The use of virtual expert panels: formulary decision-making in the 21st century

Abstract Evidence-based medicine is the driving philosophy for pharmacy and therapeutics (P&T) committee decision-making. When the available evidence is incomplete or unclear, the decision-making process becomes complex. Evaluation of the published literature to make sound decisions regarding appropriate medication use can be a time-consuming process. The use of expert panels or subcommittees may facilitate decision-making; however, this process is often limited by the schedules and time constraints of busy clinicians. This article describes the creation and utilization of a "virtual" panel of experts utilizing anonymous electronic communications to assist the P&T committee at a university-based teaching hospital in making medication use policy. The article includes a detailed description of the P&T committee's experience in the selection of virtual panel members, methods used, advantages, potential pitfalls, and the outcomes of a virtual committee used to evaluate the appropriateness of the off-label use of nesiritide (Natrecor, Scios). (Formulary. 2005;40:78-85.)

Increasingly, evidence-based medicine is being used for sound formulary decision-making to determine which drugs should be available and, if available, how these therapies should be used.¹ In this context, evidence-based medicine uses the best clinically relevant research obtained from a systematic review of the published literature along with clinical expertise to make decisions. Published clinical studies and data may support the use of a particular drug, provide evidence that a drug should not be used, or show that there is insufficient evidence for a particular use. If the evidence is incomplete or insufficient, decisions become pragmatic and may be based on additional factors, including safety and cost. When there is insufficient evidence, a conservative approach is recommended.¹

Evaluating the available evidence for a drug-related issue can be time consuming. This is particularly problematic when evidence is insufficient. The use of expert panels or ad hoc committees may facilitate this process. A consensus, based on the available evidence and local context, can aid P&T committees as they make important drug-related decisions.

EXPERT PANELS The P&T committee is the formal line of communication between an institution and the medical staff regarding all drug-related matters. As such, it relies on the power base of the medical staff members in making important drug-related decisions. Creation of a formulary, placing limitations on drug use, and establishing drug-use policy can be difficult when the topic is controversial and/or the evidence incomplete.

The use of standing subcommittees is one approach to handling issues that require special expertise or extra deliberations that the time constraints of P&T committee meetings will not permit. However, the use of standing subcommittees can be burdensome. Our experience at Shands (Hospital) at the University of Florida (Shands at UF) has been that the demands of academic faculty and clinical staff have made it increasingly difficult to find participants for standing committees. Therefore, we prefer to use ad hoc committees or expert panels to assist the P&T committee in making complex decisions. "Experts" in this context are physicians, pharmacists, or other health-care professionals who have special knowledge or skills, but who do not have a direct interest in final recommendations of the ad hoc committee.

We believe that members of an ad hoc expert panel should be appointed based on the subject that is being considered. Others may prefer standing ad hoc committees that have agreed to handle issues as they arise. Our approach allows us to select members based on their particular areas of expertise. The standing ad hoc committee approach eliminates the time required to identify and recruit members when they are needed.

Our experience has been that when used responsibly, expert panels can be a valuable supplement to the P&T committee decision-making process. A major limitation of the use of expert panels has been logistics. It can be difficult to schedule meeting times that will fit into the busy schedules of all the participants. This challenge can decrease the effectiveness of expert panels and delay resolution of drug-related issues.

'VIRTUAL' EXPERT PANELS An alternative approach to enhance the efficiency of ad hoc committees is to use "virtual" expert panels, in which a major portion of the communication is done by e-mail. This approach avoids the logistics of finding mutually convenient meeting times and allows participants ample time to consider facts and reflect on solutions, thereby streamlining the formulary decision-making processes.

There are other potential benefits to virtual expert panels. When we considered the implementation of a virtual expert panel, we felt it would be important to preserve the anonymity of the members of the group. A designated facilitator sends all correspondence and is the only identified member of the expert panel. By keeping the members of the panel anonymous, no members are allowed to dominate the discussion. This system permits all members to have equal input on the topic. The structure also allows members to be frank without concern about jeopardizing interpersonal professional relationships with colleagues or affecting tenure and promotion decisions from supervisors.

The goal of an expert panel is to reach a consensus opinion. Well-described features of formal consensus methods include anonymity, iteration, controlled feedback, and statistical response.³ Formal methods used to establish consensus, such as the Delphi technique or the nominal method, allow responses to be anonymous, which encourages input from all members of the group. Virtual expert panels also allow for several iterations of "discussion" and thereby provide the ability for participants to change their minds along the progression of iterative discussions. The feedback is controlled (by the facilitator), a summary of all opinions is expressed in anonymity, and more than just a consensus statement can be included.

