The implications of the CONsolidated Standards of Reporting Trials 2025 revised guidelines for data sharing on clinical trial reporting

Introduction The CONsolidated Standards Of Reporting Trials (CONSORT) Statement was published in 1996 in an effort to improve the quality of reporting of randomized trials. It was subsequently updated in 2001 and 2010.[1] A number of substantial changes were made in the CONSORT 2025 checklist. Additionally, the checklist was restructured to include a section on open science comprising trial registration (item 2), trial protocol and statistical analysis plan (SAP) access (item 3), sharing of de-identified participant-level data (item 4), and funding and conflicts of interest (item 5). This article reviews item 4 of the checklist in the context of the benefits of data transparency and potential risks of data sharing.[1] Data Sharing: Item 4 of the CONsolidated Standards Of Reporting Trials 2025 Checklist Data to be shared include de-identified participant data, data dictionaries, analysis codes, and study materials. The guideline states that the reporting should specify where the data will be accessible and the application process. If there is no plan for sharing data, that must be stated with a clear justification.[1] The Benefits and Challenges of Data Sharing and the Way Forward Clinical trial data sharing could have a remarkable impact on data transparency and could potentially improve the quality of research.[2] Sharing of participant-level data, protocols, and SAPs along with the published articles can pave the way for novel secondary analyses by independent researchers and decision-making on future clinical trial designs. The FAIR (Findable, Accessible, Interoperable, and Reusable) Guiding Principles for scientific data management and stewardship emphasise the importance of data transparency for this purpose.[3] Data from clinical study reports (CSRs) support the conduct of meta-analyses, which is crucial for evidence-based medicine and decision-making in the formation of treatment guidelines and health policies.[4] With the undeniable benefits of data sharing, it was inevitable that CONSORT 2025 Statement: updated guideline for reporting randomised trials would make it an essential item on their checklist. Data sharing, however, has its fair share of considerations. Although there is no argument to be made against its necessity, there is valid concern about the ethics of sharing sensitive information, data security, due credit given to the original researcher, and the increasing use of artificial intelligence (AI) in data management.[4] While a participant consents to the primary research, the consent cannot be extended to any subsequent research done, and, for an independent researcher, it is impractical to obtain consent. A solution to this may be to obtain explicit consent for data sharing from all participants, mentioning the general benefits of data sharing in research, how their data will be collected, stored, de-identified, and shared and who will be able to access the data. This should also be reflected in the consent forms. Additionally, a participant’s right to withdraw consent from any secondary research should also be respected. This would mean that the participants’ data will be removed from the dataset before sharing – a tedious but necessary task.[4,5] The lack of credit and fear of their work being stolen may discourage many researchers from sharing the data they have painstakingly collected, analyzed, and interpreted. There is currently not much known about whether primary researchers are required to share electronic Case Record Form (eCRF) data, but with the growing use of electronic data capture systems, the guidelines need to be updated as well.[6] Needless to say, the original researchers must be given all due credit and any compensation owed to them, which includes mandatory and explicit acknowledgment and could include authorship in the new work, based on their contribution. Data security is paramount while sharing participant information. Robust systems must be set in place to prevent security breaches.[7] Misuse of personal and sensitive data, especially genetic data, by insurance companies to deny claims or for forensic analysis, must be prevented at all costs. The risk of re-identification by adversaries also needs to be determined, especially for audio-visual recordings[8] and genomic data. There must be a clear and written commitment by the independent researchers to be transparent in their intentions of publishing their secondary analyses and sharing the analysis code. They must agree to maintain privacy, not disclose any data to third parties, and not attempt to re-identify study participants. Data should be shared only after a thorough and rigorous review of the data request has been made by an independent committee. The request must state that the data sharing would be only for scientific purposes and exclude forensic or insurance purposes. The terms and conditions should specify that commercial use of data is strictly forbidden. Once the independent committee has approved the request, a robust online sharing platform must be employed for sharing purposes with restricted access.[4,5] As per the 21st Century Cures Act, Certificates of Confidentiality are to be issued for all research funded by the National Institute of Health (NIH) that collects and uses “identifiable, sensitive information.” This was an update from the previous policy, in which only sensitive and stigmatizing data was protected.