Fact-checked by Grok 2 weeks ago

medRxiv

medRxiv is a free online preprint server dedicated to the rapid dissemination of unpublished manuscripts in the health sciences, including clinical, public health, and medical research. Launched in June 2019 by Cold Spring Harbor Laboratory in collaboration with Yale University and BMJ, it provides a platform for researchers to share complete but unpeer-reviewed studies, explicitly cautioning that such preprints should not inform clinical decisions or policy due to the absence of formal validation. The server has facilitated accelerated knowledge sharing, particularly during the , where it hosted over 20,000 submissions in its first year alone, enabling early visibility of emerging data on , treatments, and . However, this speed has drawn criticism for amplifying unverified claims, with some preprints later retracted or contradicted upon , contributing to public confusion and misguided applications in media reporting and decision-making. medRxiv employs to screen out submissions posing risks of harm, such as unsubstantiated alarms about carcinogens, yet instances of removed content critical of measures have raised questions about selective oversight in a field prone to institutional biases favoring consensus narratives. As of 2025, management has transitioned to the non-profit openRxiv, reflecting an evolution toward community-driven governance while maintaining its core mission of to preliminary research.

History

Founding and Early Development

medRxiv was founded in June 2019 through a between (CSHL), , and to establish a dedicated preprint server for health sciences research. The initiative addressed longstanding barriers in medical publishing, where traditional often delayed dissemination of findings by months or years, limiting timely access for researchers and clinicians. The server was announced on June 5, 2019, with submissions opening the following day, June 6, emphasizing a publisher-neutral platform to foster openness and in clinical and studies. Key figures in the founding included John Inglis and Richard Sever from CSHL, which had successfully operated the biology-focused since 2013; Joseph S. Ross and Harlan Krumholz from , who advocated for greater transparency in ; and Theodora Bloom and Claire Rawlinson from , bringing expertise in medical publishing standards. This partnership leveraged CSHL's technical infrastructure for preprint management, Yale's insights, and BMJ's global reach in health knowledge dissemination to create a non-profit service independent of commercial interests. In its early phase, medRxiv implemented features such as digital object identifiers (DOIs) for each to ensure citability and discoverability, alongside formation of an international to guide policies on content suitability for health sciences. The platform prioritized rapid posting while acknowledging sensitivities in , such as potential risks from unvetted clinical data, though initial operations focused on building submission workflows modeled after bioRxiv's proven system. By mid-2019, it had begun attracting submissions in areas like and therapeutics, setting the stage for broader adoption amid emerging challenges.

Launch and Initial Safeguards

MedRxiv began accepting manuscript submissions on June 6, 2019, following an announcement on June 5, 2019, by its founding partners: , School of Medicine, and BMJ Publishing Group. The platform was established to provide a dedicated server for sciences , addressing the absence of such a resource in fields like clinical trials and , where rapid dissemination was hindered by traditional delays. Unlike , which operates with minimal screening to prioritize speed in , medRxiv implemented initial safeguards tailored to the higher stakes of , including potential risks to and from unvetted clinical data. Submissions underwent a multi-step screening process: initial in-house review by staff with scientific and editorial expertise checked for completeness, , and basic scientific formatting, followed by evaluation by external clinical experts for ethical compliance, such as appropriate oversight of human subjects research and registration. Content posing health risks, lacking ethical standards, or containing offensive or non-scientific material was rejected, though scientific validity or methodological rigor was not assessed, preserving the preprint model's emphasis on timeliness over endorsement. These safeguards were developed in response to in the medical community about preprints' reliability, particularly after incidents of flawed COVID-19-related preprints on other platforms highlighting risks of in policy-influencing research. The process aimed to balance accessibility with responsibility, screening roughly 4-5 days per submission without impeding the core goal of free, open distribution of unpublished manuscripts. Early adoption was cautious, with endorsements from organizations like the underscoring the platform's commitment to mitigating harms while enabling breakthroughs in fields wary of bioRxiv's lighter-touch approach.