One limitation of virtual expert panels is that the structure does not lend itself well to defining the specific questions to be addressed. Electronic communications appear to be most efficient when questions are defined a priori. Defining underlying questions is most efficiently done with face-to-face discussions, because each person can offer and receive input from others.

As with all formulary issues, it is imperative to have good background intelligence to shape the questions to be addressed by virtual expert panels. This can be accomplished by the use of computer databases, medication use evaluations (MUEs), and input from pharmacists and physicians practicing in the areas in which an issue is occurring. After an issue has been identified, input is sought from the major prescribers and/or services in order to gain a greater appreciation for all perspectives.

Once the issues and major interested parties have been defined, a virtual expert panel may be considered. "Interested parties" include those practitioners who have a direct stake in decisions made because outcomes of the process would directly affect their clinical practice. Again, this approach will be most successful when the issue can be summarized in terms of discrete evidence-based questions.

Our process at Shands at UF then follows these steps. Even before the interested parties are contacted, a preliminary evidence-based review of the literature is performed. This step helps facilitate productive, informed discussions with the interested parties. The interested parties are then contacted and asked to make their case in writing, which includes providing references to support their assertions. Further, they are asked if they are aware of any other interested parties who should be included in the process.

Once all the input from interested parties has been obtained, an expert panel is formed. The majority of the panel members should be physicians. Pharmacists may aid this process by providing evidence to consider and assisting with interpretation of the available information. Similar to the P&T committee, the authority of the expert panel is dependent on this structure. Expert panel members may also be members of the P&T committee, but omitting nonmembers from the panel may elicit criticism. In order to maintain the credibility of the committee's decisions, it is advisable to exclude individuals with actual, "potential," and/or perceived conflicts of interest from participating. This may be difficult, particularly in an academic setting in which many key opinion leaders participate in drug company-funded research or are speakers or consultants for drug manufacturers. As with our in-person P&T committee process, we ask participants to recuse themselves if they feel they have a conflict of interest relevant to a specific drug decision. We also ask for disclosures from panel members to permit those interpreting a panel's opinions to consider any relevant relationships.

Physician members of the committee must be respected members of the medical staff. The appropriate medical specialties should be represented based on the topic. The panel should also include members who are familiar with the evidence-based approach to problem-solving. We have found that it is important for the P&T committee chair to personally contact each member of the panel to explain the process, emphasize the value of the panel to the institution, and share expectations about the process and its planned duration before the panel begins deliberations.

The facilitator contacts each member of the expert panel by e-mail to reinforce the process and to make sure that the panel member's e-mail address is current and is routinely monitored. From experience, we have learned to not assume that published e-mail addresses are current or accurate. If a panel member does not reply promptly to the initial contact, the member (or the member's assistant) is contacted to verify the correct e-mail address.

Timelines are clearly established and presented in the initial correspondence, although they may be modified as needed for fair and thorough deliberation. The protocol allows for multiple iterations in the process. This gives members of the panel an opportunity to change their minds based on the comments of other panel members. Because of the panel members' anonymity, participants can change their position on an issue without being concerned about appearances or peer-pressure issues.

The first correspondence is by regular mail. It includes an evidence-based summary of the issues, including copies of all pertinent articles. The correspondence also includes a copy of the written comments of the "interested parties." Most of the subsequent correspondence is by e-mail. However, if a member of the panel asks for additional information that cannot be e-mailed, that information will be hand-delivered to each member of the panel.

The charge of the panel is also presented. The panel is asked to answer specific questions, with the ultimate success of the panel determined by how clearly these questions have been answered.

The panel members are not, however, required to limit their comments and opinions to the charges-any topic is fair game. The final report of the panel will include a summary of any "side issues" that arise during the virtual discussion of the charges.

With each iteration of the panel's virtual discussion, a turn-around time is specified. Each deadline is relatively short but sufficient for panel members to consider the issues. The turn-around times get shorter as the process proceeds. It is the responsibility of the facilitator to set these deadlines. However, any member can request an extension in order to have time to fully consider all of the issues.

After the discussions are completed, the panel members are asked to vote yes or no to the questions that have been posed. If participants have asked that additional questions be included, they also vote on those. The P&T committee receives a final report with the tallied votes of the panel and any statements that panel members feel should be included. Before this report goes to the P&T committee, it is sent back to the panel for review and final approval.

The panel does not make policy. The panel functions only in an advisory role to the P&T committee. Members of the P&T committee may be appointed to the panels, when appropriate, which provides some continuity between the expert review and the policy decisions. By removing the need for an "action" recommendation, the expert panel members can focus on simply answering the questions from an evidence-based perspective. The P&T committee considers the final report of the expert panel. However, the P&T committee has sole responsibility for determining the final action to be taken.