[6,9] In 2016, the International Committee of Medical Journal Editors proposed that participant-level data be shared within 6 months of publication of the trial results.[6] Clinical trials may be “ineligible” for data sharing due to many reasons, such as ongoing and long-term follow-up, extended embargoes or other pending regulatory requirements, and the absence of explicit consent by participants for secondary analysis. An embargo is a hold on the dissemination of study data and results to ensure that the findings are presented responsibly and ethically. This prevents premature sensationalization of the results and allows for a more effective scientific communication.[4,10] While acknowledging that long-term follow-up data is vital for long-term safety, the Pharmaceutical Research and Manufacturers of America (PhRMA) and European Federation of Pharmaceutical Industries and Associations (EFPIA) still recommend that participant-level data from any clinical trial submitted for regulatory approval should be shared immediately, and data from the follow-up study can be updated. The PhRMA/EFPIA recommends that sponsors keep a public registry of trials that are eligible or ineligible for sharing, with reasons for ineligibility.[4] A recent study by Siebert and Ioannidis found that two-thirds of clinical trials published in the top medical journals lifted the embargoes for data sharing that they had stated in their original data sharing statements.[10] Data from CSRs can facilitate meta-analyses and help healthcare professionals make more informed judgments about the treatment by providing more context and safety data. Regulatory authorities encourage the sponsors to share CSRs on their online platforms. The de-identification techniques employed could vary drastically from company to company, depending on their data management policies. This would make the meta-analysis process more difficult. A solution to this could be a standardized de-identification method, which could be decided upon by a consensus of experts in the field. Dissemination of lay summaries of study results to trial participants, and the general public fosters trust in the clinical research process, regulatory bodies, and the approved medicines. It empowers patients to make informed decisions about their treatment options and gives them more autonomy. In rare cases, where clinical trials have been terminated prematurely due to safety reasons, public communication has been necessary even before the embargo has been lifted. In such cases, the communication must be restricted to the reasons for which the study ended.[4] A recent conundrum that researchers now face is the use of AI in data management. Specifically, large language models are plagued with the “black-box problem,” which refers to the inability to understand how AI makes critical decisions, as well as how it analyzes and interprets data. This not only makes it difficult to identify biases in research but also makes reproducibility impossible. Furthermore, it is difficult to identify the root cause of any erroneous output. There has been a recent call for explicability to be added to the principles of bioethics in health care. The levels of explicability are transparency (to make the input and output data accessible), explainability (ability of the model to explain its decision-making process), and interpretability (ability to understand the logic and reasoning of the AI system).[11,12] Worldwide, agencies such as the Food and Drug Administration and the NIH, the European Medicines Agency, the Cancer Research UK, the Organization for Economic Co-operation and Development, the Wellcome Trust, and the Japan Science and Technology Agency’s National Bioscience Database Center have recognized the importance of data sharing and have given their guidelines and recommendations. In India, the National AIDS Control Organisation has protocols for data sharing to protect patient information. The Indian Ministry of Science and Technology released the National Data Sharing and Accessibility Policy to promote the sharing of data owned by the government. The Indian Council of Medical Research, in its ethical guidelines, stresses the need to de-identify patient information and give credit to the primary researchers.[6] The CONSORT 2025 update has cast yet another spotlight on data sharing and will surely drive regulatory agencies, associations like the PhRMA/EFPIA, and researchers to rethink the ethical, legal, and practical aspects of data sharing. Conclusion The CONSORT 2025 update in the guidelines has some notable additions, specifically regarding data sharing. Data sharing encourages transparency, facilitates secondary research, and inculcates public confidence in the clinical trial process. However, participant consent, patient confidentiality, data ownership, and data security are of top priority. The growing popularity and use of AI is a boon; however, the lack of transparency due to the black-box problem is a formidable challenge and requires efforts to make AI better at explicability. While the researchers may face many obstacles while navigating the updated guidelines, the prioritization of data sharing is a step in the right direction and will not only advance clinical research but also improve the quality and integrity of clinical trials.