Expansion During COVID-19

During the , medRxiv experienced a marked surge in submissions, driven largely by the global urgency to share preliminary research on , clinical interventions, and responses. From its launch in June 2019 through September 2020, the server received 27,413 submissions, of which 21,862 (79.8%) were posted following screening; while monthly submissions grew steadily from 100-200 pre-pandemic, COVID-19-related papers accelerated this to peaks exceeding 3,000 per month by mid-2020. This represented a roughly 10-fold increase in overall submissions between January and May 2020 alone, with COVID-19 topics comprising up to 80% of posted preprints at the height of the crisis. The influx necessitated adaptations in operational capacity, including enhanced screening protocols to handle the volume while upholding safeguards against , though the server prioritized rapid posting for non-controversial submissions to facilitate timely scientific discourse. By December 2020, annual postings reached approximately 14,290, with 8,858 (62%) focused on , reflecting the platform's role in disseminating over 16,000 such preprints cumulatively through (59.5% of total output). Usage metrics mirrored this expansion, with abstract views escalating from tens of thousands monthly pre-2020 to millions during peak periods, underscoring medRxiv's emergence as a key venue for sciences preprints amid journal backlogs. Post-acute phase, COVID-19 preprints declined to about 7% of submissions by 2023, yet the pandemic-induced growth persisted in non-COVID domains, solidifying medRxiv's infrastructure for sustained higher volumes and broader adoption in clinical research dissemination. This expansion highlighted both the platform's utility in crisis response and challenges in maintaining quality amid rapid scaling, with only 0.17% of preprints withdrawn overall, including some COVID-related retractions due to errors or ethical concerns.

Transition to Independence

On March 11, 2025, medRxiv transitioned from operation under (CSHL) to independent governance as part of the newly established non-profit organization openRxiv, which also oversees . This shift marked the maturation of the preprint servers, originally launched as experiments at CSHL, into a self-sustaining entity focused on long-term viability in . CSHL co-founders John Inglis and Richard Sever described the change as a "natural evolution," emphasizing that while the laboratory supported the platforms' mission, their growth necessitated separation to ensure continued rapid dissemination of research without institutional dependencies. The openRxiv structure establishes a researcher-led framework with a drawn from the , aimed at enhancing operational and broadening sharing in life and health sciences. This transition preserves core features like non-peer-reviewed posting with screening for policy compliance, while enabling expanded funding models, such as and partnerships, to support amid rising submission volumes post-COVID-19. CSHL retained no ownership or control, viewing the as aligned with the platforms' original goal of advancing scientific discovery through accessible preprints.

Purpose and Scope

Core Objectives

medRxiv's core objectives center on accelerating the dissemination of in the health sciences by providing a platform for sharing complete but unpublished manuscripts, known as preprints, prior to formal . This enables researchers to rapidly communicate findings on human health studies, including clinical trials and epidemiological analyses, where delays in traditional publishing can hinder timely application to challenges. By prioritizing speed, medRxiv addresses the limitations of journal-based systems, which often impose months-long review periods, thereby fostering earlier feedback and iteration in sensitive fields like . A key aim is to enhance the and of scientific findings, allowing researchers to preliminary without paywalls or subscription barriers. This promotes broader collaboration, as authors can solicit input from peers worldwide, potentially refining studies before submission to journals. medRxiv also seeks to document the of ideas, establishing timestamps for priority and reducing disputes over originality in fast-moving research areas. Unlike peer-reviewed outlets, medRxiv does not endorse content through validation but focuses on enabling discourse while implementing safeguards against , such as mandatory screening for ethical compliance, , and basic scientific plausibility. These objectives position medRxiv as a complementary tool to formal , emphasizing provisional sharing to advance without supplanting rigorous evaluation. The server specifically targets health-related research excluded from general preprint platforms, ensuring domain-appropriate handling of clinical and policy-sensitive topics.

Differences from bioRxiv and Other Servers

medRxiv specializes in preprints describing human health research, including clinical trials, epidemiological studies, and analyses, whereas focuses on broader biological sciences such as , , and . Clinical trials and studies with direct implications for patient care or policy are directed to medRxiv, with explicitly recommending submission of such papers to its medical counterpart since June 2019. Manuscripts cannot be posted to both servers, and dual submissions result in withdrawal to maintain distinct scopes. Unlike , which employs basic screening for , ethical compliance, and scientific content validity, medRxiv implements more rigorous pre-screening protocols to mitigate risks associated with unvetted , such as potential misinterpretation leading to harm in medical practice. This includes mandatory disclaimers on all medRxiv preprints warning readers that findings are preliminary and not peer-reviewed, with heightened scrutiny for studies involving human subjects or therapeutic claims. 's lighter touch reflects its origins in basic biological research, where immediate clinical risks are lower. Compared to general preprint servers like (physics and mathematics) or SSRN (social sciences), medRxiv's health-specific focus incorporates domain-expert review during screening to ensure alignment with standards, a feature absent in non-specialized platforms that prioritize rapid dissemination without field-specific safeguards. Other health-oriented servers, such as Research Square, may allow broader interdisciplinary uploads but lack medRxiv's integration with biomedical indexing like or its emphasis on excluding non-human health topics.

