Editorial Commitment to Trust and Integrity in Science: Implications for Pain and Anesthesiology Research

Tonya Palermo^1,2*Didier Bouhassira³Karen D Davis^4,5Hugh Hemmings⁶Robert W Hurley⁷Joel Katz⁸Jaideep Pandit^10,9Theodore John Price¹¹Michael E Schatman¹²Stephan K W Schwarz^13,14Dennis C Turk¹Marc Van De Velde¹⁵Matthew D Wiles^16,17Tony L Yaksh^18*David Yarnitsky¹⁹

• ¹University of Washington Department of Anesthesiology & Pain Medicine, Seattle, United States
• ²Seattle Children's Research Institute, Seattle, United States
• ³INSERM, Paris, France
• ⁴University of Toronto Institute of Medical Science, Toronto, Canada
• ⁵Krembil Research Institute, Toronto, Canada
• ⁶Weill Cornell Medicine Department of Anesthesiology, New York, United States
• ⁷Wake Forest University School of Medicine, Winston-Salem, United States
• ⁸Department of Psychology, York University, Toronto, Canada
• ⁹Nuffield Department of Clinical Neuroscience, University of Oxford, Oxford, United Kingdom
• ¹⁰Nuffield Department of Anaesthesia, Oxford University Hospitals NHS Foundation Trust, Oxford, United Kingdom
• ¹¹Department of Neuroscience, The University of Texas at Dallas Center for Advanced Pain Studies, Richardson, United States
• ¹²Department of Anesthesiology, Perioperative Care, and Pain Medicine ; Department of Population Health – Division of Medical Ethics,, New York University Grossman School of Medicine, New York, United States
• ¹³Department of Anesthesiology, Pharmacology & Therapeutics,, The University of British Columbia, Vancouver, Canada
• ¹⁴Department of Anesthesia, St Paul's Hospital, Vancouver, Canada
• ¹⁵Katholieke Universiteit Leuven Departement Cardiovasculaire Wetenschappen, Leuven, Belgium
• ¹⁶Department of Academic Anaesthesia, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, United Kingdom
• ¹⁷Centre for Applied Health and Social Care Research (CARe), Sheffield Hallam University, Sheffield, United Kingdom
• ¹⁸University of California, San Diego, La Jolla, United States
• ¹⁹Department of Neurology, Rambam Medical Center, and Laboratory of Clinical Neurophysiology, Technion Israel Institute of Technology Faculty of Medicine, Haifa, Israel

We are a group of journal editors i dedicated to advancing discoveries and innovations in basic, translational, and clinical research across anesthesiology and pain-related disciplines, which play a crucial role in reducing the burden of pain, improving health, enhancing perioperative outcomes, and optimizing healthcare delivery. Across scientific disciplines, concerns have been raised about research quality and trustworthiness 1,2 . While these challenges are not unique to pain and anesthesiology research, we recognize this as a judicious opportunity to raise awareness and collaborate across our journals to align and strengthen initiatives to enhance research integrity, trust, and impact across our field.In a 2005 landmark paper, John Ioannidis concluded with the dramatic and troubling assertion that "most published research findings are false," stimulating a large focus in the biomedical research community on understanding issues of integrity, reproducibility, and replication that continues to be relevant to this day 3 . Indeed, there are many instances in which authors, institutions, funders, publishers and journals have failed to embody the core values that produce trustworthy science. The trustworthiness of research is affected by both intentional actions (e.g., fabrication and falsification of data, lack of rigor, image manipulation) and unintentional actions (e.g., inadequate oversight, awareness and understanding of both technical and scientific issues). Most concerning are instances of research misconduct including fabrication, falsification, or plagiarism sometimes revealed by failure to replicate or reproduce results, duplication of publications, a rise in the number of retractions, 4,5 and calls for larger numbers of papers to be retracted (e.g., 2 ). In support of Ioannidis's disquiet, some reviews (e.g., 6 , 7 ) report low replication rates of positive findings in the social and life sciences across clinical trials, epidemiological research, and molecular studies.In anesthesiology specifically, low agreement has been found between randomized clinical trials (RCTs) and meta-analytic findings for clinical pain interventions, where positive findings in meta-analyses were often not confirmed by subsequent large RCTs. For example, using individual patient data from RCTs published in Anaesthesia, Carlisle 8 demonstrated that almost half of the databases had false data as detected from the duplication of figures, tables, and other data from published work; the duplication of data in the rows and columns of spreadsheets; impossible values; and incorrect data analytic strategies and calculations.Reproducibility, clinical validity, and utility in pain and anesthesiology research are often compromised by non-representative samples (e.g., limited representation on characteristics such as race, ethnicity, age, sex/gender, or socioeconomic status that do not match populationlevel data of those most affected by pain) [9][10][11] , reliance