A literature review and reflection on clinical trials: current challenges, future directions, and potential strategies to overcome practical barriers

Introduction

Background

Clinical trials are at the top of the pyramid of evidence-based medicine and play a significant role in healthcare decision-making. In recent decades, there has been a remarkable increase in the number of registered and published clinical trials across countries, regions, and disciplines (1-6). However, only a small proportion of the tested regimens have been successfully applied in clinical practice. While there is undoubtedly an inherent chance in scientific investigations, increasing evidence indicates that there are multiple common problems in the design, implementation, analysis, and reporting of clinical trials. Among these problems, the following are particularly troubling: many clinical trials are unnecessary (7), a considerable amount of published clinical trials are of low quality (5,8-13), and serious publication bias (14,15) and result spinning (16-18) are prevalent, alongside worrisome retractions and integrity issues (19-21). For example, studies have found that 12.7% to 17.2% of randomized controlled trials (RCTs) included in meta-analyses were redundant (22). Even more strikingly, a substantial number of clinical trials were found to be unnecessarily duplicated, as high-grade Cochrane systematic review evidence already existed prior to their initiation (7). Admittedly, clinical trials undergo various time-consuming processes such as review, approval, and governance, and the resulting significant delays may partly explain such redundancy and duplication. However, this cannot fully account for the redundancy and duplication issues currently reported in clinical trials. Furthermore, published clinical trials that were of poor quality and contained avoidable design flaws have been reported across multiple specialties, with rates ranging from over 40% to nearly 80% (23-25). Strikingly, one study found that over 60% of retracted clinical research over a 10-year period were clinical trials, with data/data analysis errors (38.2%) being the most common reason and data fabrication accounting for over 90% within the cases (19). The VITALITY study further quantified the impact of clinical trial retractions on one aspect of the healthcare evidence ecosystem, revealing that the exclusion of retracted trials would alter the direction of effect in meta-analyses by 6.8% to 10.1% (20). Regrettably, these issues have emerged and persisted despite the existence of a series of policy initiatives governing clinical trials at multiple levels—including the World Health Organization (WHO), national and regional organizations, academic institutions, and journals (26-29). These concerns have led to renewed scrutiny of how RCTs are designed, conducted, and interpreted within evidence-based medicine, rather than a rejection of their foundational role (30). For clarification, in this review, trial redundancy refers to studies addressing questions already adequately answered by existing high-quality evidence, trial duplication denotes repeated testing of substantially similar interventions or hypotheses, research waste encompasses trials that are unnecessary, poorly designed, underpowered, or unusable, and low-quality trials specifically denote studies at high risk of bias due to methodological or reporting deficiencies.

Rationale and knowledge gap

Given these facts, some reviews have elaborated on how to conduct rigorous and reliable clinical trials and have outlined journal criteria for RCTs (31-34). There are also reviews that have reflected on some of the challenges mentioned and proposed recommendations for improvements (25,35,36). However, no comprehensive review has provided a broad picture of the current state of clinical trials, their landscape, registration, quality, peer review workflow, reporting, dissemination, and emerging trends. This review attempts to address this gap. In addition, it focuses on reflections from the evolving landscape of clinical trials and research over the last 5 years. Finally, this review goes beyond a mere literature review by offering critical insights and providing recommendations and practical strategies for addressing major concerns.

Objective

This review aims to review the current landscape of clinical trials, issues of transparency, quality, pre- and post-publication appraisal, reporting and dissemination, and emerging trends in clinical trials. In addition, we aim to propose solutions and attempt to offer practical strategies considering real-world challenges. Finally, we hope this review will function not only as a critique of current practice but also as a learning resource for clinicians, investigators, reviewers, and regulators.

In the following sections, we first outline key challenges across the trial life cycle, then discuss practical recommendations and potential implementation strategies, and finally integrate these into a unified framework for stakeholders. We present this article in accordance with the Narrative Review reporting checklist (available at https://actr.amegroups.com/article/view/10.21037/actr-25-132/rc).

Methods

In this review, we searched for relevant reviews, original articles, comments and perspectives that reflect on clinical trials. Using PubMed, we searched for articles published between January 1^st, 2021, and July 31^st, 2025, that reflected on trial quantity, quality, registration, protocol, risk of bias, barriers, reporting, dissemination, and up-to-date trends including artificial intelligence (AI). Only articles published in English were considered eligible for inclusion (Table 1). Although the primary search focused on 2021–2025 reflections, relevant studies published earlier were included to contextualize persistent systemic issues. A detailed search strategy can be found in Appendix 1. The literature screening process and corresponding inclusion details can be found in Figure S1.

We also retrieved data on registered clinical trials as of August 2025 from the International Clinical Trials Registry Platform, ClinicalTrials.gov, ChiCTR, ITMCTR, EU-CTR, JapicCTI, CTRI, and ISRCTN platforms, along with the number of published clinical trials in PubMed, Cochrane Library, Embase, Web of Science, and Scopus. Furthermore, we searched the EQUATOR website (https://www.equator-network.org/) for reporting guidelines listed as of July 19, 2025, and examined reporting guidelines related to clinical trials.

In addition, we reviewed all commentaries published in AME Clinical Trials Review between its inaugural issue in October 2023 and December 2025, extracting points where the commented clinical trial was mentioned as flawed and areas for further improvement, then compiling and quantifying the mention frequency. Specifically, the dimensions extracted include whether the commentary identifies deficiencies in clinical trials. If deficiencies are identified, it further examines whether the commentary addresses issues related to sample size, trial design, appropriateness of the control group, selection bias and population representativeness, trial bias, statistical analysis, outcome measures, follow-up, interpretation of results, the trial’s generalizability and clinical applicability, trial reporting, and other aspects beyond these dimensions. We also reviewed the interviews conducted with 12 experts who authored commentaries in the journal AME Clinical Trials Review. These interviews were carried out independently from our review and we did not participate in its expert selection or interview topic setting. This source of information serves as supplementary content from an expert perspective for our review.

Based on the above integrated information, we identified recurring challenges and specific aspects frequently discussed in clinical trials. Subsequently, we proposed targeted recommendations for each identified challenge. Following this, we conducted brainstorms and literature reviews to identify practical barriers. Finally, we proposed tailored potential implementation strategies for each identified barrier.

In this review, no AI or large language models were applied to the setting or refining of the search strategy, searching, choosing references, summarizing reference content, collecting data, analyzing data, outlining article, drafting articles, generating figures and tables, or generating references.

The landscape of clinical trials and regional variations

Situation overview and existing challenges

As of August 2025, preliminary searches across the International Clinical Trials Registry Platform, ClinicalTrials.gov, ChiCTR, ITMCTR, EU-CTR, JapicCTI, CTRI, and ISRCTN platforms revealed that the number of registered clinical trial records has now surpassed one million. And, as of August 2025, the number of published clinical trials and trial protocols across databases is consistently impressive among different databases—PubMed (960,000 and 30,000), Cochrane Library (1,000,000 and 120,000), Embase (520,000 and 110,000), Web of Science (300,000 and 200,000), and Scopus (1,000,000 and 210,000). Given the heterogeneity in data sources, definitions, and denominators, these registry counts and publication volumes should be interpreted as indicative trends rather than directly comparable or causally linked measures.

This high volume of clinical trials has been accompanied by marked variations in trial conduct and access across countries, characterized by deficiencies in equity, accessibility, and relevance to real-world needs at regional and national scales. Specifically, while clinical trials in oncology achieved an average annual absolute growth of 266.6 trials between 2000 and 2021, and low- and middle-income countries (LMICs) showed notable increases in early phase trials, the highest trial density remained in high-income countries (HICs). Moreover, 76.4% of countries had not initiated any new trials by 2024 (1). Among LMICs where trial volumes have grown significantly, it is hard to ignore the contributions from Asia. Between 2005 and 2018, the number of cancer clinical trials in four Asian countries (China, India, Japan, and South Korea) increased by more than twofold (37). Among Phase I tumor trials conducted in Australia, the proportion of Asian contributions surged from 6% to 39% over the decade from 2012 to 2022 (38). Moreover, while the United States remained the first contributor to the number of tumor clinical trials conducted globally (46.5% of all trials) during 2010–2019, with China ranking second (2), by 2023 China’s total number of trials had far exceeded that of the United States, which ranks first (3). The number of RCTs in surgery has also increased primarily in Asia, particularly in China; however, as of 2019, the countries publishing the most surgical RCTs remained developed nations such as Finland and the Netherlands (5). In contrast, Pakistan conducted only 2,273 trials between 2000 and 2022 (39). Despite the emergence of two major epidemics in Africa between 2010 and 2018, Nigeria’s prevalent measles and South Africa’s prevalent tuberculosis, neither disease was among the top ten clinical trials in those countries (40). Further evidence supports this observation. From 1997 to 2019, the number of RCT articles in MEDLINE increased from 10,360 to 22,384 in the top 10 countries in the Nature Index 2019 rankings. However, this substantial increase in RCTs was almost exclusively published in non-PubMed core clinical journals, meaning these trials appeared predominantly in journals with limited relevance to clinical practice (41).

The expansion of clinical trials has been accompanied by trial redundancy, waste, and quality concerns, with similarly significant regional variations. For instance, among the bariatric surgery RCTs registered on ClinicalTrials.gov from 2000 to 2022, 83.6% were found to be research waste, and 43.8% of these were published (42). One study found that during the decade from 2010 to 2021, only two out of the 10 Association of Southeast Asian Nations member nations had over 50% of their trials meeting high quality standards (43). Approximately 50% of trials conducted in China, India, Japan, and South Korea from Asia, between 2005 and 2018, had a small sample size of only one to 50 participants that can hardly guarantee sufficient power (37). Trials led by LMICs demonstrated larger median effect sizes than those in HICs (44). In contrast, a study that analyzed two decades of surgical RCTs found that the 10 countries generating surgical RCTs with the best methodological standards were all in Europe, where trials exhibited the highest proportion of low risk of bias (30.5%), with the United Kingdom ranking first (5). In addition, likely due to the high quality of United Kingdom clinical trials, the citation rate of trials originating from United Kingdom institutions is higher than that of other countries, including the United States (45).

Consequently, more newly approved drugs and bioagents are based on fewer critical trials or less rigorously designed trials (46). It is estimated that 80% of key clinical trials that were the rationale for approving novel cancer treatments from 2017 to 2021 in China had flaws in their design, conduct, or reporting (47). Among RCTs of drugs in oncology that led to FDA approval between 2009 and 2021, 55% included comparisons within trials that used a design favoring the test group, raising concerns over whether the test group truly demonstrated superiority (48). In key trials underpinning European Medicines Agency approvals for new oncology drugs from 2014 to 2016, 49% of primary outcomes exhibited high risk of bias (49).

In short, the global growth of clinical trials has not translated into global equity, and redundancy is common.

Key recommendations on the challenges

To address the above-mentioned clinical trial landscape and regional variations, we propose two key recommendations. For clarification, the following recommendations are presented as normative proposals informed by the accumulated evidence reviewed above, rather than as direct empirical conclusions derived from any single dataset.

• Managing the uncontrolled expansion of clinical trials, minimizing redundant trials and redirecting efforts toward quality assurance and critical trials that match local needs.
• Trials initiated by HICs should consider the inclusion of both developed and developing regions in multicenter trials to fully reflect the global burden of disease, population diversity, and accessibility of trial results in resource-constrained settings. Trials of dose reduction, efficiency, and practicality in developing regions should be prioritized (50).

Barriers and potential practical strategies

LMICs face various obstacles in conducting clinical trials. For instance, it has been reported that Pakistan suffers from funding deficits, weak infrastructure, shortages of technical personnel, absence of regulatory bodies such as local registries, low literacy rates hindering informed consent processes, and cultural and religious barriers (39). However, 49 trial sites across 21 LMICs in sub-Saharan Africa, South Asia, Latin America and the Caribbean, the Middle East and North Africa, and East Asia and the Pacific have established infrastructure, research personnel, ethics and recruitment, and data management services, but they lack trial network development (51). Although regional variations do exist, a survey of clinicians with clinical trial experience in LMICs identified shared factors affecting trial implementation, including difficulties in obtaining trial funding (78%), lack of dedicated research time (55%), and insufficient training (91% of trainees and early career researchers rated this as moderate or significant impact) (50). Regarding funding, research indicates that RCTs in HICs are more likely to receive industrial funding than those in LMICs (73% vs. 21%) (44). Correspondingly, the most frequently mentioned strategies for addressing these obstacles were building capacity and human resources (46%), fostering national political will (29%), and strengthening funding opportunities or infrastructure (21%) (50).

Thus, the following practical strategies may be considered to address these practical barriers (political will, funding, human resources, training, research time, and supervision).

• Recognizing the importance of initiating clinical trials in LMICs and joining trials initiated by HICs to address the overall health burden in LMICs.
• Providing policy support, including initiating trial opportunities and grants to address gaps in local clinical needs through proactive efforts by the governments and local institutions. Such efforts can supplement commercially sponsored trials, avoiding excessive reliance on pharmaceutical companies. Additionally, fostering a friendly policy environment for conducting trials initiated by HICs is necessary.
  Enhance funding diversity through multiple sources, including government grants, international charity funding, and actively identifying funding gaps through collaboration with commercial companies.
• Offering more multifaceted pathways for nurturing and developing human resources within the healthcare system, including the establishment of specialized qualification standards for clinical trial professionals.
• Enriching more flexible training, such as developing customized courses for specific regions, has proven highly effective (52), and making greater use of freely available online training resources.
• Building upon point 3, a collaborative network and mechanism between trial professionals and clinicians should be established to address time shortages.
• Building institutions to monitor clinical trial implementation, with a focus on reviewing trial rationale and necessity prior to implementation.

Transparency in clinical trials

Situation overview and existing challenges

To enhance the transparency of clinical trials, many international and regional regulations have been implemented over the past two decades. These include prospective registration, publication of trial protocols on registration platforms, statistical analysis plans (SAP) provided on registration platforms, submission of ethical approval documentation, and publication of trial results on publicly accessible registration platforms (27,29,53,54). For example, in 2017, the WHO established 11 best practices for clinical trials, including prospective registration, keeping registration records updated, providing trial results on the registration platform within 12 months, providing the protocol on the registration platform within 12 months, publishing trial results in journals, including trial identification numbers in published studies, publishing trial results in an open-access manner, examining the principal investigator’s past reporting records, monitoring trial registration records, monitoring trial result reporting, and ensuring public disclosure of trial results (29). The European Union Commission Guideline 2012/C 302/03 and Regulation (EU) No. 536/2014 require sponsors to disclose registered trial results to the European Medicines Agency via the EUCTR platform within 12 months of the trial’s completion. For pediatric clinical trials, this deadline is shortened to 6 months, and trial results must be accompanied by a public-facing summary of the trial findings (53). The U.S. Food and Drug Administration and the United Kingdom Medicines and Healthcare Products Regulatory Agency have similar regulations. Moreover, most trial registration platforms—particularly those designated as primary registries within the WHO network—explicitly require registrants to provide proof of ethical approval prior to trial registration (29). Furthermore, evidence has demonstrated the value of prospective registration for clinical trials (55,56) and the effectiveness of policy mandates in encouraging registration (57). For example, the 2007 FDA Amended Act mandated clinical trial registration and results reporting on ClinicalTrials.gov. A comparison of trial registrations before and after the act’s release between 2005 and 2014 revealed that trials were more likely to be registered following the act’s implementation (57).