Operational Features

Submission and Screening Procedures

Authors submit preprints to medRxiv through the online Processing System at submit.medrxiv.org, where they create an account or log in via if available. The preferred method involves uploading a single PDF file containing the full , including , list with affiliations and corresponding contact, abstract, keywords, main text, references, and any figures or tables. Manuscripts must report original research or clinical protocols in the health sciences, excluding reviews, commentaries, or case reports, and authors are required to declare ethical approvals, funding sources, conflicts of interest, and—for clinical trials—registration details from a WHO-compliant registry such as . No submission fees apply, and authors retain copyright, licensing content under a Attribution-NonCommercial-NoDerivs (CC BY-NC-ND) license by default, though alternatives like CC0 or CC BY are permitted. Upon submission, preprints undergo a multi-step screening process rather than , designed to exclude offensive, non-scientific, or potentially harmful without evaluating scientific validity. The first step, conducted by in-house with scientific and expertise, verifies that submission matches the , confirms scope alignment with health sciences research, checks for via automated tools, ensures absence of identifiers, and confirms declarations of ethical oversight and registration where required. If discrepancies or violations are identified—such as unregistered trials, missing ethics statements, or out-of-scope material—the submission may be rejected or returned for correction; escalated cases involving potential risks are discussed by the team, founders, or external advisors. The second screening step involves medRxiv affiliates reviewing for broader compliance, including whether the content pertains to health-related and poses no risk of public harm, such as that could influence adversely. Screening typically completes in 2–4 days but may extend over weekends, holidays, or for complex reviews; successful preprints are posted online without , , or formatting changes, assigned a , and made publicly accessible immediately unless authors request a brief hold. Authors can submit revisions to existing preprints via the author dashboard, updating the version while preserving the original and prior iterations. This process prioritizes rapid dissemination while mitigating risks, as emphasized in medRxiv's disclaimers that preprints are preliminary and not suitable for guiding clinical decisions.

Content Policies and Restrictions

MedRxiv maintains content policies centered on screening submissions for relevance to sciences, ethical compliance, and absence of harm, without formal . Manuscripts must pertain to in or , excluding narrative reviews, commentaries, case reports, or non-research materials. Submissions require complete author details, ethical oversight declarations (such as approval or exemption), and identifiers where applicable; failures in these areas lead to rejection. The screening process, conducted by in-house staff and medRxiv affiliates, occurs in stages and typically spans 2–4 days. Initial checks verify submission completeness, appropriate article type, and scope alignment, while scanning for patient identifiers or via automated tools. Further evaluation assesses potential public harm, such as content that could discourage or measures, or promote unsafe practices; offensive, non-scientific, or previously published material is also flagged for rejection. Escalated cases involve consultation with a content team, founders, or advisors, but no embargoes or scheduled posting times are permitted. Once posted, preprints receive a and cannot be removed, preserving the archival record; authors may withdraw manuscripts if they retract findings, marking them as withdrawn while retaining public access to . Revised can be submitted at any time before journal acceptance, superseding prior iterations without altering the original posting. These policies aim to balance rapid dissemination with safeguards against or ethical lapses, though screening does not certify accuracy or novelty.

