This policy describe the principles and procedures for sharing of data from trials,
observational studies and other research projects. This documents aims to:
There has been a move towards greater transparency in design, conduct and reporting in both the public and private sectors over recent years. Better registration of trials has led to a clearer understanding of ongoing research. There is more reporting of results and clearer presentation of data within those reports, e.g. CONSORT. Many funders and sponsors now insist on open access publication whereby results are free to the reader, either immediately or within a set time.
In parallel, a number of funders and sponsors have moved towards greater access to data from trials and other projects they have supported. They have done this by implementing policies that encourage or mandate access to results or data, either via the host institution or through appropriate data repositories. There are a number of advantages for research transparency including the ability to:
(i) Place the results of the study in a larger context, e.g. as part of an individual patient data meta-analysis (IPD MA).
(ii) Make secondary use of the dataset, e.g. using the trial or cohort as a convenience sample of high quality, prospectively collected data to address a different question.
(iii) Collaborate directly with other researchers where data need to be transferred to an alternative location for planned analyses, e.g. as part of biological sub-studies.
(iv) Provide supporting evidence to plan a new trial, e.g. estimating the expected event on their control arm, or estimating rates of recruitment
(v) Develop and validate new methodologies, e.g. new statistical methods or biomarker/prognostic indicators
(vi) Independently verify the analyses that the trial team has published or presented.
The scope of this document is to cover data from any type of study, including RCTs, cohort studies and meta- analyses, and both individual patient data and aggregate data. It covers how studies led by the OUCRU network and sponsored by University of Oxford can be made discoverable and the process for sharing data. This includes studies run by the Hanoi Unit at the National Institute of Infectious & Tropical Diseases (NHTD), as well as the satellite research units in Kathmandu (Nepal) and Jakarta (Indonesia).
The scope of this data sharing policy is focused on data sharing requested by external parties (not including journals, sponsors or collaborators).
For this purpose, “data” covers raw data, processed data on individuals or aggregated, summarised data in the form of reports.
The document also covers the release of stored samples. A dataset is a collection of data to be used for a specific purpose, as formulated in the proposal. Under some definitions (e.g. EMA), data has a broader definition, including files, test results, etc.
The DMC, usually independent (→IDMC; can also be termed DSMB), is the only body to routinely see accumulating comparative data from a trial. The IDMC is considered the guardian of the interim analyses, and is well placed to consider the implications of early requests for data release.
“Data sharing” encompasses data release and data transfer to external applicants.
The Data Sharing Plan is the term used here to describe a document, developed at an early stage, detailing some aspects of data sharing methods and principles.
The ability for the existence of a dataset or trial to be discovered.
TMG comprises the key internal and external people involved in running a trial, including the Principal/Chief Investigator. These members have been key in developing the trial protocol, conducting the trial and collecting and analysing the data.
The TSC is the trials’ executive body ensuring compliance to Good Clinical Practice (GCP), and is central to major decisions for the trial. The committee may be comprised of a number of independent members, including the chair, plus key members of the TMG.
The OUCRU SC determines the strategic objectives of OUCRU and defines the financial and managerial policies to achieve these objectives. The OUCRU SC has the final responsibility for reviewing and approving external data sharing requests.
This policy sets out the principles for OUCRU-coordinated research and applies where OUCRU or Oxford University is the Sponsor. Explicit input should be sought at an early stage in the project where another organisation is the Sponsor. A data sharing or clinical trial agreement should be in place prior to any data collection.
The process of receiving and processing applications, developing agreements, producing and transferring data, and responding to subsequent requests for clarification of data involves a broad range of functions across the unit. In the order of the Standard Operating Procedures (SOP) training matrix, this could be:
(i) OUCRU employs a controlled access approach for data sharing requests by external parties.
(ii) No data should be released that would compromise an ongoing trial or study.
(iii) There must be a clear scientific or other legitimate rationale for the data to be used for the requested purpose.
(iv) OUCRU’s ability to share high-quality data in a timely manner will require considerable resources. We will therefore levy a charge on all external requests (independent from a standard scientific collaboration) for OUCRU data, which will be used to support the internal infrastructure required to share data effectively and efficiently.