Although these policies have improved clinical trial registration, it remains suboptimal. Key issues include continued unsatisfactory registration rates, severe deficiencies in prospective registration, and high rates of non-disclosure of results after the registration. For example, clinical trial registration rates were found to be 69% for nutrition-related trials (58), 33.7% for orthopedic trials (59), and 31% for multidisciplinary trials (55). Additionally, prospective registration has been demonstrated by multiple studies to be effective in reducing trial bias risks (55,56), while retrospective registration has been found to exhibit over six times the incidence of publication bias (60). However, a study analyzing 1,177 clinical trials revealed that only 36.7% of the registered trials were prospectively registered (55). Furthermore, the disclosure rate of trial results is poor. The disclosure or publication rate of reported trials across various disciplines ranges from 18.4% to 78% (61-67), with trials with a significant primary outcome showing a higher disclosure rate (68). Taking pediatrics—where clinical trials are inherently challenging—as an example, research found that between 2007 and 2016, 16.54% of 1,088 RCTs involving children were not completed after registration, with many failing to explain why the trials were abandoned (66). Among the 908 completed trials, 44.49% had no data disclosure available. The consequences of this are significant. In pediatrics, due to the unique nature of pediatric treatment regimens and challenges in recruiting trial participants, many drugs used for pediatric treatment are not specifically approved for children and are instead administered off-label (69). The non-disclosure of completed trial results, whether in peer-reviewed journals or trial registries, may lead to an incomplete risk perception of drugs, potentially triggering adverse events that could have been mitigated or avoided. Although some of these data originate from a single journal (59), involve small sample sizes (64), or are derived solely from one platform (65) and thus may not represent an entire discipline or all trial platforms, the aggregated findings from multiple studies collectively reveal an unsatisfactory picture.

The availability of clinical trial protocols and SAPs has also improved, although it remains suboptimal. Key features include continuing relatively low accessibility rates, weak willingness among trial correspondents to provide protocols and SAPs, and significant risks associated with protocols being published only in journals without disclosing the original protocol/amended protocol, and revision timelines on registration platforms. Clinical trial protocols are primarily disclosed in three ways, in order of release timing are: as files on trial registration platforms, as standalone peer-reviewed publications, and as supplementary materials to an official trial publication. One study analyzed the availability of protocols from 347 RCTs across the three methods and found that protocol accessibility increased from 36% in 2012 to 66% in 2016 (70). However, another systematic review analyzed 600 randomly selected RCTs on stroke and transient ischemic stroke from the Cochrane Database between 2008 and 2020, finding that only 5% had a protocol and only 1% had a SAP (71). Similarly, a random sample of 400 RCTs in the field of nutritional interventions revealed protocol and SAP availability rates of only 14% and 3%, respectively (58). Furthermore, after identifying 201 trials where 48% lacked publicly available protocols or SAPs, researchers sent up to four emails to trial correspondents; however, 76% did not respond, 7% declined to share protocols or SAPs, and 8% of emails were invalid (72). Moreover, among the three accessible ways to obtain trial protocols and SAPs, publishing protocols (whether as supplements to trial publications or as standalone journal articles) offered no significant additional benefits compared to directly disclosing original protocols on trial registration platforms from the beginning. Specifically, this study analyzed 589 RCTs published in journals with the highest impact factors in internal medicine. It was found that trials that publicly disclosed their protocols upon publication exhibited discrepancies of up to 33.1% between the protocols and the primary outcomes or sample sizes of the trials themselves (73). Although this study retrieved only RCTs published within a relatively short timeframe, another analysis of 4,754 trial protocols spanning a decade also identified similar findings (74). Another study analyzed clinical trials related to acupuncture anesthesia registered with the ICTRP from 2001 to 2023 in 21 countries. The findings revealed poor consistency between registered protocols and published articles, including 39% inconsistency in sample size, 36.6% inconsistency in blinding methods, and 24.4% inconsistency in secondary outcomes (75).

Furthermore, both the quality of clinical trial registrations and registration platforms have been found to have a considerable potential for improvement. Key manifestations include substantial variations between platforms, even though a set of core items are established by WHO (https://www.who.int/tools/clinical-trials-registry-platform/network/who-data-set/archived/1-3), resulting in significant gaps in critical data and difficulties in the retrieval of information. A study analyzing trials registered on ChiCTR found that the platform addresses numerous key dimensions such as study type, registration date, institution type, location, sample size, interventions, study design, funding sources, application of randomization and blinding, and number of research centers. However, it also overlooks critical areas including trial update progress, data management, ethical disclosures, and assessment of registration content quality (76). Additionally, researchers have identified significant discrepancies between registrations on the International Clinical Trials Registry Platform and ClinicalTrials.gov across multiple aspects, including institutional review board approval, study phase, location, design-related issues, and data-sharing plans (77). In addition, researchers retrieved trials from ClinicalTrials.gov and Clinical Trials Registry-India to identify studies conducted in India. However, the current registration platform interfaces are very challenging, making it difficult to ensure data accuracy, completeness, and transparency of the data. Authors had to piece together information from multiple sections, such as locations, baseline measurements, brief summaries, detailed descriptions, sponsors, inclusion/exclusion criteria, design groups, and other fields, to attempt to make an informed judgment (78).

Key recommendations on the challenges

• Prioritizing strengthened oversight of mandatory registration for trials funded by non-pharmaceutical sponsors.This is because, unlike trials funded by pharmaceutical companies, much of the funding from non-pharmaceutical sponsors comes from the public, including tax revenues, and thus needs to maximize service to the public. However, a mandatory requirement imposed by the sponsors is likely to be the most direct and effective approach, as trial implementation inevitably involves funding issues. Furthermore, research indicates that the recent improvements in protocol accessibility have been primarily driven by industry-sponsored trials (70). However, numerous non-pharmaceutical sponsor institutions (such as universities and government funding agencies) have not similarly adequately implemented the WHO’s best practices for clinical trials. For example, policies and monitoring practices within the 21 major non-pharmaceutical European research funding agencies were found to fail to meet the WHO’s best practices for clinical trial transparency: 66.7% require prospective registration, 28.6% require trial results to be published on trial registries 12 months after trial completion, 42.9% actively monitor whether trials are registered, and 38.1% monitor whether results are disclosed (28). Research also evaluated open documents from 25 major medical research funding agencies in Europe, Oceania, South Asia, and Canada. The results revealed that these 25 non-pharmaceutical institutions implemented an average of only 49% (5/11) of the WHO’s best practices. Only 24% (6/25) of funders considered the principal investigator’s prior reporting record. The United Kingdom’s National Institute for Health and Care Research was the only funder that adopted all 11 policies, whereas some institutions did not obtain any points across all 11 policies (79) (Figure 1).
  Additionally, it is worth noting a paradox: while our review, approval, and governance of clinical trials aim to ensure ethical compliance, reliability, and translational potential, these lengthy processes may conversely cause significant delays that could lead to trial redundancy and duplication. Trial registration, particularly enhanced oversight of such registrations, becomes increasingly vital. This enables global trials to be publicly accessible and traceable at an earlier stage, significantly reducing current redundancy and duplication.
  

  Figure 1 Adoption level of WHO’s 11 best practice criteria among the world’s largest 25 public or charity healthcare research funders. ‘YES’ indicates mandatory policies. ‘NO’ indicates complete absence of policies. ‘NO*’ indicates non-bounded policies (encourage but do not mandate). ‘NO#’ indicates incomplete policies (policies only apply to certain trial types). This figure is reused under the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0). Original source: O’Riordan M, Haslberger M, Cruz C, et al. Are European clinical trial funders policies on clinical trial registration and reporting improving? A cross-sectional study. J Clin Transl Sci 2023;7:e166. WHO, World Health Organization.
• For industry-sponsored trials, a prioritized ranking of sponsors should be considered to encourage better registration. The AllTrials.net initiative publicly ranks sponsors based on their performance in trial reporting, encouraging better registration of industry-sponsored trials by enabling sponsors to avoid reputation damage (80).
• Making it mandatory for trials to disclose their original protocols and SAPs upon registration on the platform, rather than waiting until subsequent publications.
• Optimizing the registration platform, including establishing minimum standards for core registration information led by the WHO and implementing user-friendly structural adjustments to the platform. For example, clearly defining essential information as drop-down options while reserving manual entry fields for unusual fields to accommodate design variations; apply adaptive tuning to mandatory fields for each stage based on trial progress logic: once trial registration status changes to ‘Initiated’, the Ethics Number field becomes a mandatory entry. When the trial period reaches one year after initiation, the system automatically sends an email to the registered email address reminding to update the trial status and supplement trial outcome data. Upon changing the trial status from ‘In Progress’ to ‘Completed’ in the drop-down menu, the ‘Trial Outcome’ section automatically becomes a mandatory field.

Barriers and potential practical strategies

There are many practical barriers to more transparent clinical trials, and two key obstacles are listed here. First, there is a lack of execution and a reward-and-punishment framework for policy compliance. For instance, a study evaluating the outcomes of 1,093 rheumatology RCTs found that among journals claiming to support the recommendations of the International Committee of Medical Journal Editors (such as requiring prospective registration of clinical trials), 16% of trials published in 2022 lacked prospective registration (81). Due to the persisting lack of transparency in clinical trials, some scholars argue that complete trial transparency is an unachievable goal unless regulatory institutions, journal editors, and trial sponsors can further expand requirements and actually implement them (82). Second, there are knowledge gaps and time constraints. After interviewing authors of 1,093 RCTs, the reasons given for retrospective registration included: being unaware of the need for prospective registration; believing that only drug-related or industry-sponsored trials required prospective registration; being unaware that their study met the definition of a clinical trial; and knowing registration was required but facing time constraints (81).

In response to these obstacles, four practical strategies are recommended below, and an integrated scheme is visualized in Figure 2.

• A feasible structural reform would be to require public trial registration before ethics review, thereby ensuring transparency from the outset. Specifically, ethics committees could require a registry identifier and public protocol summary as part of the submission. Thus, trial initiators should first complete registration before applying ethical certification/trial approval. Additionally, ethics review/trial approval processes should verify whether the trial has been registered and whether the original protocol and SAP are attached to the registration platform.
  Although the WHO’s current 11 best practices require ethical approval to be obtained prior to registration—meaning ethics precedes registration—there is a major practical flaw. This is because ethical approval and trial approval from local regulatory agencies carry greater practical compulsion; if a trial lacks ethical approval or local regulatory approval, it cannot proceed. Moreover, once a trial receives ethical approval, subsequent registration is not as strictly enforceable in practice. This recommendation is supported by evidence. Israel implemented a prerequisite requirement in 2016 mandating local registration on platforms prior to ethics review, with ethics committees directly verifying trial registration status during review. Consequently, trial registrations significantly increased from 2016 to 2021 (83). In India, despite legal mandates requiring approved trials to register, audits revealed that 12% of 381 approved trials had no registration records at all (84). Furthermore, another study randomly sampled registration records for 188 COVID-19-related RCTs and found that only 41% fully disclosed ethics information. This issue was not isolated to a single registration platform; both ChiCTR and ClinicalTrials.gov exhibited particularly pronounced deficiencies. This persists despite current WHO best practice requirements mandating ethics approval prior to registration (27).
• Establishing rewards and penalties for compliance, strengthening monitoring, and conducting public audits. This can be achieved by incorporating trial registration, ethical approval, and the recording of ethical approval numbers on registration platforms, as well as the public disclosure of protocols and SAPs on trial registration platforms, and the publication of trial results, into the academic integrity assessment of trial sponsors and the journal quality evaluation of those publishing corresponding trials. Alternatively, it can be implemented through administrative pressure at the institutional level.
  For example, trial transparency can be categorized as excellent, adequate, or inadequate based on the completeness, timeliness, consistency, disclosure, and adherence to standards of trial registration (76), thereby reflecting the academic integrity of trial correspondents. For journals publishing clinical trials, journal evaluation agencies (such as Web of Science, PubMed, and other regional journal assessment bodies) could incorporate the publication of trials that fail to meet these transparency standards into their journal quality assessments. This would motivate journal editors to take concrete action and enforce compliance. Another example demonstrates that administrative pressure at the institutional level has proven effective in ensuring trial result disclosure: In January 2019, the House of Commons Science and Technology Committee issued a letter to United Kingdom universities urging compliance with reporting requirements for clinical trials in medicinal products by June 2019. By June 2021, the proportion of reported trial results increased from 29% to 91%. Five universities achieved 100% reporting compliance, and all 20 universities reported over 70% of required trial results on the European Clinical Trials Register (85).
• To address knowledge gaps, enhancing outreach and education on clinical trial transparency is essential. For instance, clearly defining clinical trials and specifying requirements for trial registration and prospective registration should be publicly disseminated across multiple platforms—including trial registration portals, national/regional trial review agency websites, and author and reviewer instructions in journals. Detailed, concrete requirements should be provided rather than vague phrasing such as ‘adhering to WHO best practices’ or ‘International Committee of Medical Journal Editors guidelines’.
• To address the time-consuming issue of registration, platforms should consider more user-friendly drop-down designs or developing AI tools to assist registration. For example, first upload the trial protocol and SAP. Then, allow the AI tools equipped on the registration platform to automatically read the files and extract key matching information. The matched content can automatically fill in the corresponding fields within the registration system. Finally, registered users can directly verify the accuracy and make modifications. Alternatively, a one-stop platform integrating trial design and registration could be actively developed. This would enable trial correspondents to input the framework structure provided by the platform during the design phase. Once the trial design is finalized (including the protocol and SAP) and saved for confirmation, these systems could automatically archive the documents. After author final verification, a single click would initiate the registration process, completing the initial registration. There are undoubtedly many more solutions. In short, platforms need to make greater efforts to provide more user-friendly and time-saving registration alternatives.

Figure 2 Proposed scheme for optimizing trial transparency. AI, artificial intelligence; SAP, statistical analysis plan.

Quality and integrity of clinical trials

In this section, we summarize common threats to trial quality, outline existing tools and standards, and then propose a more integrated, stage-based approach to safeguarding quality across the trial life cycle.

Situation overview and existing challenges

The quality of clinical trials is the key to determining whether they can truly stand at the pinnacle of evidence-based medicine. We know that, according to the GRADE evidence quality grading system, while randomized trials are assigned a default high-quality rating, issues such as risk of bias can lead to downgrading of their evidence quality (86). Currently, due to numerous quality and integrity issues in clinical trials, scholars advocate that we must maintain a more cautious and healthy skepticism toward trials (30).