Technical Infrastructure

medRxiv operates as a web-based platform managed by , with submissions processed through an online portal that accepts PDF manuscripts along with associated metadata such as titles, authors, abstracts, and keywords. Manuscripts are required to be in English and formatted according to specific guidelines, including a maximum and exclusion of formats to ensure accessibility. Following submission, preprints undergo a non-peer-reviewed screening process by an editorial team to check for compliance with content policies, after which approved manuscripts are posted in the order of approval, assigned a unique prefix of 10.1101, and made freely accessible via the web interface without subscription fees. The platform supports versioning, allowing authors to upload revised manuscripts that supersede prior versions while preserving access to historical iterations, and facilitates withdrawals or corrections as needed. Storage and dissemination rely on a centralized server infrastructure that hosts preprints primarily as PDF files, with full-text versions automatically generated and added 48-72 hours post-initial PDF posting to enhance readability and machine readability since October 19, 2020. Search functionality is provided through integrated indexing of and content, enabling retrieval by keywords, authors, or subjects, while programmatic access is supported via / feeds in XML format for recent posts and third-party for querying archives, as evidenced by open-source tools like the medrxivr . The shared operational framework with , both under Cold Spring Harbor Laboratory's oversight until their transition to the independent nonprofit OpenRxiv in March 2025, implies a unified backend for handling high-volume traffic, with annual operational costs around $3 million supporting scalability for over 100,000 combined preprints across the servers. No public details disclose the specific software stack or cloud versus on-premise hosting, prioritizing reliability for global access over open-source transparency.

Usage Statistics and Growth

medRxiv posted 913 preprints in 2019 following its launch in June of that year. Volumes surged in 2020 to 14,070 preprints, driven primarily by the , which accounted for a substantial portion of submissions during this period. By the end of 2021, cumulative posted preprints totaled 27,674, with approximately 12,691 added that year. Post-2021, annual volumes stabilized at elevated levels compared to pre-pandemic baselines but below the 2020 peak. medRxiv's official reporting data indicate roughly 10,575 new preprints in 2022 (bringing the cumulative total to about 38,249 by year-end), followed by 11,046 in 2023 (cumulative 49,295). In 2024, postings approached 12,863, reflecting sustained growth in non-COVID topics amid a decline in pandemic-related research. The following table summarizes annual posted preprints:
YearPosted Preprints
2019913
202014,070
202112,691
2022~10,575
2023~11,046
202412,863
Early in the , preprints comprised up to 80% of medRxiv's output, but this proportion fell to 7% by 2024 as focus shifted to broader health sciences research. Monthly postings fluctuated between 700 and 1,400 during peak periods, with higher volumes in late 2020 (e.g., 1,399 in ). Overall, the platform's growth underscores preprints' role in rapid dissemination, though volumes reflect screening rates of about 83% of submissions.

Author and Download Metrics

As of early 2025, medRxiv had posted nearly 64,000 preprints from more than 380,000 authors. This reflects substantial growth since its launch in June 2019, with early contributions by June 2020 already encompassing 57,096 unique authors across 124 countries. Author participation surged during the , driven by urgent dissemination needs in health sciences, before stabilizing; in 2024, the server posted 12,863 preprints, approaching its 2020 peak volume. Download and view metrics underscore medRxiv's reach, with cumulative PDF downloads totaling 51,943,342 and abstract views reaching 132,900,392 as of December 2021. These figures capture exponential early uptake: in 2019, downloads numbered 107,772 alongside 241,528 abstract views; 2020 saw 27,963,915 downloads and 61,613,928 views, reflecting pandemic-related acceleration. Per-preprint metrics, including article views, PDF downloads, and attention scores (aggregating altmetric data), are provided to authors via the platform, though these are noted to underestimate full usage due to limited tracking of article-to-article shares. Post-2021 aggregates remain less publicly detailed, but sustained posting rates indicate ongoing engagement in clinical and public health research dissemination.

Impact on Scientific Communication

Acceleration of Research Dissemination

MedRxiv facilitates the acceleration of research dissemination in the health sciences by allowing authors to post preprints shortly after submission, following a screening process that typically takes 2-4 days. This contrasts sharply with traditional journal peer review, where the interval from submission to publication often exceeds six months, as evidenced by a median time of 6.0 months (interquartile range, 3.0-8.0 months) observed between medRxiv preprint posting and subsequent peer-reviewed publication for clinical studies. By bypassing these delays, medRxiv enables immediate access to unpublished findings, promoting rapid feedback, collaboration, and iteration among researchers. The platform's emphasis on speed proved critical during public health emergencies, such as the starting in early 2020, when medRxiv became a leading source for sharing breaking research on epidemiology, treatments, and vaccines. For example, randomized clinical trials addressing interventions were posted expeditiously, allowing global scientific communities to engage with results in real time and refine ongoing studies ahead of formal validation. This timely availability supported faster knowledge integration into clinical decision-making and policy formulation, with preprints enabling clinicians to access a broader range of trial outcomes promptly. Overall, medRxiv's model prioritizes the swift establishment of scientific priority and openness, documenting idea provenance while enhancing accessibility for diverse stakeholders in health research. Launched on June 5, 2019, by in partnership with and , it addresses the unique needs of clinical and fields where delays in dissemination can hinder timely responses to evolving challenges. Such acceleration has been recognized for driving quicker breakthroughs by allowing critical engagement with preliminary data before peer-reviewed certification.