on surrogate outcomes with limited clinical relevance, underutilization of common data elements and core outcome sets, underpowered studies prone to false-negative results, and flawed statistical analysis plans that generate misleading conclusions 12 .To ensure integrity of the literature, retraction of articles may be necessary due to such issues as major errors, data fabrication, plagiarism, or unethical research practices. Authors are encouraged to identify errors in their own work and may request a corrigendum to correct the literature. However, when ethical issues are brought to a journal's attention, they have a duty to investigate, and when there is conclusive evidence, to impose a retraction to alert readers that the findings and conclusions cannot be relied upon 13 . Retractions, when reported, can have a widespread impact due to the interconnectedness of studies attributed to the same authors 14 . In the field of anesthesiology, the Retraction Watch Leaderboard 15 indicates four of the top ten authors are anesthesiologists, and two of these individuals occupy the top two positions (https://retractionwatch.com/the-retraction-watch-leaderboard/. Systematic reviews have summarized characteristics of retracted publications for research misconduct in pain (e.g., 16 ) and anesthesiology research (e.g., 17 ). Concerns regarding retractions in all scientific fields are particularly noteworthy because they undermine trust in science, can have a lasting impact on conclusions made about treatments and, ultimately, impact clinical practice. In one study by O'Connell et al. 18 , a set of 8 untrustworthy trials (i.e., identified due to concerns including data anomalies and implausible results), in spinal pain was determined to substantially impact the results of subsequent recommendations made in systematic reviews and international clinical practice guidelines in management of spinal pain. Recommendations were made for short-term as well as more extended-term actions and behaviors for several different stakeholder groups (e.g., researchers, institutions, publishers, funders, policymakers and regulators, peer reviewers) to support trustworthy research within each of the core values of ENTRUST. These recommendations are intended to guide the development of a strategy for enhancing trustworthy research, rather than serving as a mandated policy.From the perspective of engagement with our journals, here we focus on recommendations for researchers and editors/publishers.We strongly recommend that researchers thoroughly review the proposed framework, which we as editors endorse, and explore the full suite of resources available through the ENTRUST-PE network project. These can be accessed at https://entrust-pe.org and on the Open Science Framework (https://osf.io/cua7g/?view_only=ec1d9e6b1d774dbca9306ff5ae4dec67). The initiative is designed to support researchers to understand how to conduct and report science in a manner that enhances the transparency and trustworthiness of their work. By following these recommendations, researchers can provide the highest quality of research and facilitate confidence in pain science. Moreover, peer reviewers and consumers of research can be alert to potential issues of methodological rigor, transparency, lack of equity and inclusivity, and markers of potential data inauthenticity or research misconduct that play a critical role in raising concerns to editors and publishers when these are identified. Recently, both the International Association for the Study of Pain (IASP) 23 and the European Pain Federation (EFIC) 24 endorsed the ENTRUST-PE initiative, recognizing that trustworthy research not only benefits investigators and healthcare professionals but also serves patients and the public by promoting science that produces more effective pain management strategies.For a concise, actionable summary, we reproduce the guidance provided by O'Connell et al. 22 , which outlines practical suggestions researchers can implement immediately to align with the core values of the ENTRUST-PE framework (see Table 1).We plan to undertake an updated and coordinated self-assessment process across our 15 journals using the procedures outlined by Cashin et al. 20 . This will provide a critical update on current engagement efforts with transparency standards across a larger number of pain and anesthesiology journals. Such an assessment will provide the journals with a list of potential areas for improvement to guide their efforts.