This caution is partly driven by the concerning quality of many current clinical trials, characterized by high risk of bias, avoidable design flaws, statistical errors and inadequate sample sizes/power, with results that are fragile. For example, using the RoB 2.0 tool to assess risk of bias in oral and maxillofacial RCTs revealed that 50% showed high risk of bias, 36% raised some concerns about bias risk, and only 14% were classified as low risk of bias (87). Specifically, a study evaluating 840 RCTs on pre-neoplastic gastric lesions published between 2011 and 2021 found uncertain risks of up to 98.93%, 98.69%, and 100% for allocation concealment, blinding of participants or panelists, and blinding of outcome evaluators, respectively (88). And this has been observed in acupuncture RCTs conducted in Japan since the 1990s: apart from improvements in specific areas like sequence generation, the remaining five domains assessed by the Cochrane RoB tool do not seem to have improved (89). For trial design, the identified issues manifest in multiple dimensions. Avoidable design errors in trials had high rates across various domains—weight-loss surgery RCTs at 48.6% (42), dermal scar RCTs at 48.9% (24), vascular RCTs at 68.4% (90), and gastric cancer RCTs at 77.8% (23 trials) (23); not using overall survival as a trial endpoint was found to be associated with a high rate of proportional hazards violation, with approximately 23.8% of phase III superiority clinical trials demonstrating proportional hazards violation (91); RCTs with sequential primary or secondary endpoints published in leading journals were found to have little assured adequacy in such designs (9); 68.7% of antifungal non-inferiority RCTs up to 9 September 2020 failed to justify the use of non-inferiority designs (11), whereas 12.9% of superiority RCTs published in 2021 in the five medical journals with the highest impact factors should have been designed as non-inferiority trials (92); Adaptive designs, increasingly favored by researchers in recent years, have also been found to have similar design issues, including interim analysis timing and frequency, consideration of outcomes for interim analysis, and unplanned adjustments (93); Insufficient consideration may exist in trial protocols from the very beginning—a cross-sectional appraisal of phase III clinical trial records on ClinicalTrials.gov revealed that 77.7% lacked a complete description of randomization methods (10). For the statistical dimension of trials, issues are also multifaceted. A study that examined the appropriateness of statistical methods in acupuncture RCTs found that as many as 70.7% of trials employed suboptimal statistical methods rather than using intention-to-treat analysis or modified intention-to-treat analysis (8); A study that examined trials using repeated measurements of continuous variables as primary outcomes found that 94% of trials used inconsistent statistical methods for sample size calculation and primary analysis, 25.8% used incorrect formulas when adjusting sample size calculations for loss-to-follow-up, and 49.5% of re-calculated sample sizes were found to be larger than the reported sample sizes (12); Analysis of limitations across 120 RCTs revealed that the most commonly self-reported limitations involved: small sample size (47.5%), duration and follow-up restrictions (33.3%), and inadequate controls (32.5%) (94); About half of phase III clinical trials registered on ClinicalTrials.gov by May 26, 2023 did not reach an 80% power level for sample size calculations (13); Re-calculating the sample sizes for 242 RCTs across 44 Cochrane systematic reviews revealed that approximately 90% of trials either overestimated or underestimated the required sample size (95). The fragility index (FI) represents the minimum number of non-events required to reverse the effect direction toward the intervention or control group, thereby shifting the outcome from statistical significance to non-significance. A larger FI indicates greater robustness and reliability of trial results. For FI, clinical trials have been found to exhibit very low FI in multiple domains: a median FI of 3 in colorectal cancer trials (96), 2 in spinal surgery (with FI often lower than the number of lost to follow-up) (97), 4 in hip arthroscopy RCTs (98), and 4 in COVID-19 trials (with over half have an FI of less than 1% of each sample size) (99). Most paradoxically, we typically assume randomization ensures baseline comparability between treatment and control groups in controlled trials. However, a study reviewing 142 RCTs found that among the 84.5% claiming baseline comparability, 57% failed to report how balance was assessed, and post-hoc calculations revealed that 34.5% of these trials exhibited severe imbalance in at least one baseline variable between groups (100).

Integrity is another area of concern in clinical trials beyond trial quality, characterized by issues such as poor transparency, data fabrication or errors, and high retraction rates due to integrity concerns. While trial transparency has been thoroughly discussed earlier, it bears repeating that transparency itself constitutes a vital component of academic integrity. Researchers who analyzed over 6,000 RCTs up to November 2021 for quality review found various credibility and integrity issues throughout the entire RCT lifecycle, with the most common challenges being ethical and transparency concerns (21). Additionally, an analysis of 209 initial drafts of RCTs submitted to an anesthesia journal revealed fabricated data as the most prevalent cause for concern, occurring in a high incidence of 82% (101). Although data from this one journal may not be representative of other journals, such a high rate of data falsification in initial drafts is deeply alarming. Moreover, instances of fabricated or problematic data persist at a similarly high rate in RCTs that proceed through peer review and are ultimately published. One study evaluated the reasons for retractions of clinical research published in PubMed-indexed journals between 2012 and 2022. Among 242 retracted research articles, 62% were clinical trials. The most common reasons for retraction were data or data analysis errors (38.2%), plagiarism (11.8%), duplicate publication (11.1%), ethical issues (7.3%), and methodological flaws (7%) (19). The VITALITY study analyzed RCTs retracted for any reason on Retraction Watch as of November 5, 2024, finding that 70.8% of retractions were data-related, while 12.9% were labeled as paper mills (20). The retraction of clinical trials serves as a corrective mechanism for problematic studies, yet even when retracted, the negative consequences of flawed trials may not be fully mitigated. The 68 conclusions distorted by problematic trials identified in the VITALITY study have been incorporated into 157 clinical guidelines (20), and one study found that the net effect of retractions due to integrity issues in clinical trials is not zero but negative after fitting the data (102).

Both the quality and integrity issues discussed above indirectly reflect that the current peer review process for clinical trials may be sub-optimal. As we know, the traditional quality control system involving journal editors and invited reviewers is critical. One study analyzed 209 initial submissions of RCTs in the field of anesthesia to journals and found that falsified data were the most common cause for concern, occurring in 82% of cases (101). This fully justifies the importance of editors and peer reviewers, as identifying problematic trials often depends on individual patient-level data, which is frequently difficult to obtain after publication. However, it seems we have not fully utilized this advantage, resulting in a less satisfactory situation. For example, a survey of 1,733 reviewers found that only 34.3% checked trial registration websites, while some respondents expressed a desire for editors to independently verify whether trials were registered and to check for discrepancies between registration details and manuscript content prior to sending manuscripts to peer reviewers (103). A study evaluated peer review reports for parallel RCTs across 62 open-access journals published by BioMed Central. Sampling 26 RCTs yielded 59 review reports, revealing that these comments were short (median length of 276 words, with some even consisting of only one sentence), superficial (rarely addressing conclusions, protocol deviations, or data accessibility), and rarely constructive (only 19%). Furthermore, 96.6% of reviewers failed to mention whether they had reviewed the corresponding trial protocol or registration information (104).

In contrast, a great potential exists for quality review of manuscripts at three distinct stages: prior to submission, during the preprint phase, and after formal publication. Regarding pre-submission review, the establishment of independent central review, separate review and monitoring of trial protocols (whether by institutional centers or sponsor-designated bodies), and the creation of data monitoring committees show clear necessity. A comparison of 13,928 pediatric RCT records on ClinicalTrials.gov during 2008–2021 revealed that trials with data monitoring committees more frequently terminated trials for scientific data-related reasons (105). And third-party data monitoring may be preferable, despite its challenges, as studies have found that in phase II oncology clinical trials, local researchers significantly overestimated overall response rate data compared to paired blinded reviewers (106). The National Cancer Institute requires a scientific review of clinical trial protocols at designated cancer centers, in addition to review by Institutional Review Boards. Analysis of reviews conducted by the Protocol Review and Monitoring Committee at the Harold C. Simmons Cancer Center from January 2009 to June 2013 revealed that the proportion of trials requiring modifications reached 27% for industry-sponsored trials and 54% for investigator-initiated trials (IITs) (54). Similarly, the Bill and Melinda Gates Foundation established a Data Analysis Committee for their funded clinical trials, dedicated to conducting scientific reviews of the trial protocols. A retrospective analysis of Data Analysis Committee scientific review feedback for 52 trial protocols from 2020 to 2022 revealed a total of 1,537 comments (36). The median number of comments per protocol was 28, with the most frequently addressed areas being data collection, outcome measurement and endpoints, endpoint analysis, adjustment analysis, other analyses, sample size and power, design, inclusion/exclusion criteria, randomization, statistical simulation, and intervention information (Figure 3). These align with the most frequently identified issues in problematic trials after publication. Regarding preprint-stage and post-publication peer review, a study analyzed comments from systematic reviews, post-preprint discussions (from PubPeer, medRxiv, Research Square, etc.), and post-publication peer review comments. The analysis revealed that systematic review evaluators identified methodological and reporting issues in 89% of RCTs. Additionally, 15% of trials received comments on at least one preprint server or PubPeer, and post-preprint and post-publication comments identified significant problems in 9% of trials (25). Among these concerns: research design issues accounted for 46%, while others included incomplete reporting (39%), errors (31%), sample size issues (23%), statistical analysis problems (19%), applicability of results (18%), inconsistencies in methods analysis and reporting (16%), spin (14%), ethical issues (12%), and inability to access raw data or protocols (7%) (Figure 4). Similarly, an analysis of 1,037 preprint reviews found that comments on preprint servers were comparable in their content to those reviewed through formal peer review (107). Even in retraction analyses, concerns raised about 150 clinical trials retracted between 2012 and 2022 included 72.61% originating from editors or chief editors, while the remaining 27.39% originated from other sources, including authors, readers, and post-publication peer reviewers (19). Journals such as AME Clinical Trials Review specialize in post-publication commentary and review of clinical trials, publishing critiques of published trials and review articles reflecting clinical trials. We conducted a retrospective analysis of all articles published in the journal from its inaugural issue in October 2023 through the December 2025 issue. After excluding commentaries addressing post-hoc analyses, subgroup analyses, or secondary analyses of trials, and removing review articles, we identified 184 commentaries targeting 161 original trials (https://actr.amegroups.org/issue/archive). These commentaries spanned multiple specialties including lung cancer, gastric cancer, colorectal cancer, esophageal cancer, breast cancer, cardiovascular diseases, hematologic malignancies, and prostate cancer. Among the 161 trials, 69 (42.9%) were identified as having issues that could have been resolved prior to publication by the commenting experts. Specifically, the four dimensions mentioned most frequently among these 69 trials were small sample sizes, lack of confidence in trial design dimensions such as group allocation/blinding design and implementation/center considerations/endpoints, poor generalizability and applicability, and inadequate control of trial bias and confounding (Table 2). Furthermore, we interviewed authors of these commentary manuscripts (https://actr.amegroups.org/post/category/interviews-with-outstanding-authors). Among the 12 experts interviewed, 8 (66%) re-emphasized their concerns regarding trial quality issues. Among these, the interpretation of trial results and the frequent neglect of quality-of-life concerns in trial outcome measures are particularly common. For instance, Dr. Toyoaki Hida, Dr. Takehiro Uemura, and Dr. Jiaxin Niu all raised concerns about result interpretation in trials, including overinterpretation of subgroup analysis results and insufficient attention to long-term follow-up data and long-term safety. Dr. Savvas Lampridis, Dr. Takehiro Uemura, and Dr. Igor Gómez-Randulfe all emphasize that quality of life and patient-reported outcomes are areas frequently overlooked in current trials.

Figure 3 Dimensions of trial issues identified during the Data Analysis Committee’s review of trial protocols before trial initiation. This figure is reused under the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0). Original source: Burford B, Norman T, Dolley S. Scientific review of protocols to enhance informativeness of global health clinical trials. Trials 2025;26:85.

Figure 4 Dimensions of trial issues identified through preprints, post-publication peer review, and systematic review comments. This figure is reused under the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0). Original source: Davidson M, Korfitsen CB, Riveros C, et al. Post-publication peer review and the identification of methodological and reporting issues in COVID-19 trials: a qualitative study. BMJ Evid Based Med 2025;30:233-40. PICO, Population, Intervention, Comparison, Outcome.

Although, as noted above, the current quality and integrity of clinical trials remain a great concern, we are not entirely without experience. Over the years, we have accumulated considerable lessons and have increasingly prioritized these issues, introducing many requirements and review tools to enhance clinical trial standards (32-34,108-120) (summarized in Table 3). It is evident that the standards or tools we use to evaluate the quality and integrity of clinical trials are no longer our primary concern. The challenge we now face is how to establish a widely accepted, validated standard. As Bordewijk et al. noted, ‘apart from the methods to locate textual plagiarism and image manipulation, all other methods, be it theoretical or empirical, are based on examples, are not standardized, and lack formal validation’ (121). Beyond this, the next challenge we face is how to overcome practical obstacles to ensure these tools are used.

Key recommendations on the challenges

In response to these challenges, four key recommendations are listed below, and an integrated scheme is visualized in Figure 5.

• Quality control for clinical trials demands front-loaded efforts, strengthening pre-submission evaluation. The focus should be on independent review of trial protocols and SAPs by the trial sponsor or a third-party review body, ensuring quality control from study conception.
  This is because we currently place excessive emphasis on peer review of clinical trials while neglecting peer review of clinical trial protocols. If protocols were mandated to undergo pre-trial peer review, just like the trials themselves, it would significantly enhance trial quality. And, as mentioned earlier, only a small portion of protocols are published—that is, undergo peer review. Many clinical trials either do not upload protocols during registration or have protocols that have not undergone peer review. Considering certain specific circumstances in clinical trials, it may not be necessary to publish all protocols. Instead, mandatory review mechanisms could be added during the registration, funding, or approval stages.
• Consider incorporating preprint reviews, post-publication reader comments, post-publication letters, post-publication commentaries, and systematic reviewers’ critiques of trials into the clinical trial review pipeline. Establish an ecosystem that enables linkage and feedback for corrections across all quality control stages of clinical trials.
  The quality control requirements for clinical trials demand more than individual action, but rather an era requiring structural reform. World Health Assembly resolution WHA75.8 (2022) has called for ‘a strengthened global architecture for a coordinated and high-quality clinical trials’ (122). Notably, studies indicate that the median time interval between preprint release and receiving comments is 10 days (interquartile range 2–65 days) (25), which closely matches typical peer review delays. Before final acceptance, editors could ensure authors have fully addressed feedback from both formal and preprint peer reviews. Of course, such integration can enhance research quality and encourage broader scientific participation to proactively identify abnormalities in data or expose potential research fraud. However, it is often perceived as lacking accountability and, when conducted anonymously without formal discussion, may be labeled as ‘lynching’. Therefore, establishing a central coordination and oversight mechanism is essential to prevent discriminatory criticism and unethical conduct. Additionally, post-publication comments, letters, commentaries, and systematic reviews related to clinical trials often remain scattered and fragmented. These opinions have not been systematically brought to the attention of trial authors or journal editors for further corrections. For instance, PubMed’s practice of linking commentaries to the original articles they critique and labeling them with ‘comment on’ and ‘comment in’ represents a promising approach. Similarly, Clinical Trials Review Database (https://ctrd.amegroups.com/) from AME Publishing Company serves this purpose by linking reviewed trials with their corresponding commentaries. However, these approaches still lack a mechanism for providing feedback that enables the original trial authors and the editors of the journal where the trial was published to be informed and make ongoing corrections. It is necessary to explore how to design feedback templates to streamline communication and revisions between reviewers and authors/original journal editors. Perhaps this could be achieved by testing the integration of such evaluations into digital platforms that facilitate the revision process—meaning that publication is not a one-time event but rather a continuous process of public oversight and ongoing correction.
• During the editorial and peer-review quality control process, develop structured review guidelines for clinical trials, including items editors must verify and specific dimensions that peer-review reports must address.
  For example, establish editorial review requirements for clinical trials by incorporating core dimensions addressed by key tools such as TRACT and INSPECT-SR (32,33,111,112,115) into routine review items. These include verifying trial registration status, checking for significant discrepancies between registries and manuscripts (cross-referencing registry information + original registered protocol + published protocol + article text for thorough verification), checking if the actual sample size is less than 85% of the planned size, and confirming whether endpoints have been altered. This enables effective labor division, reserving more specialized reviews for peer experts. Regarding peer review, for instance, authors could be mandated to report FI when reporting trial results. Key issues frequently identified in Figures 3,4Table 2, along with essential tools like JBI (108), RoB 2.0 (109), are incorporated into mandatory peer review feedback items. These could include whether randomization was genuinely ensured; whether there were significant deviations from the intended intervention; whether sample sizes were too small or power issues existed; whether concerns about bias were present; whether major flaws existed in the design or statistical methods; and whether data were falsified.
• For integrity reviews, consider implementing a ‘one-to-all’ review mechanism. Upon identifying one trial with integrity concerns, initiate a systematic review of other trials conducted by the same authors, with the journal editor and the authors’ institutional affiliations playing key roles.
  For example, a systematic review of all RCTs by a particular researcher, initiated after one of their RCTs raised concerns, revealed integrity issues in 14 RCTs involving 1,405 participants (123).