Influence on Policy and Practice

MedRxiv preprints gained significant traction in shaping responses during the , where rapid evidence needs outpaced traditional peer-review timelines. The RECOVERY collaborative group's preprint, posted on June 22, 2020, reported that dexamethasone reduced 28-day mortality by one-third among hospitalized patients receiving invasive (from 41.4% to 29.3%) and by one-fifth among those receiving oxygen without ventilation. This finding prompted immediate policy shifts, including endorsements by the UK's and the World Health Organization's updated guidelines in July 2020 recommending corticosteroids for severe cases, contributing to dexamethasone's global adoption and an estimated prevention of over 3 million deaths by mid-2021 according to modeling analyses. Vaccine development similarly benefited from medRxiv's speed. Pfizer and BioNTech's phase 3 efficacy , released November 9, 2020, detailed 95% efficacy against symptomatic after two doses in over 43,000 participants, with adverse events comparable to trials. This informed use authorizations by the FDA on December 11, 2020, and equivalent bodies worldwide, accelerating rollout amid surging cases and enabling policies for prioritized of high-risk groups. Such instances illustrate how preprints facilitated causal evidence for interventions, bypassing delays that could exacerbate outbreaks, though final peer-reviewed confirmations largely aligned with preliminary results in these high-profile cases. Despite these successes, medRxiv's influence on practice has been tempered by inherent limitations, as the platform explicitly advises against relying on unverified reports for clinical decisions. In non-emergency contexts, adoption remains cautious; for instance, while 29.3% of articles in major journals like and NEJM in 2020 cited medRxiv preprints, subsequent analyses revealed discrepancies in 13.6% of matched publications, including changes in results or interpretations that could mislead if acted upon prematurely. Policy-makers and practitioners have thus increasingly integrated preprints with from multiple sources, mitigating risks of flawed data propagation observed in retracted works. Overall, medRxiv's role underscores a shift toward provisional in crises, enhancing causal responsiveness while demanding rigorous validation.

Reception and Criticisms

Positive Assessments

MedRxiv has been praised by proponents in the scientific community for facilitating the rapid dissemination of health sciences research, particularly during the COVID-19 pandemic, where it emerged as a key platform for sharing preliminary findings ahead of formal peer review. The Chan Zuckerberg Initiative recognized medRxiv's role by awarding it $2 million in June 2020, citing its status as a top source for breaking COVID-19 research that improved openness, accessibility, and collaboration among researchers. This funding underscored its value in documenting the provenance of ideas and enabling early feedback, which supporters argue enhances scientific progress without supplanting peer review. Empirical analyses have bolstered positive views by demonstrating substantial concordance between medRxiv preprints and their subsequently peer-reviewed publications. A 2022 in examined 1,006 medRxiv clinical studies from 2019–2021 that later appeared in journals, finding that 82.5% had identical or similar study characteristics, 85.9% matching results, and 87.4% aligned interpretations, suggesting preprints often provide reliable previews of validated work. Such findings counter concerns over prematurity, with researchers like those involved in medRxiv's development, including Harlan Krumholz, advocating for its embrace as a tool to accelerate knowledge sharing and teach discernment in evaluating emerging science. Institutional endorsements highlight medRxiv's growing importance in biomedical communication. A 2025 Nature article noted that the platform has "become so important" for preprint distribution that it warranted independent nonprofit to support expansion, reflecting researcher demand for swift, to findings. Similarly, reviews from library evaluators, such as Doody's in July 2021, described it as a "valuable resource" for and rapid research avenues, emphasizing its clarity in distinguishing s from peer-reviewed content. These assessments align with broader arguments that medRxiv promotes collaboration, establishes priority, and fills gaps in traditional publishing timelines, particularly in urgent fields like .