(v) Data exchange complies with Information Governance and Data Security Policies as well as relevant institutional and ethics regulations in all of the relevant countries.
A key issue for sharing is, knowing what information is available and when it will be available. To this end, OUCRU research should be “discoverable”.
All OUCRU clinical trials will be registered on an appropriate clinical trials register and will be discoverable through this process.
Trials will also be listed on the OUCRU website.
OUCRU funded projects and studies other than randomised controlled trials (RCTs) may be found on the OUCRU website.
All applicants submitting funding proposals to the Medical Research Council, Wellcome and the Gates Foundation are required to include a Data Sharing Plan as an integral part of the application. Other funders may also ask or may be expected to follow suit in the future.
The Data Sharing Plan presents early information on data sharing principles and methods. It may also include a description of data collection, management, storage and curation, and covers security risks that would later be covered in a risk assessment and a Data Management Plan.
Standardised terms and definitions should be used if appropriate and possible, including from the design stage of a study in order to facilitate future data sharing. For example:
In line with journal and sponsor’s regulations OUCRU has several options for data sharing, including but not limited to
A number of stages may be required in the assessment process of data requested by external parties. The process for access to data should be made clear to potential applicants. A template request form is provided which includes a diagram of the default process.
Data requests may continue to be received long after the trial has closed. The Principal/Chief Investigator should remain reachable for as long as is feasible in order to consider such requests.
As with any application, there must be a strong scientific rationale or other legitimate reason. Key points should be considered:
The applicants should detail their request on a request form.
Further items may be added to this if required and variations are acceptable for particular studies providing the process is well documented in the SC.
Completing the request form can be an iterative process, with updates from each stage of the review. The applicants can be asked to specify the information that will allow the reviewers to consider all relevant aspects:
The comments of the reviewers should be recorded at each stage. A form for this purpose is provided as a template, which may be sent out to reviewers. Feedback should be provided, as a summary, to the applicants.
The contracts team should be notified at the appropriate stage of the review process so that suitable agreements can be drawn up promptly after a successful review. Agreements will be made at an institutional level rather than an individual level in all but the most exceptional circumstances.
Researchers in receipt of data must store it securely, according to Information Governance and Data Protection regulations. They are expected to process and publish reports according to their initial plans.
Agreements should specify the boundaries of use of the data, including, for example, that the data can only be used for the purposes for which they have been released.
Applicants should not use the data to identify individual patients, unless this is a pre-specified goal for the purpose of record linkage.
The data protection legislation of all countries involved in the transfer must be considered. Particular care must be taken in preparing or releasing identifiable or sensitive data on individuals. Identifiers should be removed prior to disclosure wherever possible e.g. no names or detailed geographical locations; using ages or years of birth instead of full date of birth. Consideration should be given to further anonymising the data by including new ID numbers and breaking the link to the original dataset. This choice will largely depend on how the data are to be used and whether further linkage would be required. The time and effort required to prepare the appropriate dataset should not be under-estimated. Data should be transferred in an appropriately secure method, after discussion with the Data Management and IT department.
Recipients should acknowledge receipt of data and should check immediately for any problems.
The recipients of data should update OUCRU SC at time of completion of their project and share arising publications.
An audit trail of applications and approvals is required. It is important to record in a tracking log that data have been requested and, if released: when, to whom and for what purpose. A Unit-wide database for requests will be developed and maintained by the CTU.
OUCRU datasets may include linked data from other sources e.g. HTD data on hospital invoices. These are considered to be part of the dataset and can be included in data released from OUCRU, subject to agreement and approval by the original source.
OUCRU data that are provided to other people may be subject to further requests for sharing. For example, a trial dataset may be released to Researcher B for inclusion in a meta-analysis, and Researcher C might then request access to the full dataset of Researcher B for a further analysis.
Onward sharing is not permitted without OUCRU consent. Generally, onward sharing to Researcher C will be permitted if Researcher C wishes to use the data in a project that has a purpose in line with the original release to Researcher B. If Researcher C wishes to use the data for a purpose that was not covered by the data release agreement, Researcher C should contact OUCRU directly for data.
This policy was last reviewed in 2017. The web version was updated in 2023.
For questions regarding Data Sharing, please contact the Clinical Trials Unit at OUCRU.