Figure 5 Proposed collaboration and ecosystem of quality control and integrity inspection in clinical trials across stages. SAP, statistical analysis plan.

Barriers and potential practical strategies

There are many practical barriers to better quality and integrity in clinical trials. Here is a discussion of four key obstacles. First, there are barriers related to inadequate proficiency among those conducting trials and reviewers at various stages, coupled with limited training resources. A study surveyed 120 expert members of the World Stroke Organization across 42 countries regarding their understanding of non-inferiority RCTs. A significant 62% reported limited knowledge of the study design (124). A study involving 260 participants from seven countries—ranging from medical students to professors—asked them to identify discrepancies in data charts within given articles. It found that 95.3% of discrepancies were omitted, with 62% of participants failing to detect any differences. Participants who spent more time on the task noticed more discrepancies, and those with more experience publishing papers also identified more discrepancies (125). Although 50% of participants in this study were medical students and only 31% held senior academic titles, it is important to note that participants were explicitly informed of data errors and asked to identify them. Without such prompts, the detection of discrepancies might be even lower. Additionally, a retrospective analysis of online training materials on peer review identified 42 training opportunities, with only 15% of the 20 open access resources covering clinical trials (126). Second, the barriers include insufficient human resources and limited time among stakeholders involved in various stages of trial quality control and integrity checking. Third, there may be structural institutional barriers to integrating preprint comments, post-publication reviews, and systematic reviews into the peer review pipeline, as well as establishing linkages and feedback correction ecosystem. Fourth, trials in specific specialties or trial types face unique professional challenges. For example, challenges in surgical RCTs involve variations in defining interventions and executing them across centers (127).

In response to these practical obstacles, the following strategies are recommended.

• To address issues of insufficient competence and limited training resources, efforts should be intensified to develop more open-source training resources and establish reviewer training mechanisms through journals or professional institutions. Successful examples already exist in this regard, such as the Web of Science’s reviewer training platform Publons Academy and the online open-source training resources provided by various publishers. Of note, training around non-inferiority designs, complex interventions, and risk-of-bias assessment etc. may be prioritized as these are recurring themes in many specialties.
• Regarding the limited human resources and time available for reviewing trials, incorporating preprint comments into the review system as suggested above can itself alleviate some of the tension on human resources and time. Additionally, when inviting reviewers, busy senior reviewers could be permitted to cultivate core team members by having younger team members conduct initial reviews, while more experienced and busier senior reviewers (those invited by editors) perform the actual quality control. This approach would increase the pool of future reviewer candidates and address the time constraints faced by senior reviewers. To ensure review quality, incentive and oversight mechanisms could be established, such as requiring mutual signatures on review comments or making review comments open access. This manuscript will later discuss in detail the significant progress made by AI in peer review, which could partially alleviate human resource and time constraints. However, it is essential that both editors and reviewers implement manual checks and oversight when utilizing AI for trial review, to prevent information leaks and incorrect comments.
• Regarding structural institutional barriers, consider having authoritative, well-established, trusted, and influential institutions or organizations take the lead in facilitating collaboration. For instance, the International Committee of Medical Journal Editors at the journal level, the Publons platform that already possesses robust reviewer and review comment databases, the PubMed platform provided by the National Institutes of Health, and PubPeer as a non-governmental alternative are all potential desirable leading parties. Crossref, a widely used tool and platform across various journals, has taken a promising first step by incorporating review reports from different stages and contributors into its content types and linking them to the articles under appraisal (https://www.crossref.org/documentation/schema-library/markup-guide-record-types/peer-reviews/).
• For specific professional barriers, stakeholders should develop more detailed criteria for operations and evaluations tailored to specific challenges. For instance, in surgical clinical trials, the Surgical Technique Reporting Checklist and Standards can be used to define and report surgical interventions more accurately, reducing variability between centers (128). The American Heart Association’s specific standards for designing, conducting, and analyzing cardiac surgery trials also represent an excellent initiative and model (129).

Reporting and dissemination of clinical trials

Situation overview and existing challenges

Regarding the reporting of clinical trials and their protocols, we have the CONSORT and SPIRIT reporting guidelines (https://www.consort-spirit.org/). Both sets of guidelines have been updated multiple times, with the latest version being the 2025 edition. Moreover, as of November 2025, there are 30 extensions to SPIRIT and CONSORT, covering how to report clinical trials with various study designs, different types of data, and diverse intervention approaches (https://www.consort-spirit.org/extensions).

However, reporting of clinical trials remains unsatisfactory across multiple specialties, trial types, and data sets, with variations observed across different levels of journals and regions. Across different specialties, CONSORT usage or reporting compliance was reported at only 32.11% in periodontal trials (130) and 54.3% in COVID-19 treatments (131)—even when we have the CONSERVE 2021 reporting guidelines (132) specifically tailored to accommodate trial reporting adjustments deemed extenuating under emergency circumstances like COVID-19. Only 51.6% of stroke clinical trials reported trial registration numbers, while 60.7% exhibited selective reporting (133). In sepsis research, as many as 91.6% of intervention trials failed to report outcomes (134). Across different trial designs, non-inferiority trials were found to have significant room for improvement in reporting (135). Even among expansion trials published in high-impact-factor journals, adherence to CONSORT-DEFINE remains unfavorable (136). Similarly, compliance with the corresponding CONSORT extension reporting guidelines among stepped-wedge cluster randomized trials from eight high-impact-factor journals was disappointing. For instance, 35% of trials failed to clearly display the duration of each time period in figures as required, and 30% omitted mention of the stepped-wedge design type in titles (137). Regarding reports on various data, a systematic review analyzing 2,025 RCTs found significant discrepancies between abstracts and full texts—particularly in results and conclusions—which is highly misleading since many readers primarily review abstracts rather than the full text (138). Conference abstracts of clinical trials were also found to score an average of only 17.6 points out of 50 (139), although this data was only for RCT abstracts from the European Orthodontic Society Congress. The reasons for this might be partly attributed to the moderate to low compliance found with the CONSORT-A extension guidelines for clinical trial abstracts (140). Similarly, the CONSORT-HARMS extension guidelines for harms were found to be used in only 25% of applicable trials, with zero studies citing the CONSORT-HARMS 2004/2022 statement (141). Disclosure of conflicts of interest in trials was even worse, with only 34% of trials found to adequately disclose conflicts of interest of funders (142). Across different levels of journals and regions, 497 trials published in NEJM, Lancet, JAMA, and BMJ generally complied with the CONSORT 2010 statement, achieving an average compliance rate of 90%. However, reporting on allocation descriptions, concealment, and implementation of randomization still requires improvement (143), particularly in journals from less developed regions. In a study from Iran, only 43.2% of CONSORT items were reported (144). Moreover, the reporting quality of trial protocols and their adherence to the SPIRIT guidelines are similarly no better. For instance, 75.5% of SPIRIT items were found to be unreported in orthodontic trial protocols (145), while the median number of items compliant with SPIRIT-PRO—an extension guideline for patient-reported outcome trials—was only 7.5 (146).

It is noteworthy to emphasize separately that among the challenges analyzed in numerous articles regarding clinical trial reports, spin reporting is mentioned with high frequency. For example, among pilot surgical trials conducted from January 1, 2011, to December 31, 2021, spin reporting was identified in 81.9% of studies, with similar proportions observed in abstracts (80.6%), results (81.9%), and discussions (80.6%) (16). Spin reporting in trials involves not only omitting statistically non-significant outcomes but also introducing new significant primary outcomes (147). Among 162 RCTs published in the 20 highest-impact-factor anesthesia journals, 40% of statistically non-significant trials exhibited misleading spin reporting in abstract conclusions, with spin descriptions reaching as high as 89% (17). Spin reporting in statistically non-significant RCTs was even higher in plastic surgery trials, reaching 85% (148). Visual abstracts were not immune to this problem either. A study randomly sampled RCTs from 15 high-impact-factor journals and found that 57% of visual abstracts for trials with non-significant primary outcomes exhibited spin reporting (149). Regarding spin reporting terminology for negative trial results, an analysis of over 500,000 RCTs found that the overall prevalence of such phrasing remained stable over the 30-year period from 1990 to 2020. The most common expressions were ‘marginally significant’, ‘all but significant’, ‘a nonsignificant trend’, ‘failed to reach statistical significance’, and ‘a strong trend’ (18). The underlying reason is that the academic community still tends to publish statistically significant trial results. Studies have found that statistically positive trials are more likely to be published than negative ones (14,15,150) and are published more quickly (150), whereas an estimated half of trial results are never published, primarily due to ‘negative results’ (151). This reflects an excessive reliance on and misunderstanding of the 0.05 statistical significance threshold among researchers, clinicians, editors, and reviewers. In fact, the 0.05 P value threshold is highly dependent on sample size, leading to a high probability of false positives or P-hacking in trials (152). Correspondingly, scholars have suggested abandoning the common P value threshold and associated terminology like ’statistical significance’, instead judging the clinical relevance of trial results in context.

Why do reporting issues persist in clinical trials and trial protocols despite the existence of numerous reporting guidelines? At least two key reasons have been identified. One issue is the underuse of reporting guidelines. A study that surveyed over 100 journals found that only 46% of journals mention the use of reporting guidelines when inviting peer reviewers to evaluate their manuscripts (153). This reflects the ambiguous and unclear recommendations for adopting CONSORT guidelines by journals. A study that searched the author instructions of 165 high-impact-factor journals found that only 38% of journals mentioned the CONSORT statement in their author instructions, and only 37% explicitly required its use (154). Furthermore, 63% of journals provided insufficiently clear recommendations—failing to explicitly state whether its use was required—and only 47% of editors responded that they had incorporated the CONSORT reporting guidelines into their editorial processes. Another reason is that a large number of newly released reporting guidelines impose an overwhelming workload on researchers. With many newly published guidelines lacking evidence of their effectiveness, they have garnered limited endorsement from both editors and researchers. As of July 19, 2025, the EQUATOR website (https://www.equator-network.org/) lists 672 published reporting guidelines. Among these, 17% are extensions of existing guidelines, with the most common extensions being for the CONSORT guidelines on clinical trial reporting. However, a study interviewing 150 researchers found that participants perceived using reporting guidelines as burdensome and complicated, while remaining uncertain about the actual benefits of adhering to them (155). Although this study did not specifically target the CONSORT or SPIRIT reporting guidelines, it reflects to some extent the significant imbalance between the workload imposed on researchers by current reporting guidelines and the perceived effectiveness of these guidelines in achieving their intended purpose. Additionally, a study interviewed journal editors, with 60% indicating they would be more inclined to consider adopting reporting guidelines in the future if there were more evidence on their effectiveness, including evidence evaluating the effectiveness of the CONSORT series of expanded reporting guidelines (156).

Issues surrounding clinical trials and their protocol reporting also extend to dissemination towards the public, including inadequate accuracy, information bias, lack of standards for an objective and fair lay summary, and systemic structural deficiencies that hinder public engagement and access to trial results. An analysis of YouTube videos created by specialized doctors, dentists, and medical students found that 27.6% were rated as low quality (157). Although the sample size was small and the data did not specifically relate to clinical trials, it highlights the issue of low-quality information dissemination on social media. As clinical trials have a greater impact on social media compared to other research types (158), they have become increasingly important in public information dissemination. Research has found that among 262 clinical trial results disseminated on social media, trials with positive outcomes yielded higher Altmetric scores, indirectly motivating mass media to bias reporting toward positive results (159). Additionally, while the CONSORT 2025 guidelines require a plain language summary, currently we do not have standardized requirements for preparing such summaries (160). Similarly, the access to trial information for the general public is also difficult on registration platforms. A cross-sectional analysis of disclosed participant information sheets on ClinicalTrials.gov revealed that 22% of trials did not include any mention of disseminating trial results to participants, whereas 41% of trials used only the same brief standardized text (four sentences) to inform participants about the trial (161).

Key recommendations on the challenges

• More research should be conducted on the effectiveness of CONSORT and SPIRIT extension reporting guidelines, and this evidence should be widely disseminated to enable researchers to recognize the potential benefits of using these guidelines and to encourage journal editors to truly integrate reporting guidelines into their editorial and peer-review processes.
• Introducing more explicit and specific mandatory requirements from journals for the use of CONSORT, SPIRIT, and their series of extension reporting guidelines.
  As early as 2001, research has confirmed that the publication of CONSORT guidelines correlates with improved quality of RCT reporting (162). Subsequent studies have repeatedly demonstrated that journal policies mandating authors to use reporting guidelines—whether through compulsory language in author instructions or requiring completion of a reporting checklist during manuscript revision—can enhance the quality of research reporting (163,164).
• Reducing focus on statistical significance thresholds and encouraging reporting of effect sizes, P values, and 95% confidence intervals, as well as emphasizing the clinical relevance of statistical differences identified in trials.
• Developing standardized and more accurate requirements for plain language summaries and figure abstracts of clinical trials, and establishing a structured pathway for public participation and access to accurate trial results.

