Publishing Consensus Conference Proceedings: Can We Avoid Missed Opportunities While Effectively Managing Bias?

Le résumé fourni par la source

Every year there are innumerable consensus development conferences on topics related to medical diagnosis and treatment. These meetings, which typically convene multiple experts within a discipline, often involve substantial organization, commitment, and financial expenditure. They are labor-intensive and usually represent a high level of discussion, producing topical and sometimes compelling results. Unfortunately, much of this work is lost to the field after adjournment, which ultimately represents, in our view, a waste of time, effort, and potentially valuable clinical guidance. We recognize and value the results of some consensus development conferences that would be of interest to the readers of Pain Medicine; thus, we accept summary transcripts of these meetings to be peer reviewed and, if suitable, published. Although there may be import in such work, it is critical to consider the implicit (or inherent) and explicit bias of these scientific consensus-forming groups. Implicit bias is the unconscious expectations of the participants of a consensus group, whereas explicit bias is the conscious and overt bias/expectations of the sponsors, organizers, and participants of the consensus process. In assessing bias, it can be informative to classify consensus development conferences by type of funding source: federal/national funding agency–sponsored, philanthropic and/or advocacy group–sponsored, or private industry–sponsored. These three types of consensus development groups differentially represent degrees of expertise, value to the field, and bias. Federally sponsored consensus development groups represent perhaps the least degree of overall implicit and explicit bias, as they are designed to include the full range of expert and shareholder opinion. These consensus meetings usually concern some particular topical aspect of a given field and are usually published in a federal publication or registry. They are often “public” and by design welcome lay as well as expert input and comment. Lay input introduces predictable stakeholder bias, but this is often managed formally and informally by including a wide range of participating stakeholders, and an effort is usually made to ensure that all relevant points of view are included. The Department of Health and Human Services Pain Management Best Practices Inter-Agency Task Force, which published a report in 2019, is a recent example [1]. A broad array of professionals and laypersons convened in-person and by phone and e-mail over a 12-month period to develop recommendations that aimed to clarify the confusing landscape of pain management for the public. Consensus meetings sponsored by philanthropic and/or advocacy groups have an intermediate degree of bias. These meetings often have the specific aim of developing guidance on specific diagnostic techniques or treatments. They are often closed, and bias is managed by the diversity and professionalism (or lack thereof) of the participants. The participants usually are selected on the basis of interest, often with expenses paid, but usually no consulting fees. This type of consensus meeting, unfortunately, is not often published in the medical literature, and the results are used specifically by the sponsors. Pain Medicine has published summaries of some of these types of consensus development meetings, such as a detailed pragmatic review of ketamine for complex regional pain syndrome [2] and consensus guidelines on the use of urine drug screens [3]. Consensus meetings sponsored by private industry have the highest degree of implicit and explicit bias. The industry sponsors have a vested interest in aggregating participants of the highest level of expertise, and thus the value of the discussions and consensus can be very high. Commonly, the experts are paid consultants, working for the sponsors in highly organized formats. In most cases, these high-level discussions are designed to develop and plan specific drug or device regimens in a medical specialty. The high likelihood of significant bias inherent in these types of consensus development groups can be partially mitigated with some effort, yet “managing” bias is not the sponsors’ primary purpose. Rarely published, this type of meeting may sometimes be of great value to the field. In the midst of an unprecedented opioid epidemic, the paper by Webster et al. in this issue of Pain Medicine [4] contributes to an important dialogue on the safe and effective use of buprenorphine products that are Food and Drug Administration–approved for the management of chronic pain. At the same time, this review highlights the potential for implicit and explicit bias in the identification and recruitment of participants, literature review and synthesis, and consensus recommendations. First, most of the participants on the consensus panel received funding (i.e., advisory board, consultant, and/or research funding) from the conference sponsor, whose drug was the topic of the meeting. These relationships were appropriately disclosed. Second, the conference suffered from insufficient preconference literature review and synthesis. Systematic reviews provide an important foundation for consensus development. For this conference, literature reviews were not systematic and were further limited by a restrictive search strategy that yielded few articles, with apparently no grading of the quality or strength of the evidence. As a result, in formulating their opinions, participants may have had to rely disproportionately on their past, albeit expert and considerable, experience. Third, a key consensus recommendation—that clinicians consider switching directly from high-dose opioids to the sponsor’s buccal buprenorphine product without necessarily tapering—was made in the absence of published evidence. It is not clear why, or under what circumstances, this would be necessary. and notwithstanding the authors’ recommendation, the potential for precipitated withdrawal or difficult switching, necessitating the use of immediate-release pure mu-opioid receptor agonists, alpha-2 blockers, or both, can be challenging outside the hands of highly experienced clinicians who also can provide close and frequent follow-up during the (possibly prolonged) switching process. The editors of Pain Medicine recognize the value of facilitating the appropriate and timely dissemination of consensus recommendations to clinical practice. All research, and thus all publications, have a level of inherent bias, but we must take steps to recognize and transparently present the bias inherent in the consensus process if we are to publish reports of these efforts. Authors will be required to specifically acknowledge the sponsors’ intent and role and disclose in detail potential and actual conflicts of interest. Ideally, authors should discuss in detail the types of bias and the attempts and procedures designed to minimize and mitigate this bias, both in the conference process and in the final recommendations. The reviewing editors must recognize these issues. In general, editors and reviewers must be made aware of potential bias and maintain vigilance throughout the management of submitted manuscripts. Ultimately, readers must be educated about and informed of these sources of bias so that they may properly judge the value of the information presented.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.