February 2025 issue
February 28, 2025 | The recent NANS meeting in Orlando offered numerous non-scientific sessions. One in particular was a panel discussion bringing together some of the leading editors of scientific journals in the sector, including Chris Gilligan, Pain Practice, Robert Hurley, Pain Medicine, Robert Levy, Neuromodulation: Technology at the Neural Interface, and Michael Schatman, Journal of Pain Research. The session was moderated by Ryan D’Souza from the Mayo Clinic, and Nasir Hussain from Wexner Medical Center.
The panelists tackled pivotal topics such as the proliferation of journals, medical deadlines across journals/societies, authorship and conflicts of interests, and the use of AI in manuscripts.
The discussion sparked a thought-provoking dialogue on these topics and more.
The Role and Limitations of Medical Guidelines
Medical guidelines play a critical role in standardizing clinical practice, yet their effectiveness is often undermined by payer decisions and regulatory inconsistencies. As Michael Schatman pointed out, “All the guidelines in the world are wonderful for forming practice. Yet at the same time, the real crisis is all about payers making decisions and declaring everything we want to do as experimental, irrespective of the evidence basis that we present.”
Historically, there was a federal agency in the U.S. that reviewed, accepted, or rejected guidelines and maintained a registry. However, Levy highlighted a key shift: “Until about six years ago, the government had oversight, but now anyone can put out any kind of guideline without consistency or direction.” This lack of centralized oversight has led to a proliferation of inconsistent and sometimes contradictory guidelines, making it challenging for clinicians to follow best practices.
The Need for Standardization and Collaboration
A major obstacle in guideline development is the fragmentation of medical societies. Hurley underscored the issue: “There are way too many pain societies and spine societies, and there’s way too much overlap. That, in and of itself, poses a problem for guidelines because it’s very hard to come together when each society has its own interests and leadership.”
Despite these difficulties, multi-society collaboration has shown promise. The NIH Low Back Pain Task Force guidelines are an example of successful cooperation, having been published in more than 15 journals simultaneously. This demonstrates that when stakeholders align their goals, a unified and widely accepted standard is possible.
Conflicts of Interest in Medical Publishing
One of the most pressing issues in guideline development is the influence of industry. Levy explained, “Industry has managed to foster key opinion leaders in the field who are very biased toward their approaches or products. If you tease apart several of these guidelines, you find differences that are really based on the members of the panel and their interests with industry.”
Collaboration with industry is often necessary, as research in neuromodulation and other fields is prohibitively expensive without external funding. “Unless you have unlimited financial resources, collaboration with industry—even to a minimal degree—is required in our field,” stated one panelist. However, transparency is essential. Some organizations, such as the American Society of Hematology, prohibit any individual with industry conflicts of interest in the past three years from serving on guideline panels. Other fields, however, accept industry funding without the same level of scrutiny.
To address these concerns, for example, Neuromodulation: Technology at Neural Interface, has implemented disclosure policies requiring all authors to sign affidavits detailing potential conflicts of interest. While this system is not foolproof, falsifying such legal documents can carry severe professional consequences.
Addressing the Proliferation of Low-Quality Guidelines
The increasing number of guidelines—many lacking rigorous scientific foundations—has led to confusion among practitioners. “The quality of guidelines has decreased over the past several years,” Hurley states. “There’s a tremendous amount of heterogeneity, and in interactions with payers about patient access, there is a clear hierarchy of which guidelines carry weight, and which are quickly dismissed.”
Levy advocates for a more standardized approach, similar to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses). “We should establish ‘guidelines for guidelines’ to ensure consistency and reliability,” he said. By implementing stricter standards, the medical community can enhance the credibility and usefulness of clinical recommendations.
The Cascading Effect in Publishing
With the proliferation of new journals, many of which have minimal acceptance criteria or operate on a pay-to-publish model, the quality of journal publications today varies significantly. Some journals employ a “cascade” system, where a rejected manuscript is passed along to other affiliated journals for consideration. One panelist observed that many articles should not be published at all, yet some authors persist in submitting their work to multiple journals until they find one willing to accept it, even if it is of lower quality.
Regarding the respected Wiley journals, Gilligan explained, “If we are rejecting the paper, we can reject it and refer it to some of the other Wiley journals or we can reject it. And not refer it depending on the quality.”
The Impact of AI on Medical Publishing
The rise of artificial intelligence has introduced new challenges in medical publishing, particularly in the realm of fabricated data. “We’re seeing an increasing number of AI-generated manuscripts with falsified data,” Levy warned. “It’s becoming incredibly hard for editors, reviewers, and section editors to detect these issues.”
A significant concern is the creation of seemingly legitimate references that do not exist. AI can generate citations that appear credible but lead nowhere. To combat this, some journals such as Neuromodulation: Technology at the Neural Interface, have implemented affidavits requiring authors to disclose any AI-generated content. Hurley also noted the development of software capable of detecting AI-written text, though its accuracy remains a point of contention. “Some programs claim a paper is 100% AI-generated, but how do they determine that? It’s not pulling a watermark from ChatGPT. It’s essentially saying, ‘This is written too well for a human.”
The problem extends beyond individual papers. Some authors have submitted more than 100 articles in a single month, flooding journals with AI-generated content. This necessitates stricter oversight and verification processes to maintain the integrity of medical literature.
Ensuring Ethical and Effective Guidelines for the Future
Despite these challenges, there are clear steps that can be taken to improve the credibility and utility of medical guidelines. Greater collaboration among societies, increased transparency in conflict-of-interest disclosures, and the implementation of standardized guideline development frameworks are all crucial measures. Additionally, developing robust AI detection methods and stricter editorial policies can help safeguard the integrity of published research.
Hurley underscored the importance of humility and cooperation in guideline development: “If we can decrease our ego a little bit and increase our collaborations, we can put out guidelines that have national and international impact.”
By addressing issues such as payer influence, conflicts of interest, AI-generated content, and the proliferation of low-quality guidelines, the medical community can move toward more effective and trustworthy clinical guidance. The future of medical guidelines depends on a collective commitment to rigor, transparency, and ethical integrity.