For example, uploading standardized lay summaries to trial registration platforms to create a user-friendly public portal that better integrates with social media information, which could facilitate more accurate trial information access for patients, their families, journalists, and others (165). Alternatively, proven AI tools might be employed to optimize the accurate dissemination of trial information to the public. Research has developed the Trial Promoter tool to test the accuracy of its generated public dissemination messages for clinical trials. Results showed that during a 10-week test period across Twitter and Facebook, 97.7% of the 525 social media posts generated were accurate (166). Another example is Zannad et al., who proposed a more detailed systematic approach to ensure rapid and accurate dissemination of trial results among the general public. This includes encouraging more journals to offer publication options for plain language summaries and figure abstracts (167).

Barriers and potential practical strategies

The above recommendations may encounter several key practical obstacles. First, continued increases in the number of reporting guidelines persist alongside a lack of evidence evaluating their effectiveness. Second, traditional journals may face pressure to publish clinical trials that fail to meet statistical significance thresholds for reasons of journal profits or impact. Third, more compulsory requirements for CONSORT, SPIRIT, and related extensions—along with their integration into peer review processes—may pose time-related challenges for authors, editors, and reviewers.

In response to these obstacles, the following three practical strategies are recommended.

• The EQUATOR Network serves as the centralized collection of reporting guidelines, currently operating in its 1.0 era—the collection phase. The next phase could consider transitioning to the 2.0 era—the evaluation or screening era. Since there are currently sufficient reporting guidelines for clinical trials, what is lacking are good, effective, high-quality reporting guidelines. The EQUATOR Network could consider establishing a dedicated evidence section evaluating the quality and effectiveness of reporting guidelines to motivate guideline developers and evaluation researchers to conduct more evaluative studies.
• Consider establishing more journals specifically dedicated to publishing ‘negative’ clinical trials that do not meet statistical significance thresholds.
  This would foster a more rational academic culture regarding statistical significance thresholds and alleviate concerns among traditional journals about declining impact factors when publishing such results. It must be emphasized that such journals would also require robust methodological standards, including prospective registration, protocol and SAP availability, and adequate sample size. Furthermore, the proposal for establishing these journals is intended to provide a credible platform for publishing well-designed, neutral studies, rather than a repository for rejected manuscripts.
• Reasonably using AI tools to detect the completeness of clinical trial and protocol reports to improve efficiency, while allowing researchers, editors, and reviewers to focus more time on information verification.

Emerging trends in clinical trials

Situation overview and existing challenges

Emerging trends in clinical trials are numerous, with three primary areas discussed here: registry-based randomized controlled trials (RRCTs), IITs, and AI. Unlike traditional RCTs, which typically employ strict inclusion criteria, RRCTs support research by capitalizing on registries, thereby reflecting real-world conditions more accurately, enhancing generalizability, and expanding recruitment sources. Compared to RCTs, RRCTs employ significantly broader inclusion criteria and encompass multiple intervention modalities. By embedding randomization within registries, RRCTs combine the advantages of randomization with large-scale registry data, offering benefits such as reduced financial costs, enhanced efficiency, straightforward design, practicality, strong external applicability, data linkage, and addressing the gap where traditional RCTs are primarily funded by pharmaceutical companies (168). Currently, studies in Europe and the United States have demonstrated the success of RRCTs in this regard. However, RRCTs also face significant challenges, including ethical/informed consent issues, patient withdrawal, and regulatory boundaries (169). Moreover, despite increasingly widespread support for RRCTs among patients, clinicians, and research coordinators (170), concerns about data quality issues (e.g., missing data, measurement errors, recording bias, coding errors) have been raised, limiting the full realization of registries’ potential and corresponding advantages (170). Moreover, there exists significant variation in the definition and understanding of RRCTs among trial researchers, and education surrounding this methodology remains inadequate (171). Meanwhile, while doctors and trial coordinators acknowledge the advantages of RRCTs, they perceive them as less exciting than RCTs. RRCTs also entail more burdensome research management requirements and face recruitment challenges posed by commercially sponsored RCTs competing with RRCTs (169).

IITs are clinical trials initiated and conducted by researchers from academia or research institutions. Their emergence primarily aims to address commercially insufficient scientific questions and generate applicable solutions for clinical practice, differing from trials sponsored by pharmaceutical companies that focus on commercial value and product launches. IIT can also mitigate, to some extent, the current situation where businesses hold an absolute advantage in trial numbers, influence independent decision-making by intervention scholars (such as whether to publish trial results), and engage in academic misconduct like ghostwriting. For example, a review of trial publications by Portuguese authors found that industry-sponsored trials primarily focused on neurology, while IITs were more prevalent in gastroenterology and infectious diseases. These trials targeted different populations and pursued distinct objectives. The authors called for enhancing the capacity of national clinical research teams to mobilize the independence and value of trials distinct from commercial industry trials (172). Phase I clinical trials in the United Kingdom, Germany, France, Spain, and Italy were found to have significantly lower proportions of non-profit trials compared to industry-sponsored trials. In Italy, non-profit Phase I trials accounted for less than 20%, while industry-sponsored trials exceeded 80% (173). In Hungary, industry-initiated trials constituted as much as 97% of all trials approved in 2012 (174). Moreover, research indicates that industry sponsors influence trial outcome reporting through multiple ways, including prematurely halting trials and withholding results of discontinued trials, owning and controlling data access, and negotiating clinical trial agreements in multicenter trials where investigators’ publication rights are not fully protected (175). This research system creates a dependency on funding from industry sponsors, which may weaken the ability of researchers and institutions to negotiate terms that fully protect publication rights with these sponsors. Moreover, although industry-sponsored trials generally come with professional medical writing support and are found to have better compliance with CONSORT (176), there are many potential pitfalls involved, including the high risk of ghostwriting. For instance, the author has attended numerous academic conferences sponsored by pharmaceutical companies. At some events, companies directly allocate time to promote their writing services under the title of ‘writing support’. However, discussions with fellow attendees revealed that some companies actually provide direct ghostwriting services, constituting serious academic misconduct. Moreover, the author has reviewed job postings from pharmaceutical companies in China, where positions such as medical writer explicitly involve drafting trial manuscripts—not merely providing support as advertised in meetings. Amsterdam et al. also provided case studies detailing corruption in pharmaceutical industry-funded clinical trials: ‘medical ghostwriting and data misrepresentation’, ‘key opinion leaders conducting drug promotion under the flag of ‘medical education’, and collusion by journal editors failing to uphold scientific and peer-review standards (177). However, IITs also face their own challenges. IITs have been found to exhibit lower quality compared to industry-sponsored trials. For instance, in the National Cancer Institute (NCI)’s scientific review of clinical trial protocols, 27% of industry-funded trials required design modifications post-review, whereas this proportion rose to 54% for IITs (54). Additionally, IITs have been found to lack robust scientific review and monitoring systems compared to trials sponsored by pharmaceutical companies, particularly in less developed regions including India. Significant issues also persist regarding adverse event reporting and participant compensation (178). In this regard, China took a significant step forward in 2024 by enacting management regulations for IITs (https://www.medicalresearch.org.cn/login). These regulations establish the main responsibilities of various institutions, a detailed system for scientific review, ethical review, institutional project initiation and completion, and the uploading and disclosure of research information. They also strengthen classification management, prohibit meaningless duplicate research, and enhance overall research efficiency. Furthermore, they establish a systematic regulatory framework for administrative and technical oversight, featuring quality management, an approval system, and information retention periods extending to over 10 years.

AI represents another promising emerging direction in clinical trials. This field is advancing rapidly, accompanied by a mixed landscape of both pros and cons. This includes AI-intervention trials being found to have poor performance in quality, reporting, and transparency, alongside the immense potential and room for development in using AI to address the numerous issues mentioned earlier in this paper—including assisting peer review. Specifically, as of 2023, over 3,000 AI clinical trials had been registered on ClinicalTrials.gov, with more than half being randomized trials, primarily conducted in HICs (179). As of March 4, 2024, trials involving AI medical devices have been conducted in over 60 countries/regions, with 80% being interventional trials. The United States, China, and Canada rank as the top three in terms of trial volume. The primary areas of application include stroke, cognitive impairment, Alzheimer’s disease, depression, schizophrenia, and anxiety disorders (180). However, the quality of AI intervention clinical trials has been found to be concerning. A study evaluating AI trials in ophthalmology from 2010 to 2022 revealed that these trials all had a moderate risk of bias or raised some concerns (181). Reporting and transparency for AI intervention trials are equally concerning. Such trials were found to have a mere 53% compliance score with CONSORT-AI (181). At the journal level, an evaluation revealed that among the 52 journals publishing 65 AI intervention clinical trials, only 3 explicitly endorsed or mandated CONSORT-AI (182). Moreover, among AI trials systematically retrieved, only 15.8% of trial registration platforms publicly disclosed results—significantly lower than the 37.6% dissemination rate via peer-reviewed articles. In addition, 53.6% of articles were found to exhibit selective reporting, and 75% of trials failed to report adverse events altogether (183). Against these concerns are the immense potential for using AI throughout the entire clinical trial lifecycle. For instance, in enhancing trial efficiency and recruitment, preliminary AI systems built on natural language processing technology have been deployed to automatically match suitable clinical trials based on analysis of patients’ daily records, achieving nearly 90% specificity and nearly 80% accuracy (184). Meanwhile, TrialGPT can reduce screening time for trial patient recruitment by 42.6% (185). For trial information integration and retrieval, Trialstreamer (https://trialstreamer.ieai.robotreviewer.net/) automates the identification and categorization of RCTs by continuously monitoring PubMed and the WHO International Clinical Trials Registry Platform. This effectively addresses the issues of fragmented information across trial registries and missing links between registered trials and published articles (186). As of September 2, 2025, the platform contained 908,993 RCT records. More importantly, it claims daily updates, open-source availability, and free download access. Users can input Population, Intervention, Comparison, Outcome (PICO)-structured clinical queries, with results automatically ranked based on predicted risk of bias—prioritizing larger, higher-quality trials. Regarding risk of bias assessment in trials, RobotReviewer was developed to automatically evaluate the risk of bias in RCTs. Results showed that in evaluating 1,955 RCTs, its agreement rate with manual assessments (the gold standard involving two independent groups and two additional rounds of verification) ranged from 63.07% to 83.32%. Although the agreement rate still needs improvement—primarily in allocation concealment and consistency of outcome assessor blinding—it can be considered to have met the standard for a useful assistant tool (187). Regarding trial integrity reviews, researchers fed GPT-4o the TRACT checklist for evaluating trial trustworthiness, full trial PDFs, trial registration documents, and trial registration URLs. GPT-4o achieved an 84% agreement rate with human evaluators on 16 out of 19 TRACT items. However, its limitations include remaining semi-automated, requiring repeated user prompts, needing individual queries for each TRACT item (as batch-querying all 19 items is ineffective), still lacking access to certain subscription-based databases like Retraction Watch, and necessitating substantial manual verification and data security management (188).

Key recommendations on the challenges

• Conducting more educational outreach to raise awareness of RRCTs to increase patient enrollment rates, including providing monetary compensation to enrolled participants (169). AI tools can be used to further boost enrollment rates and shorten recruitment times.
• Registration authorities and RRCTs should establish sustainable funding models and develop cross-regional data governance collaboration mechanisms, including systematic data quality management systems (171). For education, prioritization may be considered for consistency and completeness of routine data, consent processes in pragmatic settings, and governance.
• Reinforcing support and management for IITs, including preserving scientific independence, scientific review, monitoring, quality control, and managing conflicts of interest.
• For AI-intervention trials, adopt the corresponding recommendations and implementation strategies outlined earlier to enhance transparency, quality, integrity, trial reporting, and dissemination.
• Detailed regulations should clearly specify where AI is used in the workflow, how systems are validated, and what human oversight is in place. For instance, the use of AI requires promoting and educating on core principles such as enhanced human oversight, safety management, and transparency, as exemplified by the FAIITH principles proposed at ASCO 2024 (https://www.asco.org/news-initiatives/policy-news-analysis/asco-sets-six-guiding-principles-ai-oncology).

The FAIITH principles include:

• Fairness and equity (AI developers and users must address bias in AI model design and implementation while ensuring equitable access to AI tools);
• Accountability (AI systems must comply with legal regulations and ethical requirements; AI developers bear responsibility for AI system decisions and adherence to legal, regulatory, and ethical standards);
• Informed stakeholders (when AI is used for clinical decision-making and care, patients and clinicians must be notified);
• Institutional oversight (decision-makers should establish structural compliance policies for AI management, including safeguarding decision-making autonomy and information privacy for clinicians and patients);
• Transparency (AI must maintain transparency throughout its entire lifecycle);
• Human-centered application (human-to-human interaction remains a fundamental element of healthcare. AI should not eliminate the need for human interaction or serve as a substitute for sensitive interactions).

Barriers and potential practical strategies

Regarding the above suggestions, the most significant practical obstacles seen by the author currently are the use of AI without disclosure or transparency, and the abuse of AI without any real human oversight, verification, or review. For instance, numerous manuscripts that utilize AI without any disclosure by the authors are identified in our daily editorial work. A study surveying 1,600 academics found that over 50% have used AI tools to assist with peer reviewing manuscripts (189). The American Association for Cancer Research announced findings at its 10th International Conference on Peer Review and Scientific Publishing in 2025, revealing that 23% of abstracts and 5% of peer review reports may contain text generated by large language models. However, fewer than 25% of authors complied with the requirement to report AI usage during submission. Additionally, we have observed at work that some reviewers return AI-generated feedback without any modifications, revealing that certain reviewers never even verify the AI-generated comments. Pangram Labs, a company specializing in detecting AI-generated content, conducted an exhaustive analysis of AI usage among reviewers at International Conference on Learning Representations (ICLR) 2026. Among the review comments for 75,800 papers, 21% were entirely AI-generated, 4% were heavily edited by AI, 9% were moderately edited by AI, and 22% were lightly edited by AI, with only 43% written entirely by human reviewers.

In order to address these significant practical obstacles, the following strategic recommendations are proposed.

• Developing tools with improved accuracy and detection rates for identifying AI usage. As AI systems learn and evolve, an increasing number of AIs will no longer rely on hard-coded language patterns but rather directly mimic human habits and phrasing. Correspondingly, detection tools must undergo efficient and timely updates and iterations.
• Cultivating a batch of specialized professionals who are responsible for managing AI across all stages of clinical trials, specifically tasked with manual oversight, manual verification, and manual review.
• Implementing AI usage standards and requirements at the legal level, and accelerating the refinement of legal provisions—including duty determination and penalty schemes—rather than merely relying on current core principles and ethical appeals.

Strengths and limitations

The strength of this review lies in a comprehensive overview of major challenges across the entire clinical trial life cycle, from trial registration to publication and dissemination. In addition, this review goes beyond a summary of existing literature by proposing potential solutions and practical strategies that account for real-world barriers, thereby facilitating real action toward resolving these challenges. Finally, this review may also serve as a teaching resource for journal clubs, methodology courses, and reviewer training.

