In practice
Below are illustrations of topical aspects to help studies meet the requirements and expectations in practice. Various sections of the Policy on Sharing of Research Data from Population and Patient Studies refer to these practical illustrations.
- Study purpose and priorities for sharing
- Privileged use of data by study team
- Categories of data availability
- Bona fide research
- Confidentiality and data security
- Meeting the costs of data-sharing
- Commercial use of data
- Criteria for considering data access requests
- Models for independent oversight of research data access
Study purpose and priorities for sharing
A study policy and the supporting information should include:
Study purpose
- The research themes and questions for which the study is (and has been) funded so that potential users can identify the unique potential of the study for productive new-uses.
Specific opportunities and priorities for data-sharing
Study policy in relation to maximising value through responsible sharing is clear, and specific priorities for sharing are transparent, e.g.:
- Whether priority is given for new uses that complement the study objectives and/or for which the study is a unique resource.
- Whether the study is open to unforeseen opportunities – especially for interdisciplinary synergy, research translation and knowledge transfer.
- Whether there are opportunities for non-research uses such as teaching and dissemination.
- Whether, when and how biomedical materials (e.g. human tissue, DNA) and other resources may be available for sharing.
- How requests from projects with closely similar research questions are handled; for instance, whether such groups are encouraged to collaborate with each other, and if not how competing projects are prioritised (see access governance).
- Illustrative examples of previous/current data-sharing collaborations and their most significant outputs (datasets and publications).
- Whether and how additional questions or measures can be added to future waves of data collection - including indications of the likely costs involved.
Privileged use of data by study team
Over the data lifecycle, a balance needs to be struck between the needs of the study team and those of potential new users, reflecting their actual and potential contribution to maximising the value of the data.
The intellectual and managerial efforts that go into study design, achieving funding, data collection, management and analysis need to be incentivised and rewarded. It is reasonable for principal investigators (PIs) to have specific arrangements for them to analyse the data and to publish results. These arrangements may comprise a limited period of exclusive use.
In practice, facilitated collaboration with external researchers can significantly increase the value and quality of the data even early on, e.g. through contributing to data cleaning and creating derived variables.
The study policy on sharing should define the terms of privileged use by the PIs, such that they are transparent and can be verified. If the policy provides for a period of exclusive use, it should take into account the following considerations:
- Where a study’s principal investigator(s) are funded to deliver a specific programme of research, a limited period of privileged use of the data they have acquired and to which they have added value is reasonable.
- For practical reasons this time period may be indicative and need to be revised if delays occur (e.g. in recruitment). Different periods may be applied to different datasets, e.g. to take account complexity of cleaning and documentation.
- Timings will depend on a study’s data collection patterns.
- However, extended privileged use by PIs – e.g. beyond that approved through peer review - could be controversial and open to challenge.
- In relation to timing, the terms could, for instance, be expressed as follows: “6-months after the end of the current grant period”, “12-months after new data collection to allow for data cleaning and documentation” or “3-months following the first publication of findings based on the data.”
The principles of a study’s privileged use of data should be explicit in the study’s funding proposals, as part of the Data Management Plan (if such a plan is required by the funder).
Categories of data availability
Studies may find the following categorisation of data availability helpful:
1. Unrestricted availability: Anonymised data (e.g. summary tables) for which the risk of disclosure (identification of individual participants) directly or through association with other data sources is extremely low, which can safely be made readily accessible without restriction (“public”).
2. Independently available: Data available in principle for use by independent new, bona fide research, within the terms of participant consent and not restricted by IPR, prior collaborations or other reasons, and for which the necessary metadata are well documented and available.
3. Dependently available: Data restricted by the scope of historically obtained participants’ consent, or other reasons such that sharing (and possibly analysis) can be effected only with significant facilitation by the study team (e.g. where data and metadata quality is not sufficient for independent use).
4. Unavailable: Data not available for sharing because of ethical, IPR, prior exclusive agreements or other constraints.
Bona fide research
“Bona fide” is frequently used in relation to sharing of research data, but is rarely defined. For the purposes of this guidance, key characteristics of bona fide research can be considered to be as follows:
- An intention to generate new knowledge and understanding using rigorous scientific methods. (This includes discovery research, development and validation of methodology and technology, validating and challenging previous findings, and pilot research). And…
- An intention to publish the research findings and share the derived data in the scientific community, without restrictions and with minimal delay, for wider scientific and eventual public benefit. (Recognised constraints include a short prepublication delay to ensure proper management of intellectual property). And…
- The intended activities are not inconsistent with legal and ethical requirements or widely recognised good research practice.
In practical terms, a research project or proposal that has been approved by a recognised funder should normally be considered to be “bona fide”.