Concerns Over Quality and Reliability

MedRxiv preprints undergo no formal prior to posting, which can result in the dissemination of manuscripts containing errors, incomplete analyses, or . This absence of pre-publication scrutiny has prompted concerns that such content may mislead researchers, policymakers, and the public, particularly in high-stakes fields like clinical medicine where premature conclusions could influence decisions. Empirical analyses have identified discrepancies between medRxiv preprints and their eventual peer-reviewed publications. For instance, a of COVID-19-related studies found differences in titles, , or conclusions in nearly half of the preprint-published version pairs that were cited in leading medical journals. Similarly, among 67 COVID-19 studies from medRxiv that reached journal publication, only 34% showed no discrepancies in results reporting, while 22% exhibited at least one major discrepancy, such as changes in or effect sizes. These alterations often stem from methodological refinements or error corrections during , underscoring the provisional nature of preprint findings. Retractions of medRxiv preprints highlight additional reliability issues, with faster retraction timelines compared to peer-reviewed articles—median of 29 days versus 139 days—but frequently tied to ethical or procedural in 70% of downstream cases. Moreover, retracted preprints are commonly cited without critical acknowledgment; in one analysis of papers, 86% of citations to retracted works treated them as valid, potentially perpetuating flawed evidence in subsequent research. Critics argue that medRxiv's rapid posting model, while accelerating dissemination, amplifies risks in areas prone to or , such as pandemic response research, where unvetted preprints garnered outsized and policy attention before validation. Although many preprints align with final publications—e.g., concordant results in most clinical study pairs—the potential for influential errors persists, as not all submissions proceed to , leaving low-quality work in circulation. MedRxiv mitigates some risks through author attestations and post-publication withdrawal options, but these measures do not substitute for rigorous vetting.

Specific Controversies in Citation and Interpretation

A 2022 analysis of 205 research articles published in 2020 by The , NEJM, , and found that 60 (29.3%) cited at least one medRxiv , but 25.6% of these citations were deemed inappropriate due to mismatches between the cited preprint and its eventual peer-reviewed version, such as alterations in titles, , or conclusions. Nearly half of the cited preprints exhibited substantive differences upon publication, underscoring the hazards of relying on unverified preliminary findings. Moreover, medRxiv's server did not indicate links to published versions for 42.7% of these preprints at the time of , complicating efforts by citing authors. Critics have highlighted systemic flaws in preprint citation practices, including the frequent omission of disclaimers noting the absence of , which can foster undue authority for provisional results. In discourse, this contributed to interpretive disputes, as media and policymakers sometimes amplified claims without caveats, leading to public confusion when revisions emerged. For instance, abstracts of randomized trials on medRxiv often displayed ""—exaggerated interpretations of results—more frequently than their published counterparts, potentially biasing early citations toward overstated or signals. A prominent case involved (HCQ) for treatment. An April 2020 medRxiv preprint analyzing U.S. Department of Veterans Affairs data (n=1,868 patients) reported no reduction in mortality or ventilation risk with HCQ, alone or combined with , prompting the to suspend related trials on May 6, 2020. This , cited over 1,000 times by mid-2020, faced scrutiny for unadjusted confounders like disease severity and comorbidities, with subsequent analyses questioning its causal inferences due to reliance on electronic health records without . Proponents of HCQ argued the preprint's rapid uptake exemplified premature dismissal of a potentially repurposable , while opponents viewed it as evidence against off-label promotion amid evolving evidence; the debate persisted despite peer-reviewed confirmations of null effects in randomized trials. Ivermectin preprints on medRxiv similarly sparked interpretive conflicts. Multiple early submissions, including systematic reviews aggregating small trials, suggested prophylactic or therapeutic benefits against , fueling advocacy for widespread use despite methodological flaws like high risk of bias and inconsistent dosing. A 2021 medRxiv identified bias as a primary driver of positive results across trials, with low-quality studies overrepresented in favorable citations; later large randomized trials, such as one posted on medRxiv in June 2022 (n=1,432), found no meaningful symptom reduction or hospitalization prevention at standard doses. These discrepancies led to polarized interpretations, with some jurisdictions authorizing based on aggregates while regulatory bodies like the FDA cautioned against extralabel use, citing insufficient causal evidence from robust designs. Such cases illustrate how citations, absent rigorous , can amplify selective interpretations aligned with preconceived views rather than empirical consensus.