This review also has its own limitations. First, as it is not a systematic review, there are certainly omissions in the literature, so the challenges identified may also be incomplete. Also, since there are too many components in the whole clinical trial life cycle, the review cannot fully cover each component in its entirety. Second, this review is based on reflective articles on clinical trials. Such articles may also suffer from publication bias, as papers highlighting problems like trial registration issues, quality concerns, or reporting deficiencies may be more likely to be published than those reporting the absence of problems. Consequently, certain issues may be overestimated or described with bias. Finally, although we have provided recommendations and solutions to the best of our ability while considering practical challenges and implementation barriers, it must be acknowledged that some suggestions may still prove difficult to implement or remain idealistic in specific countries or contexts. Even so, we have retained these recommendations. This is because, even if only some of them are adopted to some extent in specific countries or contexts, they still hold value in potentially improving the current situation.

Conclusions

In summary, while the number of clinical trials has surged dramatically, significant unnecessary redundancy and waste persist, compounded by global inequities in trial access. Strengthening oversight of trial quality, transparency, integrity, reporting, and information dissemination is now paramount. Even the rise of emerging fields like RRCTs, IITs, and AI has not yet addressed these issues well and has introduced new challenges. In conclusion, the key recommendations and potential strategies for practical barriers (Figure 6) may facilitate improvements if adopted and implemented. And, even modest, incremental changes along the lines we propose can help reduce waste and improve the informativeness of clinical trials.

Figure 6 Recommendations for major challenges in clinical trials and potential implementation strategies for practical barriers. AI, artificial intelligence.

Acknowledgments

The authors would like to thank Yijie Chen and Fangyuan Jiang for their assistance with copyright checking and table refinement. The authors applied AI tools DeepSeek and DeepL for language polish. The prompt used is ‘please help with English language polishing, ensuring that the original meaning is not altered, academic tone is maintained, and readability is improved’. The authors have reviewed and verified the accuracy of all AI-polished content word by word.

Footnote

Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://actr.amegroups.com/article/view/10.21037/actr-25-132/rc

Peer Review File: Available at https://actr.amegroups.com/article/view/10.21037/actr-25-132/prf