Nesiritide was reviewed and added to the formulary at Shands at UF in September 2001, based on its ability to optimize hemodynamic parameters and improve symptoms in patients with acute decompensated CHF. No restrictions were placed on nesiritide use. Initially, its utilization was modest, but the use gradually increased over time.

A potential problem with our institution's use of nesiritide was identified in the summer of 2003. Utilization data from other University Hospital Consortium hospitals indicated that our institution's use was considerably more than other hospitals. Additional benchmarking data confirmed that our expenditures on this product greatly exceeded similar hospitals.

Information gathered demonstrated that nesiritide was being used for its labeled indication (ie, acute decompensated CHF) and extensively for off-label indications. The 2 off-label indications identified were long-term use for patients awaiting heart transplantation (ie, bridge to cardiac transplantation) and prevention of acute renal failure in patients undergoing cardiothoracic surgery. Although there were theoretical pharmacological reasons to support the role of nesiritide in these situations,⁶ an evidence-based review suggested limited published information existed to support off-label uses (Table 1).^7-10

Based on the increasing utilization of nesiritide, it was projected that fiscal year expenditures for 2003-2004 would exceed $2.6 million for this drug alone (Figure 1). This was many times the projected pharmaceutical impact that had been anticipated when nesiritide was added to the formulary and many times the average use by other comparable hospitals.

Based on the investigation of nesiritide's use, the chair of the P&T committee sent a letter to the services using nesiritide for off-label indications. They were given a deadline to provide in writing their rationale for the continued off-label use of nesiritide, including any published evidence or information that had been presented at clinical meetings. This information and an independent evidence-based review completed by Shands at UF's Drug Information and Pharmacy Resource Center were presented at the September 2003 P&T committee meeting.

The P&T committee decided that an expert panel would be formed to determine whether there was sufficient evidence to justify the off-label use of nesiritide. In October 2003, members were selected, and the P&T chair contacted each participant to review the charge of the panel, the expected commitment, and how the findings of the panel would be used.

Nine people were invited to participate in the expert panel, and only 1 physician declined the invitation. There was broad representation from medical specialties, including cardiology and nephrology. Three physicians had expertise in evidence-based medicine. Two pharmacists were appointed to the panel: 1 facilitator, who was considered 1 of the 8 panel members; and 1 non-voting participant, who assisted in the evidence-based review of nesiritide. Although members were screened for potential conflicts of interest, a formal disclosure form was not required. This may be required for future expert panels.

The facilitator contacted members by creating an e-mail group that included all the participants, but all correspondence was sent blind carbon-copied (bcc) so that members remained anonymous. The facilitator was the only unblinded member of the panel and the e-mail group. Each member replied individually to the facilitator, who then sent all comments to the entire panel.

Throughout the virtual discussion, the panel members redefined the questions to be answered (Table 2). The panel then examined the evidence available for each question.

The expert panel members agreed on recommendations by the end of November 2003 and provided the information to the P&T committee that met in December 2003. Based on the recommendations of the expert panel, the P&T committee voted to restrict nesiritide use to its labeled indication and for use as a bridge to cardiac transplantation. The P&T committee recommended the development of a Nesiritide Order Form that would facilitate ongoing monitoring of use and enable a clinical pharmacist to do a concurrent MUE to determine compliance with the approved restrictions. Although nesiritide would not be denied when ordered, continued therapy would have to be approved based on these criteria. The services most impacted by these limitations were notified of the new restrictions before their implementation.

The restrictions and the Nesiritide Order Form were implemented in February 2004. This had a profound effect on the utilization of nesiritide (Figure 2). The off-label use of nesiritide decreased dramatically, and the use of nesiritide for acute decompensated (fluid overloaded) CHF also decreased.

UNEXPECTED BENEFITS OF EXPERT PANELS The services using nesiritide felt strongly that it was effective for the off-labeled indications in the cardiothoracic surgery population despite the lack of published evidence. Based on the pathophysiology of cardiothoracic bypass and the pharmacology of nesiritide, the impression was that nesiritide decreased acute renal failure in this patient population. The evidence-based review of nesiritide highlighted the lack of objective science to support this contention.

These discussions led to the formation of a group of faculty who were interested in studying nesiritide use in this clinical setting. Because there had been considerable off-label use of nesiritide before limitations were imposed, an additional recommendation of the nesiritide expert panel was that observational research be considered to determine if there was merit for further research. Based on the recommendations of the expert panel, the hospital agreed to fund an observational study to examine this issue.

