arrangements.

The Industry Liaison Panel emphasized the importance it attaches to neutrality of presentation by the agency. Such neutrality includes both the adoption of a dispassionate tone and the withholding of any judgment about a submission, even if asked.

Agency responses to the implications of this view vary. Some FDA officials counter that sponsor presentations routinely lack neutrality and reflect (charitably) a "best foot forward" posture that minimizes the problems and highlights the promises of an application. Those holding this view see themselves as obligated under the statutes governing FDA to act to protect the public health and to expose the weaknesses of a sponsor's submission and presentation. Neutrality, under the circumstances, is neither possible nor desirable.

Other FDA officials view the agency's responsibility as one of presenting a thorough, fair critique in a scientifically objective way and withholding any expression of a judgment about what should be done with the application. If asked for a judgment, some officials feel obliged to respond, while others decline to do so.

The response of Dr. Robert Temple to the Industry Liaison Panel report deserves quotation:

The IOM committee believes that agency presentations of its reviews of a sponsor's application should reflect a critical but fair evaluation of the data. If the agency has problems with the application, the identification of which reveals the agency's conclusion (tentative or fixed), the committee sees no reason to object to the communication of those concerns. Indeed, failure to do so might be construed as manipulation by not apprising the committee of factors that will be important in the agency's official decision.

On the other hand, the IOM committee strongly believes that agency presentations about a sponsor's application should be professional in tone, and thorough, fair, and scientifically objective in their critiques. The agency should acknowledge its public health responsibilities to ensure that sponsors provide evidence of product safety and effectiveness.

One question that has been raised is how FDA staff should interact with advisory committees in the conduct of a meeting. As a general proposition, agency staff should not attempt to dominate committee discussions but should seek (with the help and under the control of the chair) to elicit the views of advisory committee members. The chair of a committee should be instructed about the importance of appropriate use of discussion time by FDA staff and should exercise control over the meeting accordingly.

Complaints have been made that some seating arrangements at advisory committee meetings suggest an effort to influence the outcome of the deliberations. The IOM committee believes that the general principle that should govern seating arrangements is that they should be made, to the extent physical facilities permit, to facilitate the deliberations of the committee. The division director should not sit next to the committee chair. Other agency personnel should be placed so that a demarcation between committee members and FDA staff is apparent.

Many advisory committee members, both past and present and including those serving on the IOM committee, those interviewed by the committee, and others, have voiced the complaint that they are always required to function "on stage" and without the opportunity for closed discussions. The preference for the latter stems most often from a desire of scientists to engage in critical, often vigorous back-and-forth technical argument, which is impeded and often thwarted by the constraints of a televised, open meeting before an audience of several hundred individuals.

Regrettably, from the standpoint of facilitating scientific deliberations, the requirements of FACA and the Government in the Sunshine Act make it impossible to close committee meetings save except in defined circumstances. These include the discussion of proprietary information or, on occasion, confidential information about individuals.

As indicated in Chapter 3, the CDRH reorganized its advisory committee system in 1990, formally disestablishing 17 committees and reconstituting them as panels under a single Medical Devices Advisory

Committee. Subsequently, the combination product requirements of the Safe Medical Devices Act of 1990 led the FDA to issue product jurisdiction regulations in November 1991; in that same month it announced three inter-center agreements on the same subject. These efforts led to a rechartering of both the CDER and CBER advisory committees, a process that is being completed this year.

Under these new arrangements, each of these three centers will have the authority to draw needed experts from a long slate of committee members to serve as full voting members at specific panel meetings. This policy is intended to minimize the difficulties of assuring a quorum, to minimize conflict-of-interest problems by expanding the pool from which advisory committee members are drawn, and to allow a better match of expertise to agenda items.

In addition, however, these actions may engender a tension between increased flexibility to match expertise to need and the possible appearance of agency efforts to choose members of an advisory committee to achieve a desired outcome. On the one hand, the Industry Liaison Panel recommended the creation of a large pool of experts on whom FDA could draw for advisory committee members as a means to ensure a match between expertise and agenda item. This proposal reflects the view that the competencies sought on an advisory committee are often specific to a given agenda.

On the other hand, composing an advisory committee from a larger pool of members, however that pool is constituted, is not without problems. Building an advisory committee around a specific agenda leaves the agency open to accusations that the agency is manipulating the committee to achieve a desired outcome. Given, however, that the committees are only advisory and that the FDA is not bound by their advice, the argument that the FDA would deliberately construct a committee around a specific point of view is not compelling.

The IOM committee commends the agency for its rechartering of the CDRH, CDER, and CBER advisory committees to permit greater flexibility in composing a committee in which expertise is related to the subject matter of the agenda. However, the IOM committee also urges care in composing committees from the larger pool of available members.

Advisory committee members frequently complain that FDA provides no feedback to them on the results of their deliberations. They have no direct knowledge of the effect of their contributions and, in some measure, regard their input as diminished as a result.

The FDA does not strongly defend its current practice. It often notes that resource limitations on professional staff prevent it from fulfilling this function adequately. It also notes that the progress of an NDA approval is closely followed by the financial investment community, and simple prudence argues against informing committee members of a forthcoming approval in advance of notifying a sponsor.