Funding: This research was funded in part through the NIH/NCI Cancer Center Support Grant (No. P30 CA008748).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://actr.amegroups.com/article/view/10.21037/actr-25-132/coif). K.Z. serves as the Editor-in-Chief of AME Clinical Trials Review and full-time staff of AME Publishing Company (the publisher of AME Clinical Trials Review). Y.P., B.S., Y.L., Y.C., and F.Y. are full-time staff of AME Publishing Company and serve as unpaid Associate Editors of AME Clinical Trials Review. S.L. serves as an unpaid Section Editor of AME Clinical Trials Review from August 2025 to July 2027. V.G.S. serves as an unpaid Section Editor of AME Clinical Trials Review from November 2025 to December 2027. C.B.S. serves as an unpaid International Advisory Board member of AME Clinical Trials Review from December 2025 to December 2027. I.N. and K.K.Z. serve as unpaid International Advisory Board members of AME Clinical Trials Review from November 2025 to December 2027. The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Izarn F, Henry J, Besle S, et al. Globalization of clinical trials in oncology: a worldwide quantitative analysis. ESMO Open 2025;10:104086. [Crossref] [PubMed]
2. Akiki V, Troussard X, Metges JP, et al. Global trends in oncology research: A mixed-methods study of publications and clinical trials from 2010 to 2019. Cancer Rep (Hoboken) 2023;6:e1650. [Crossref] [PubMed]
3. Lee EH, Lee S, Shin JI, et al. Registered clinical trial trends evolved differently in East Asia vs the United States during 2014-2023. J Clin Epidemiol 2025;183:111791. [Crossref] [PubMed]
4. Nielsen J, Sleaby R, Kumarakurusingham E, et al. Randomised controlled trials in women’s health in the last two decades: A meta-review. Eur J Obstet Gynecol Reprod Biol 2022;278:11-5. [Crossref] [PubMed]
5. Pronk AJM, Roelofs A, Flum DR, et al. Two decades of surgical randomized controlled trials: worldwide trends in volume and methodological quality. Br J Surg 2023;110:1300-8. [Crossref] [PubMed]
6. Shen G, Xu X, Li M, et al. The Clinical Trials Landscape for Alzheimer’s Disease. CNS Neurosci Ther 2025;31:e70492. [Crossref] [PubMed]
7. Vergara-Merino L, Verdejo C, Franco JVA, et al. Registered trials address questions already answered with high-certainty evidence: A sample of current redundant research. J Clin Epidemiol 2021;134:89-94. [Crossref] [PubMed]
8. Liu J, Luo X, Yao M, et al. Use of statistical methods among acupuncture randomized controlled trials was far from satisfactory. J Clin Epidemiol 2022;152:1-12. [Crossref] [PubMed]
9. Selman CJ, Lee KJ, Ferguson KN, et al. Statistical analyses of ordinal outcomes in randomised controlled trials: a scoping review. Trials 2024;25:241. [Crossref] [PubMed]
10. Lai D, Wang D, McGillivray M, et al. Assessing the quality of randomization methods in randomized control trials. Healthc (Amst) 2021;9:100570. [Crossref] [PubMed]
11. Komorowski AS, Bai AD, Cvetkovic A, et al. Methodological and reporting quality of non-inferiority randomized controlled trials comparing antifungal therapies: a systematic review. Clin Microbiol Infect 2022;28:640-8. [Crossref] [PubMed]
12. Ren Y, Jia Y, Yang M, et al. Sample size calculations for randomized controlled trials with repeatedly measured continuous variables as primary outcomes need improvements: a cross-sectional study. J Clin Epidemiol 2024;166:111235. [Crossref] [PubMed]
13. Chen H, Pang J, Zhao Y, et al. Investigating started sample size, completed sample size and drop-out rate in 10 252 phase III clinical trials: Insights from ClinicalTrials.gov. Br J Clin Pharmacol 2025;91:1957-64. [Crossref] [PubMed]
14. Lei LX, Li KC, Yu WH, et al. Analysis of Current Status of Clinical Trial Registrations in Andrological Diseases: Insights from ClinicalTrials.gov and ICTRP Databases. Am J Mens Health 2025;19:15579883251325478. [Crossref] [PubMed]
15. Mikelis F, Tzanetakis GN, Eliades T, et al. Publication bias in randomized controlled trials in dentistry. What factors affect statistical significance of outcomes? J Dent 2022;123:104183.
16. McKechnie T, Kazi T, Shi V, et al. Spin Reporting Is Common in Pilot Randomized Controlled Trials in Surgery: A Methodological Survey. J Surg Res 2025;306:16-25. [Crossref] [PubMed]
17. Demarquette A, Perrault T, Alapetite T, et al. Spin and fragility in randomised controlled trials in the anaesthesia literature: a systematic review. Br J Anaesth 2023;130:528-35. [Crossref] [PubMed]
18. Otte WM, Vinkers CH, Habets PC, et al. Analysis of 567,758 randomized controlled trials published over 30 years reveals trends in phrases used to discuss results that do not reach statistical significance. PLoS Biol 2022;20:e3001562. [Crossref] [PubMed]
19. Maurya M, Munshi R. Reasons for retraction of clinical research articles in PubMed indexed medical journals from 2012 to 2022. Indian J Med Ethics 2024;IX:296-300. [Crossref] [PubMed]
20. Xu C, Fan S, Tian Y, et al. Investigating the impact of trial retractions on the healthcare evidence ecosystem (VITALITY Study I): retrospective cohort study. BMJ 2025;389:e082068. [Crossref] [PubMed]
21. Núñez-Núñez M, Maes-Carballo M, Mignini LE, et al. Research integrity in randomized clinical trials: A scoping umbrella review. Int J Gynaecol Obstet 2023;162:860-76. [Crossref] [PubMed]
22. Yun Q, Lin M, Jia Y, et al. Prevalence of and factors associated with potentially redundant randomized controlled trials: a cross-sectional study. J Clin Epidemiol 2024;167:111265. [Crossref] [PubMed]
23. Lu J, Xu BB, Shen LL, et al. Analysis of characteristics and trends of randomized controlled trials of gastric cancer between 2000 and 2019. Zhonghua Wai Ke Za Zhi 2022;60:478-85. [Crossref] [PubMed]
24. Chen H, Huang Z, Sun B, et al. Evaluating research waste and traits among randomized controlled trials of scars over the past 20 years: a cross-sectional study. Postgrad Med J 2024;100:925-31. [Crossref] [PubMed]
25. Davidson M, Korfitsen CB, Riveros C, et al. Post-publication peer review and the identification of methodological and reporting issues in COVID-19 trials: a qualitative study. BMJ Evid Based Med 2025;30:233-40. [Crossref] [PubMed]
26. Zhang J, Wu J, Li J, et al. Trends in drug-drug interactions for new drug clinical trials in China over the past 10 years (2013-2022). BMC Pharmacol Toxicol 2025;26:66. [Crossref] [PubMed]
27. Weber F, Pscheidl T, Sydenham E, et al. Investigation of ethics approval as part of a research integrity assessment of randomised controlled trials in COVID-19 evidence syntheses: a meta-epidemiological study. BMJ Open 2025;15:e092244. [Crossref] [PubMed]
28. Bruckner T, Rodgers F, Styrmisdóttir L, et al. Adoption of World Health Organization Best Practices in Clinical Trial Transparency Among European Medical Research Funder Policies. JAMA Netw Open 2022;5:e2222378. [Crossref] [PubMed]
29. WHO. Joint statement on public disclosure of results from clinical trials. 2017. Available online: https://cdn.who.int/media/docs/default-source/clinical-trials/ictrp-jointstatement-2017.pdf
30. Mielke D, Rohde V. Randomized controlled trials-a critical re-appraisal. Neurosurg Rev 2021;44:2085-9. [Crossref] [PubMed]
31. Kaur R, Li J. How to Conduct a Randomized Controlled Trial. Respir Care 2023;69:128-38. [Crossref] [PubMed]
32. Anderson K, Romero R, Odibo AO, et al. Quality criteria for randomized controlled studies: obstetrical journal guidelines. Am J Obstet Gynecol MFM 2021;3:100334. [Crossref] [PubMed]
33. OBGYN Editors’ Integrity Group (OGEIG). Berghella V. Trustworthiness criteria for meta-analyses of randomized controlled studies: OBGYN Journal guidelines. Acta Obstet Gynecol Scand 2024;103:2118-21.
34. Gyawali B, Eisenhauer EA, van der Graaf W, et al. Common Sense Oncology principles for the design, analysis, and reporting of phase 3 randomised clinical trials. Lancet Oncol 2025;26:e80-9. [Crossref] [PubMed]
35. Prowse SR, Treweek S, Brazzelli M, et al. Improving Trial Informativeness: A Rapid Review of Global Research on How to Ensure Trials Are Useful. J Eval Clin Pract 2025;31:e70147. [Crossref] [PubMed]
36. Burford B, Norman T, Dolley S. Scientific review of protocols to enhance informativeness of global health clinical trials. Trials 2025;26:85. [Crossref] [PubMed]
37. Doi M, Yukawa K, Sato H. Characteristics of Asian 4 countries on cancer clinical trials registered in the International Clinical Trials Registry Platform between 2005 and 2018. Chin Clin Oncol 2021;10:28. [Crossref] [PubMed]
38. Hitchen N, Shahnam A, Manoharan S, et al. Trends in phase 1 oncology clinical trials across Australia; Analysis of ClinicalTrials.gov 2012-2022. Asia Pac J Clin Oncol 2024;20:731-7. [Crossref] [PubMed]
39. Mumtaz H, Haider SMA, Neha F, et al. Clinical trials landscape in a lower-middle-income country (Pakistan). J Clin Transl Sci 2024;8:e7. [Crossref] [PubMed]
40. Strüver V, Ibeneme SC. Why are emerging countries popular for clinical research? S Afr Med J 2021;111:453-9. [Crossref] [PubMed]
41. Fukuhara S, Kataoka Y, Aoki T, et al. International Collaboration and Commercial Involvement in Randomized Controlled Trials From 10 Leading Countries, 1997 Through 2019. Cureus 2024;16:e61205. [Crossref] [PubMed]
42. Huang ZN, Qiu WW, He QC, et al. Characterization changes and research waste in randomized controlled trials of global bariatric surgery over the past 20 years: cross-sectional study. Int J Surg 2024;110:1420-9. [Crossref] [PubMed]
43. Siburian MD, Muchanga SM, Villanueva A, et al. Quality of Randomized Controlled Trials in the Association of Southeast Asian Nations (ASEAN) - A Systematic Review. Ther Innov Regul Sci 2025;59:718-27. [Crossref] [PubMed]
44. Wells JC, Sharma S, Del Paggio JC, et al. An Analysis of Contemporary Oncology Randomized Clinical Trials From Low/Middle-Income vs High-Income Countries. JAMA Oncol 2021;7:379-85. [Crossref] [PubMed]
45. Arora M, Kilcoyne A, Bolodeoku J, et al. Has the UK lost its position as a destination for world-leading clinical research? A comparative analysis of haematological cancer clinical trials performance before Brexit. BMJ Open 2024;14:e086058.
46. Zhang AD, Puthumana J, Downing NS, et al. Assessment of Clinical Trials Supporting US Food and Drug Administration Approval of Novel Therapeutic Agents, 1995-2017. JAMA Netw Open 2020;3:e203284. [Crossref] [PubMed]
47. Zhang Y, Chen D, Fu M, et al. Strength of evidence supporting cancer drug approvals in China between 2017 and 2021: a retrospective analysis. Lancet Oncol 2025;26:1252-60. [Crossref] [PubMed]
48. Olivier T, Haslam A, Prasad V. Dose modification rules and availability of growth factor support: A cross-sectional study of head-to-head cancer trials used for US FDA approval from 2009 to 2021. Eur J Cancer 2022;172:349-56. [Crossref] [PubMed]
49. Naci H, Davis C, Savović J, et al. Design characteristics, risk of bias, and reporting of randomised controlled trials supporting approvals of cancer drugs by European Medicines Agency, 2014-16: cross sectional analysis. BMJ 2019;366:l5221. [Crossref] [PubMed]
50. Eldridge L, Goodman NR, Chtourou A, et al. Barriers and Opportunities for Cancer Clinical Trials in Low- and Middle-Income Countries. JAMA Netw Open 2025;8:e257733. [Crossref] [PubMed]
51. Makama M, McDougall AR, Shalit A, et al. Identifying and assessing the capacity and experience of trial sites in low- and middle-income countries for high-quality randomised drug trials in maternal and perinatal health. BMJ Glob Health 2025;10:e018063. [Crossref] [PubMed]
52. Ejigu DA, Fekadu A, Whitty J, et al. Development, implementation, and evaluation of an innovative clinical trial operations training program for Africa (ClinOps). BMC Med Educ 2025;25:119. [Crossref] [PubMed]
53. European Commission. Commission Guideline — Guidance on posting and publication of result-related information on clinical trials in relation to the implementation of Article 57(2) of Regulation (EC) No 726/2004 and Article 41(2) of Regulation (EC) No 1901/2006. Official Journal of the European Union 2012; C 302:7-10. Available online: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52012XC1006(01)
54. Ning N, Yan J, Xie XJ, et al. Impact of NCI-mandated scientific review on protocol development and content. J Natl Compr Canc Netw 2015;13:409-16. [Crossref] [PubMed]
55. Lindsley K, Fusco N, Li T, et al. Clinical trial registration was associated with lower risk of bias compared with non-registered trials among trials included in systematic reviews. J Clin Epidemiol 2022;145:164-73. [Crossref] [PubMed]
56. Tian Y, Lu H, Zhou W, et al. Prospective registration was associated with a reduced risk of bias for randomized controlled trials: a meta-research study. J Clin Epidemiol 2025;184:111813. [Crossref] [PubMed]
57. Phillips AT, Desai NR, Krumholz HM, et al. Association of the FDA Amendment Act with trial registration, publication, and outcome reporting. Trials 2017;18:333. [Crossref] [PubMed]
58. Mello AT, Kammer PV, Nascimento GM, et al. Credibility at stake: only two-thirds of randomized trials of nutrition interventions are registered and lack transparency in outcome and treatment effect definitions. J Clin Epidemiol 2023;161:74-83. [Crossref] [PubMed]
59. Riddle DL, Goldman AH, Tarver T, et al. Examination of Randomized Trials and Corresponding Trial Registry Entries: Registration Timing and Primary Outcome Analysis in the Journal of Arthroplasty. J Arthroplasty 2022;37:1645-1649.e7. [Crossref] [PubMed]
60. Vitali FC, Santos PS, Rocha AO, et al. Adherence to Registration and Selective Outcome Reporting in Randomized Clinical Trials Published in Endodontic Journals Over the Past 5 Years: A Meta-Research Study. J Endod 2025;51:258-267.e7. [Crossref] [PubMed]
61. Rahal DA, Khodeiry MM, Dihan Q, et al. Characteristics and publication rate of registered clinical trials in amblyopia. J AAPOS 2025;29:104182. [Crossref] [PubMed]
62. Meller L, Sambo AB, Nguyen N, et al. Publication rates of registered strabismus trials from ClinicalTrials.gov. J AAPOS 2024;28:103936. [Crossref] [PubMed]
63. Lin L, Tang Y, Yang L, et al. Research waste among randomized controlled trials in ovarian cancer: A cross-sectional study. Eur J Surg Oncol 2024;50:108437. [Crossref] [PubMed]
64. Lu D, Chen X, Mu Y, et al. Discontinuation and non-publication of randomized controlled trials on cervical cancer or precancer. Jpn J Clin Oncol 2024;54:1141-9. [Crossref] [PubMed]
65. Oca M, Rajan SM, Qian ZJ. Otitis Media Trial Publication After ClinicalTrials.gov Registration. Otolaryngol Head Neck Surg 2025;173:882-7. [Crossref] [PubMed]
66. Nebeska K, Souckova L, Stepanova R, et al. Exploring the publication gap in pediatric randomized clinical trials: completed vs. uncompleted pediatric clinical trials. Front Med (Lausanne) 2025;12:1590125.
67. Pei L, Su M, Yakeya R, et al. Current landscape of dyslipidemia-related randomized clinical trials registered on the International Clinical Trials Registry Platform. Front Public Health 2025;13:1554858. [Crossref] [PubMed]
68. Lippert T, Schmucker E, Shukla A, et al. Publication Bias in Upper Gastrointestinal Oncology Clinical Trials. J Gastrointest Cancer 2024;55:950-5. [Crossref] [PubMed]
69. Wan M, Alessandrini E, Brogan P, et al. Risk-proportionate approach to paediatric clinical trials: The legal requirements, challenges, and the way forward under the European Union Clinical Trials Regulation. Clin Trials 2022;19:573-8. [Crossref] [PubMed]
70. Schönenberger CM, Chiaborelli M, Taji Heravi A, et al. Public availability of randomized clinical trial protocols: a repeated meta-research study. J Clin Epidemiol 2025;185:111865. [Crossref] [PubMed]
71. Syed M, Martin H, Sena ES, et al. Selective outcome reporting in randomised controlled trials including participants with stroke or transient ischaemic attack: A systematic review. Eur Stroke J 2023;8:923-31. [Crossref] [PubMed]
72. Campbell D, McDonald C, Cro S, et al. Access to unpublished protocols and statistical analysis plans of randomised trials. Trials 2022;23:674. [Crossref] [PubMed]
73. Mathieu S, Bouillon-Minois JB, Renard Triché L, et al. Protocol publication rate and comparison between article, registry and protocol in RCTs. BMC Med Res Methodol 2025;25:31. [Crossref] [PubMed]
74. Vorland CJ, Brown AW, Kilicoglu H, et al. Publication of Results of Registered Trials With Published Study Protocols, 2011-2022. JAMA Netw Open 2024;7:e2350688. [Crossref] [PubMed]
75. Li Y, Liu YN, Guo Z, et al. Current situation of clinical trial registration in acupuncture anesthesia: A scoping review. J Integr Med 2025;23:256-63. [Crossref] [PubMed]
76. Li S, Li J, Zeng W, et al. Re-analysis of the current status of clinical trial registration in China. Front Med (Lausanne) 2024;11:1394803. [Crossref] [PubMed]
77. Buruk B, Guner MD, Ekmekci PE, et al. Comparison of COVID-19 studies registered in the clinical trial platforms: A research ethics analysis perspective. Dev World Bioeth 2022;22:217-30. [Crossref] [PubMed]
78. Samanta AD, Mendiratta J, Saberwal G. Identifying trials run in India, registered with ClinicalTrials.gov, is not straightforward. J Biosci 2025;50:27.
79. O’Riordan M, Haslberger M, Cruz C, et al. Are European clinical trial funders policies on clinical trial registration and reporting improving? A cross-sectional study. J Clin Transl Sci 2023;7:e166.
80. AllTrials. All trials registered. All results reported. 2013. Available online: https://www.alltrials.net/wp-content/uploads/2013/09/What-does-all-trials-registered-and-reported-mean.pdf
81. Mongin D, Buitrago-Garcia D, Capderou S, et al. Prospective registration of trials: where we are, why, and how we could get better. J Clin Epidemiol 2024;176:111586. [Crossref] [PubMed]
82. Dal-Ré R. Is full transparency in clinical trials an achievable goal? Eur J Intern Med 2024;123:1-3. [Crossref] [PubMed]
83. Engel A, Cohen O. Trends in Israeli clinical trials registration “MyTrial”. Isr J Health Policy Res 2024;13:62. [Crossref] [PubMed]
84. Ravindranath AG, Benny A, Bhalla IS, et al. Clinical trial registration in India: 12% of drug regulatory trials are not registered, as required by law. Health Res Policy Syst 2025;23:19. [Crossref] [PubMed]
85. Mirjam Keestra S, Rodgers F, Gepp S, et al. Improving clinical trial transparency at UK universities: Evaluating 3 years of policies and reporting performance on the European Clinical Trial Register. Clin Trials 2022;19:217-23. [Crossref] [PubMed]
86. Balshem H, Helfand M, Schünemann HJ, et al. GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol 2011;64:401-6.
87. Kaur A, Bali RK, Patnana AK, et al. Quality Assessment of Randomized Controlled Trials Published In Journal of Maxillofacial and Oral Surgery (MAOS) From 2009-2021 Using RoB-2.0 Tool. J Maxillofac Oral Surg 2025;24:1-7. [Crossref] [PubMed]
88. Zhang Q, Wu D, Wang Y, et al. Quality evaluation of the literature on clinical randomized controlled trials of traditional Chinese medicine for treatment of gastric precancerous lesions in the past 20 years. Zhejiang Da Xue Xue Bao Yi Xue Ban 2023;52:636-45. [Crossref] [PubMed]
89. Masuyama S, Yamashita H. Trends and quality of randomized controlled trials on acupuncture conducted in Japan by decade from the 1960s to the 2010s: a systematic review. BMC Complement Med Ther 2023;23:91. [Crossref] [PubMed]
90. Chen H, Chang R, Chen Y, et al. Advancing randomized controlled trials of vascular anomalies: an analysis of trial waste. Eur J Pediatr 2024;183:5001-11. [Crossref] [PubMed]
91. Lin TA, McCaw ZR, Koong A, et al. Proportional Hazards Violations in Phase III Cancer Clinical Trials: A Potential Source of Trial Misinterpretation. Clin Cancer Res 2024;30:4791-9. [Crossref] [PubMed]