If the observational study shows promise, there is interest in pursuing a prospective, randomized, controlled trial to determine the value of nesiritide in the prevention of acute renal failure following cardiothoracic bypass surgery. The results of the observational study will aid in the design of the prospective trial (eg, inclusion and exclusion criteria).

Without the review of the expert panel and the actions of the P&T committee, the drive for the creation of new information would not have occurred as quickly as it did. When faculty review objective information and discuss the merits of drug use in this type of setting, the process can be an incubator for original research.

CONCLUSIONS In our experience, virtual expert panels can be useful to P&T committees. Used appropriately, there are several potential benefits:

This procedure works well when the charge of the panel is well defined. When the clinical issues are not as well defined, the virtual nature of the discussion may prevent the necessary interaction to define the issues. With additional experience, we will be able to determine when it is best to use virtual panels versus ad hoc committees that meet face-to-face to resolve important issues.

Identifying participants for expert panels can be a challenge. Busy faculty must balance the commitment against their other responsibilities. Therefore, panels must be used judiciously. Also, after the panel has completed its charge, the chair of the P&T committee formally recognizes the participants' contributions. A letter is sent to the immediate supervisor of the faculty member and the chief operating officer of the hospital outlining the participation in the panel and the value to the hospital and university. Panel members are copied on this correspondence, which will be considered service to the hospital during the participant's annual review.

Our experience with this process at Shands at UF is specific to a university-based teaching hospital setting. Whether this process could be adapted to other formulary settings is unknown. We believe that other formulary settings, such as large HMOs or group models, could benefit from the logistical benefits of this process. However, there may be specific obstacles and challenges in these settings of which we are unaware.

Dr Hatton is co-director of the Drug Information and Pharmacy Resource Center, Shands (Hospital) at the University of Florida (Shands at UF), Gainesville, Fla, and a clinical professor, College of Pharmacy, University of Florida. He can be reached at hatton@ufl.edu

Dr Gonzalez-Rothi is a professor, College of Medicine, University of Florida, and chair of the pharmacy and therapeutics committee, Shands at UF.

Dr Smith is a clinical pharmacy specialist in Drug Information/Formulary Management, The Methodist Hospital, Houston, Tex. (At the time this article was written, Dr Smith was the drug information specialty resident at Shands at UF.)

Mr Knudsen is director of Pharmacy Services, Shands at UF, and assistant dean for hospital programs, College of Pharmacy, University of Florida.

REFERENCES 1. Eddy DM. Evidence-based medicine: What it is, why to use it, and how to incorporate it into decision-making. Formulary. 2002;37:525-530.

2. Jenkings KN, Barber N. What constitutes evidence in hospital new drug decision making? Soc Sci Med. 2004;58:1757-1766.

3. Jones J, Hunter D. Consensus methods for medical and health services research. BMJ. 1995;311:376-380.

4. Natrecor [package insert]. Sunnyvale, CA: Scios, Inc; 2003.

5. Vichiendilokkul A, Tran A, Racine E. Nesiritide: a novel approach for acute heart failure. Ann Pharmacother. 2003;37:247-258.

6. Zineh I, Schofield RS, Johnson JA. The evolving role of nesiritide in advanced or decompensated heart failure. Pharmacotherapy. 2003;23:1266-1280.

7. Hill JA, Hsu K, Pauly DF, Schofield R, Aranda JM Jr. Sustained use of nesiritide to aid in bridging to heart transplantation. Clin Cardiol. 2003; 26:211-214.

8. Truong KM, Bernabei AF, Czerska B, et al. Nesiritide use in end stage heart failure [abstract]. J Heart Lung Transplant. 2003;22(Suppl 1):S110.

9. Moazami N, Damiano RJ, Bailey MS, et al. Nesiritide (BNP) in the management of postoperative cardiac patients. Ann Thorac Surg. 2003;75:1974-1976.

10. Radovancevic B, Vrtovec B, Yazdanbakhsh AP, et al. Perioperative use of nesiritide in heart failure patients undergoing implantation of left ventricular assist device [abstract]. J Heart Lung Transplant. 2003;22(Suppl 1):S201. Abstract 391.

11. Lazar RM, Shapiro PA, Jaski BE, et al. Neurological events during long-term mechanical circulatory support for heart failure: the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) experience. Circulation. 2004; 109:2423-2427.

12. Bentz B, Hupcey JE, Polomano RC, Boehmer JP. A retrospective study of left ventricular assist device-related infections. J Cardiovasc Manag. 2004;15:9-16.

13. Sivaratnam K, Duggan JM. Left ventricular assist device infections: three case reports and a review of the literature. ASAIO J. 2002;48:2-7.