(2) Gain access to automated tools to improve transparency and trustworthiness, while fostering innovation in new methodologies.Innovations are needed to support a range of automated processes to enhance transparency and integrity. At present, multiple checks of transparency and trustworthiness are conducted manually by reviewers and editorial teams. Journals can carry out protocols in the work flow prior to the initiation of peer review around many indicators for quality, trustworthiness, and ethics concerns such as possible image manipulation, internal inconsistencies in referral to figures and tables, text plagiarism, adherence to reporting checklists, registration of systematic reviews, identifying discrepancies between research registrations (e.g., ClinicalTrials.gov) and reporting of clinical trial outcomes, and the inclusion of relevant animal and human review board approvals, to name a few. One example of checking for random sampling in RCTs is the method suggested by Carlisle and colleagues, but (a) this is labor-intensive and (b) does not apply where recruitment has not been entirely random 27,28 . Although there are automated processes to check for duplicate text, there are none yet to assist with these data integrity checks, and this requires dedicated staff effort. In this regard, several publishers/journals have introduced advanced technology (i.e., artificial intelligence) to detect duplicate manuscript submissions across all their respective journal platforms. Others have initiated "flag alerts" for authorships that include individuals who have been associated with multiple manuscript retractions. Additional automated processes are needed to help authors, reviewers, and editors to standardize more thorough yet efficient approaches to enhance transparency of reporting and enhance trustworthiness of published work.Several approaches can be used to identify areas for improvement in this area. For example, we can engage in robust discussions with our publishers to emphasize the importance of automated tools, checks, and alerts, and advocate for their implementation in our journals. In addition, we can continue to advocate for adequate staffing to enable the critical checks needed for pre-review of submissions by the journal, which requires explicit formal training of a stable journal staff. While using advanced technology and providing journal staff entails a heightened responsibility of the publisher with possible financial consequences, it increases our confidence in the integrity of the research and builds trust in our science. We can also provide guidance and when possible, share resources (e.g., "how to guidance") with our authors to enhance their own knowledge of tools to increase trustworthy science. For example, some reference management software (e.g., Zotero [Corporation for Digital Scholarship, Vienna, VA, USA], EndNote™ [Clarivate, London, UK]) have capabilities to check references for retractions 29 .(3) Create a platform for collaboration among editors of leading pain and anesthesiology journals This editorial highlights a significant collaboration among editors of leading pain and anesthesiology journals, which can serve as a foundation for continued engagement. We suggest holding online annual meetings and developing other platforms for information exchange for this group to discuss emerging trends, ethical concerns, and resource sharing. This may also serve as a forum for discussing general or specific integrity concerns and addressing the removal of inauthentic data from the literature, while ensuring confidentiality and privacy are upheld. We also recognize that there are barriers to engaging in transparency and integrity standards and anticipate initiating dialogue to better understand these barriers and how journals can support authors without increasing burden.(4) Offer educational opportunities and resources to professional societies, forums, journal reviewers, and early-career professionals Journals can be an important resource to guide and teach researchers and consumers about transparency and integrity standards, and we see several opportunities to make an impact. For example, one opportunity to introduce standards for trustworthiness is through the system adopted by several of our journals for manuscript review mentorship and editorial fellowship that provides tutorials, training, and experience reviewing or managing manuscripts. Moreover, we can leverage our partnerships with the professional societies that are associated with many of our journals to offer training and instruction on transparency and integrity. This could include professional development programs for reviewers, as well as early-career faculty (e.g., North American Pain School), and offerings developed by groups such as the International Association for the Study of Pain's Early Career Network, (https://www.iasp-pain.org/early-career-network/), and by setting expectations for presenting and sharing research at scientific meetings (e.g., checking for retractions of any published studies discussed in presentations).Our journals can help disseminate information on tools targeting researchers directly 30 that can be made available to authors in a toolkit to assist them in pursuing values of openness and integrity. For example, statistical assessment tools to assess the accuracy of reported findings may be implemented by running simple, automated error checks, such as using the StatCheck tool 31 . It should be stressed that increasing the education provided enhances quality, reliability, and integrity.Ultimately, as a community of scientists and clinicians in pain and anesthesiology, we must recognize that trust is a dynamic and multifaceted concept. Reproduced with permission from O'Connell et al. 22