92. Rajendrabose D, Collet L, Reinaud C, et al. Some superiority trials with nonsignificant results published in high impact factor journals correspond to noninferiority situations: a research-on-research study. J Clin Epidemiol 2025;177:111613. [Crossref] [PubMed]
93. Wang Y, Yao M, Liu J, et al. A systematic survey of adaptive trials shows substantial improvement in methods is needed. J Clin Epidemiol 2024;167:111257. [Crossref] [PubMed]
94. Alvarez G, Núñez-Cortés R, Solà I, et al. Sample size, study length, and inadequate controls were the most common self-acknowledged limitations in manual therapy trials: A methodological review. J Clin Epidemiol 2021;130:96-106. [Crossref] [PubMed]
95. Iheozor-Ejiofor Z, Lakunina S, Gordon M, et al. Sample-size estimation is not reported in 24% of randomised controlled trials of inflammatory bowel disease: A systematic review. United European Gastroenterol J 2021;9:47-53. [Crossref] [PubMed]
96. Nelms DW, Vargas HD, Bedi RS, et al. When the p Value Doesn’t Cut It: The Fragility Index Applied to Randomized Controlled Trials in Colorectal Surgery. Dis Colon Rectum 2022;65:276-83. [Crossref] [PubMed]
97. Muthu S, Ramakrishnan E. Fragility Analysis of Statistically Significant Outcomes of Randomized Control Trials in Spine Surgery: A Systematic Review. Spine (Phila Pa 1976) 2021;46:198-208. [Crossref] [PubMed]
98. Maldonado DR, Go CC, Huang BH, et al. The Fragility Index of Hip Arthroscopy Randomized Controlled Trials: A Systematic Survey. Arthroscopy 2021;37:1983-9. [Crossref] [PubMed]
99. Itaya T, Isobe Y, Suzuki S, et al. The Fragility of Statistically Significant Results in Randomized Clinical Trials for COVID-19. JAMA Netw Open 2022;5:e222973. [Crossref] [PubMed]
100. Nguyen TL, Xie L. Incomparability of treatment groups is often blindly ignored in randomised controlled trials - a post hoc analysis of baseline characteristic tables. J Clin Epidemiol 2021;130:161-8. [Crossref] [PubMed]
101. Bramley P, Hulman J, Wanstall H. Risk of bias and problematic trials: characterising the research integrity of trials submitted to Anaesthesia. Anaesthesia 2024;79:1309-16. [Crossref] [PubMed]
102. Fukuma S, Ikenoue T, Yamada Y, et al. Changes in Drug Utilization After Publication of Clinical Trials and Drug-Related Scandals in Japan: An Interrupted Time Series Analysis, 2005-2017. J Epidemiol 2021;31:410-6. [Crossref] [PubMed]
103. Mathieu S, Chan AW, Ravaud P. Use of trial register information during the peer review process. PLoS One 2013;8:e59910. [Crossref] [PubMed]
104. Logullo P, MacCarthy A, Kirtley S, et al. Peer review reports of randomized controlled trials in oncology can be short and superficial. J Clin Epidemiol 2025;185:111893. [Crossref] [PubMed]
105. Machado T, Mainoli B, Caldeira D, et al. Data monitoring committees in pediatric randomized controlled trials registered in ClinicalTrials.gov. Clin Trials 2023;20:624-31. [Crossref] [PubMed]
106. Dello Russo C, Cappoli N, Pilunni D, et al. Local Investigators Significantly Overestimate Overall Response Rates Compared to Blinded Independent Central Reviews in Phase 2 Oncology Trials. J Clin Pharmacol 2021;61:810-9. [Crossref] [PubMed]
107. Carneiro CFD, da Costa GG, Neves K, et al. Characterization of Comments About bioRxiv and medRxiv Preprints. JAMA Netw Open 2023;6:e2331410. [Crossref] [PubMed]
108. Barker TH, Stone JC, Sears K, et al. The revised JBI critical appraisal tool for the assessment of risk of bias for randomized controlled trials. JBI Evid Synth 2023;21:494-506. [Crossref] [PubMed]
109. Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 2019;366:l4898. [Crossref] [PubMed]
110. Critical Appraisal Skills Programme (CASP). Randomised controlled trials (RCTs) Checklist. 2024. Available online: https://casp-uk.net/casp-tools-checklists/randomised-controlled-trial-rct-checklist/
111. Mol BW, Lai S, Rahim A, et al. Checklist to assess Trustworthiness in RAndomised Controlled Trials (TRACT checklist): concept proposal and pilot. Res Integr Peer Rev 2023;8:6. [Crossref] [PubMed]
112. WilkinsonJHealCFlemyngEINSPECT-SR: a tool for assessing trustworthiness of randomised controlled trials.medRxiv 2025;2025.09.03.25334905.
113. Wilkinson J, Heal C, Antoniou GA, et al. Protocol for the development of a tool (INSPECT-SR) to identify problematic randomised controlled trials in systematic reviews of health interventions. BMJ Open 2024;14:e084164. [Crossref] [PubMed]
114. Alfirevic Z, Weeks J. Trustworthiness assessment of published clinical trials: Literature review of domains and questions. Cochrane Evid Synth Methods 2024;2:e12099. [Crossref] [PubMed]
115. Lawson S, Doulaveris G, Anderson K, et al. Retracted vs non-retracted obstetrical randomized trials: Which quality criteria are most associated with retraction for untrustworthiness? Eur J Obstet Gynecol Reprod Biol 2025;309:137-42. [Crossref] [PubMed]
116. Khan KS, Fawzy M, Chien P, et al. International multistakeholder consensus statement on post-publication integrity issues in randomized clinical trials by Cairo Consensus Group. Int J Gynaecol Obstet 2025;169:1093-115. [Crossref] [PubMed]
117. Hayden JA, Ogilvie R, Kashif S, et al. Identification of latent classes of randomized controlled trials based on integrity and reporting characteristics. J Clin Epidemiol 2025;179:111644. [Crossref] [PubMed]
118. Zarin DA, Goodman SN, Kimmelman J. Harms From Uninformative Clinical Trials. JAMA 2019;322:813-4. [Crossref] [PubMed]
119. Van’t HooftJvan DijkCAxforsCValue of clinical research: Usefulness tool development and systematic review of 350 randomised controlled trials in preterm birth.Authorea 2021. doi: .
120. Smith EG, Grigorian HL. A System for Rapidly Yet Rigorously Evaluating the Quality of Randomized Controlled Trials. J Clin Psychopharmacol 2023;43:306-12. [Crossref] [PubMed]
121. Bordewijk EM, Li W, van Eekelen R, et al. Methods to assess research misconduct in health-related research: A scoping review. J Clin Epidemiol 2021;136:189-202. [Crossref] [PubMed]
122. World Health Organization. Extending the Thirteenth General Programme of Work, 2019–2023 to 2025. 2022. Available online: https://apps.who.int/gb/ebwha/pdf_files/WHA75/A75_8-en.pdf
123. Mol BW, Bordewijk EM, Rogozińska E, et al. Data integrity of 14 randomised controlled trials. Eur J Obstet Gynecol Reprod Biol 2024;298:98-103. [Crossref] [PubMed]
124. Katsanos AH, Lioutas VA, Yperzeele L, et al. Perception and acquaintance of stroke specialists on non-inferiority trials: An international survey. J Stroke Cerebrovasc Dis 2025;34:108132. [Crossref] [PubMed]
125. Cole GD, Shun-Shin MJ, Nowbar AN, et al. Difficulty in detecting discrepancies in a clinical trial report: 260-reader evaluation. Int J Epidemiol 2015;44:862-9. [Crossref] [PubMed]
126. Willis JV, Cobey KD, Ramos J, et al. Limited online training opportunities exist for scholarly peer reviewers. J Clin Epidemiol 2023;161:65-73. [Crossref] [PubMed]
127. Butterworth JW, Boshier PR, Mavroveli S, et al. Challenges to quality assurance of surgical interventions in clinical oncology trials: A systematic review. Eur J Surg Oncol 2021;47:748-56. [Crossref] [PubMed]
128. Zhang K, Ma Y, Wu J, et al. The SUPER reporting guideline suggested for reporting of surgical technique. Hepatobiliary Surg Nutr 2023;12:534-44. [Crossref] [PubMed]
129. Gaudino M, Chikwe J, Bagiella E, et al. Methodological Standards for the Design, Implementation, and Analysis of Randomized Trials in Cardiac Surgery: A Scientific Statement From the American Heart Association. Circulation 2022;145:e129-42. [Crossref] [PubMed]
130. Prado MC, Dotto L, Agostini B, et al. Global mapping of randomized controlled trials in dentistry. Braz Dent J 2024;35:e246233. [Crossref] [PubMed]
131. Grüßer L, Eißing C, Kowark A, et al. Poor reporting quality of randomized controlled trials comparing treatments of COVID-19-A retrospective cross-sectional study on the first year of publications. PLoS One 2023;18:e0292860. [Crossref] [PubMed]
132. Parums DV. Editorial: Reporting Clinical Trials with Important Modifications Due to Extenuating Circumstances, Including the COVID-19 Pandemic: CONSERVE 2021. Med Sci Monit 2021;27:e934514. [Crossref] [PubMed]
133. Wang X, Long Y, Zhang N, et al. Impact of selective reporting bias on stroke trials: potential compromise in evidence synthesis - A cross-sectional study. BMC Med Res Methodol 2024;24:255. [Crossref] [PubMed]
134. Hu H, Wang X, Huang P. Result reporting and early discontinuation of sepsis trials registered on ClinicalTrials.gov. J Evid Based Med 2023;16:186-93. [Crossref] [PubMed]
135. Kim KS, Tao B, Kaur H, et al. Reporting Quality of Noninferiority Randomized Controlled Trials for Skin Disease: A Meta-Epidemiological Study. J Invest Dermatol 2026;146:418-24. [Crossref] [PubMed]
136. Huang Y, Zhong J, Zhu T. Reporting quality of dose-expansion cohorts in trials of anticancer drugs could be improved: a cross-sectional study. Trials 2025;26:198. [Crossref] [PubMed]
137. Varghese E, Briola A, Kennel T, et al. A systematic review of stepped wedge cluster randomized trials in high impact journals: assessing the design, rationale, and analysis. J Clin Epidemiol 2025;178:111622. [Crossref] [PubMed]
138. Nascimento DP, Ostelo R, van Tulder MW, et al. Do not make clinical decisions based on abstracts of healthcare research: A systematic review. J Clin Epidemiol 2021;136-57.
139. Paheerathan S, Flitti D, Cobourne MT, et al. Reporting quality of randomized controlled trial abstracts presented at the European Orthodontic Society Congress between 2015-2024: has there been an improvement over time? Eur J Orthod 2025;47:cjaf039.
140. Fernández-Pires P, Prieto-Botella D, Valera-Gran D, et al. Reporting Quality in Abstracts of Randomized Controlled Trials Published in High-Impact Occupational Therapy Journals. Am J Occup Ther 2022;76:7604205120. [Crossref] [PubMed]
141. Zheng R, Tao L, Sun Y, et al. Inadequate Reporting of Harm From Randomized Clinical Trials in Top Medical Publications. J Evid Based Med 2025;18:e70006. [Crossref] [PubMed]
142. Broager MI, Faltinsen E, Hróbjartsson A, et al. Funder and authors’ financial conflicts of interest information in trial publications: adequacy of reporting and obtaining missing information. J Clin Epidemiol 2025;182:111786. [Crossref] [PubMed]
143. McErlean M, Samways J, Godolphin PJ, et al. The reporting standards of randomised controlled trials in leading medical journals between 2019 and 2020: a systematic review. Ir J Med Sci 2023;192:73-80. [Crossref] [PubMed]
144. Gohari F, Baradaran HR, Tabatabaee M, et al. Quality of reporting randomized controlled trials (RCTs) in diabetes in Iran; a systematic review. J Diabetes Metab Disord 2015;15:36. [Crossref] [PubMed]
145. Eleftheriadi I, Ioannou T, Katechi V, et al. Not enough SPIRIT shown in the registration and reporting of orthodontic trial protocols. Eur J Orthod 2023;45:29-37. [Crossref] [PubMed]
146. Shah D, Sparano F, Shah V, et al. Quality of Patient-Reported Outcomes in Multiple Myeloma Randomized Controlled Trial Protocols and Publications According to the SPIRIT-PRO and CONSORT-PRO guidelines: A Systematic Review. Eur J Haematol 2025;115:454-64. [Crossref] [PubMed]
147. Shang BH, Yang FH, Lin Y, et al. Discrepancies between pre-specified and reported primary outcomes: A cross-sectional analysis of randomized controlled trials in gastroenterology and hepatology journals. PLoS One 2024;19:e0305027. [Crossref] [PubMed]
148. Yuan M, Wu J, Li A, et al. “Spin” in Plastic Surgery Randomized Controlled Trials with Statistically Nonsignificant Primary Outcomes: A Systematic Review. Plast Reconstr Surg 2023;151:506e-19e.
149. Duran M, Boutron I, Hopewell S, et al. A cross-sectional study assessing visual abstracts of randomized trials revealed inadequate reporting and high prevalence of spin. J Clin Epidemiol 2024;176:111544. [Crossref] [PubMed]
150. Showell MG, Cole S, Clarke MJ, et al. Time to publication for results of clinical trials. Cochrane Database Syst Rev 2024;11:MR000011. [Crossref] [PubMed]
151. Chavalarias D, Wallach JD, Li AH, et al. Evolution of Reporting P Values in the Biomedical Literature, 1990-2015. JAMA 2016;315:1141-8. [Crossref] [PubMed]
152. Adda J, Decker C, Ottaviani M. P-hacking in clinical trials and how incentives shape the distribution of results across phases. Proc Natl Acad Sci U S A 2020;117:13386-92. [Crossref] [PubMed]
153. Hirst A, Altman DG. Are peer reviewers encouraged to use reporting guidelines? A survey of 116 health research journals. PLoS One 2012;7:e35621.
154. Hopewell S, Altman DG, Moher D, et al. Endorsement of the CONSORT Statement by high impact factor medical journals: a survey of journal editors and journal ‘Instructions to Authors’. Trials 2008;9:20. [Crossref] [PubMed]
155. Sharp MK, Glonti K, Hren D. Online survey about the STROBE statement highlighted diverging views about its content, purpose, and value. J Clin Epidemiol 2020;123:100-6. [Crossref] [PubMed]
156. Fuller T, Pearson M, Peters J, et al. What affects authors’ and editors’ use of reporting guidelines? Findings from an online survey and qualitative interviews. PLoS One 2015;10:e0121585.
157. Ventura JVL, da Silva CF, Lopes TDS, et al. Ameloblastoma on YouTube: content analysis and strategies to address misinformation. Oral Maxillofac Surg 2025;29:91. [Crossref] [PubMed]
158. Lorenzon L, Grossman RC, Soreide K. Impact of Randomized Controlled Trials in the Social Media: Does Science Trend As Much As Everyday Events? World J Surg 2021;45:88-96. [Crossref] [PubMed]
159. Skelin M, Katić J, Šarčević D, et al. Comparison of media and academic attention of recently published positive and neutral or negative randomized cardiovascular clinical trials. Rev Cardiovasc Med 2022;23:31. [Crossref] [PubMed]
160. Pal A, Klingmann I, Wangmo T, et al. Publishing clinical trial results in plain language: a clash of ethical principles? Curr Med Res Opin 2024;40:493-503. [Crossref] [PubMed]
161. Dal-Ré R, Caplan AL, Holm S, et al. What participants are told about receiving trial results when they consent to participate in a trial. Basic Clin Pharmacol Toxicol 2025;136:e14115. [Crossref] [PubMed]
162. Moher D, Jones A, Lepage L, et al. Use of the CONSORT statement and quality of reports of randomized trials: a comparative before-and-after evaluation. JAMA 2001;285:1992-5. [Crossref] [PubMed]
163. Did a change in Nature journals’ editorial policy for life sciences research improve reporting? BMJ Open Sci 2019;3:e000035.
164. Lin Y, Yang F, Shang B, et al. Reporting quality of animal research in journals that published the ARRIVE 1.0 or ARRIVE 2.0 guidelines: a cross-sectional analysis of 943 studies. Cardiovasc Diagn Ther 2024;14:1070-82. [Crossref] [PubMed]
165. Ogino D, Takahashi K, Sato H. Characteristics of clinical trial websites: information distribution between ClinicalTrials.gov and 13 primary registries in the WHO registry network. Trials 2014;15:428. [Crossref] [PubMed]
166. Reuter K, Ukpolo F, Ward E, et al. Trial Promoter: A Web-Based Tool for Boosting the Promotion of Clinical Research Through Social Media. J Med Internet Res 2016;18:e144. [Crossref] [PubMed]
167. Zannad F, Crea F, Keaney J, et al. Rapid, accurate publication and dissemination of clinical trial results: benefits and challenges. Eur Heart J 2023;44:4220-9. [Crossref] [PubMed]
168. Yan MK, Adler NR, Heriot N, et al. Opportunities and barriers for the use of Australian cancer registries as platforms for randomized clinical trials. Asia Pac J Clin Oncol 2022;18:344-52. [Crossref] [PubMed]
169. Prang KH, Karanatsios B, Zhang A, et al. “Nothing to lose and the possibility of gaining”: a qualitative study on the feasibility and acceptability of registry-based randomised controlled trials among cancer patients and clinicians. Trials 2023;24:92. [Crossref] [PubMed]
170. Urban L, Haller N, Pieper D, et al. A methodological review identified several options for utilizing registries for randomized controlled trials. J Clin Epidemiol 2025;178:111614. [Crossref] [PubMed]
171. Karanatsios B, Prang KH, Yeung JM, et al. A qualitative study exploring stakeholders’ perceptions of registry-based randomised controlled trials capacity and capability in Australia. Trials 2024;25:834. [Crossref] [PubMed]
172. Pinheiro Andrade M, Matias D, Batuca J, et al. Portuguese Authorship in Published Clinical Trials: Differences in Industry and Investigator Initiated Trials. Acta Med Port 2021;34:733-40. [Crossref] [PubMed]
173. Di Tonno D, Perlin C, Loiacono AC, et al. Trends of Phase I Clinical Trials in the Latest Ten Years across Five European Countries. Int J Environ Res Public Health 2022;19:14023. [Crossref] [PubMed]
174. Sándor-Bajusz KA, Kraut A, Baasan O, et al. Publication of clinical trials on medicinal products: follow-up on trials authorized in Hungary. Trials 2022;23:330. [Crossref] [PubMed]
175. Morrow RL, Mintzes B, Gray G, et al. Industry Sponsor Influence in Clinical Trial Reporting in Canada: A Qualitative Interview Study. Clin Ther 2022;44:374-88. [Crossref] [PubMed]
176. Gattrell WT, Hopewell S, Young K, et al. Professional medical writing support and the quality of randomised controlled trial reporting: a cross-sectional study. BMJ Open 2016;6:e010329. [Crossref] [PubMed]
177. Amsterdam JD, McHenry LB, Jureidini JN. Industry-corrupted psychiatric trials. Psychiatr Pol 2017;51:993-1008. [Crossref] [PubMed]
178. Kumar G, Cherian JJ, Mathur R, et al. Navigating investigator-initiated clinical trials: A call for guidelines & monitoring frameworks from an Indian context. Indian J Med Res 2024;160:141-9. [Crossref] [PubMed]
179. Maru S, Matthias MD, Kuwatsuru R, et al. Studies of Artificial Intelligence/Machine Learning Registered on ClinicalTrials.gov: Cross-Sectional Study With Temporal Trends, 2010-2023. J Med Internet Res 2024;26:e57750. [Crossref] [PubMed]
180. Lu Y, Chen J, Zhang T, et al. Analysis of the global registration status of clinical trials for artificial intelligence medical device. Sheng Wu Yi Xue Gong Cheng Xue Za Zhi 2025;42:512-9. [Crossref] [PubMed]
181. Pattathil N, Zhao JZL, Sam-Oyerinde O, et al. Adherence of randomised controlled trials using artificial intelligence in ophthalmology to CONSORT-AI guidelines: a systematic review and critical appraisal. BMJ Health Care Inform 2023;30:e100757. [Crossref] [PubMed]
182. Martindale APL, Llewellyn CD, de Visser RO, et al. Concordance of randomised controlled trials for artificial intelligence interventions with the CONSORT-AI reporting guidelines. Nat Commun 2024;15:1619. [Crossref] [PubMed]
183. He Z, Yang L, Li X, et al. Discrepancies in reported results between trial registries and journal articles for AI clinical research. EClinicalMedicine 2025;80:103066. [Crossref] [PubMed]
184. Ghosh S, Abushukair HM, Ganesan A, et al. Harnessing explainable artificial intelligence for patient-to-clinical-trial matching: A proof-of-concept pilot study using phase I oncology trials. PLoS One 2024;19:e0311510. [Crossref] [PubMed]
185. Jin Q, Wang Z, Floudas CS, et al. Matching patients to clinical trials with large language models. Nat Commun 2024;15:9074. [Crossref] [PubMed]
186. Marshall IJ, Nye B, Kuiper J, et al. Trialstreamer: A living, automatically updated database of clinical trial reports. J Am Med Inform Assoc 2020;27:1903-12. [Crossref] [PubMed]
187. Tian Y, Yang X, Doi SA, et al. Towards the automatic risk of bias assessment on randomized controlled trials: A comparison of RobotReviewer and humans. Res Synth Methods 2024;15:1111-9. [Crossref] [PubMed]
188. Au LS, Qu L, Nielsen J, et al. Using artificial intelligence to semi-automate trustworthiness assessment of randomized controlled trials: a case study. J Clin Epidemiol 2025;180:111672. [Crossref] [PubMed]
189. Naddaf M. More than half of researchers now use AI for peer review - often against guidance. Nature 2026;649:273-4. [Crossref] [PubMed]