A bona fide research organisation is one that has the capability to lead or participate in high quality, ethical research. It will have a public commitment to adhere to recognised research and information governance good practice. (It is not a requirement that such research is the primary business of that organisation, or that all of the research undertaken by that organisation is published. Nor is it a requirement that the organisation be publicly funded.)
A bona fide researcher is a person with
- the professional expertise and experience to conduct bona fide research
- a formal relationship with a bona fide research organisation that requires compliance with appropriate research governance and management systems
Confidentiality and data security
Maintaining confidentiality and data security are crucial to public confidence in research.
- The study policy on data sharing should set out the standard use conditions for secure management of confidential and sensitive information.
- The conditions should distinguish appropriately between data that are non-disclosive, potentially identifiable (although anonymised) and identifiable (not anonymised); this may relate to their availability.
- The standards required of external users and study team members should be the same, and should be appropriate to the risks.
Study policy and practice must conform to
- MRC guidance on Personal Information in Medical Research
- MRC guidance on human tissues
- MRC information security policy (for MRC units and institutions)
Meeting the costs of data-sharing
The study should set out a policy for meeting the costs of sharing. Policies may cover the following:
1. Whether publishing (or having readily available) anonymised data tables is an effective means for supporting bona fide new use at low cost to the study and ‘free’ to the research community.
2. Whether the scale and costs of data-sharing merits building an explicit budget-line into the study’s MRC next funding request.
3. Whether new users are asked to contribute (e.g. through grant support) to the costs of preparing data and facilitating use (or analysing data on behalf of new users) and if so what proportion. The greater the resource needed to support new use, the more reasonable it will be to recover some or all of the costs of data preparation and facilitation.
4. Whether the study will absorb some of the sharing costs for proposed new use that closely complements the study team’s MRC-approved research programme.
5. Whether for commercial use the study charges a premium over and above the costs of preparing the data for sharing, reflecting the value of the intellectual asset. Such arrangements must be on a non-exclusive basis. They should be consistent with RCUK guidance on management of intellectual assets.
Commercial use of data
MRC encourages researchers to work in partnerships, also with commercial organisations.
- Studies may charge for cost recovery for data sharing
- Studies may charge a premium for their Intellectual Property Rights (IPR)
- Proposed use by commercial organisations should be for bona-fide research and educational uses only, that are consistent with the public interest
- Exclusive relationships for the use of data should be avoided, both for commercial and non-commercial purposes.
- The MRC technology transfer office can advice on the charging for IPR and should be consulted by PIs.
Criteria for considering data access requests
The criteria for approving data access need to be transparent and appropriate to the study, the funder’s intentions and the participants’ interests. New use should be consistent with maximising the value of the data, across the lifetime of the study, for public good. The following should be taken into account in developing study-specific criteria:
Study purpose and priorities for sharing
- The request is consistent with the study policy on sharing.
- New use is compatible with the requirements of MRC for the study team to deliver on their MRC-approved research programme.
- Where a proposed new use is similar to an existing use, the new use should be compatible with maximising the value derivable from the data
- There is not an obviously more appropriate source of data for the request.
Capability and capacity to use the data for high quality research
- The proposed research is bona fide and is methodologically sound: e.g. the analysis plan is robust and the requested sample structure and size will be sufficient to generate meaningful results.
- For research use (rather than educational use), the requestor is a bona fide researcher working within a bona fide research organisation.
- Any other new uses of subsets of the data, e.g. for teaching, are specified and similarly the requestor can demonstrate the capability and capacity to use the data.
Lifecycle
- Relevant data of the necessary quality are held by the study and can reasonably be made available for use (e.g. collection, validation and cleaning have been completed).
- Sharing of the data now does not obviously compromise the ability to derive additional value later.
Consent, ethics and confidentiality
- The purpose of the proposed research is consistent with the study’s own ethics approval and participant consent.
- There is a clear route for regulatory permissions (e.g. ethics approval) to be sought and for approvals to be notified to the study
- There is not an unacceptable risk to the confidentiality of the participants’ identities or to other aspects of data security.
- There is a clear policy on whether, when and how re-identification of participants may be considered, and under what controls.
- Sharing with organisations outside the UK conforms to the DPA eighth principle that personal data are not transferred to another country outside the European Economic Area that does not have adequate protection for individuals’ personal information.
Partnerships and commercial use
- The relationship between new user and the study team is equitable in relation to the expected benefits and management of risks.
- The proposed use of the data does not pose an unmanageable risk to continued participation of the cohort members.
- Proposed use by commercial organisations or by individuals and organisations outside the UK is for bona-fide research, educational or other uses is consistent with the public interest, and is on a non-exclusive basis.
- Commitment to publishing the findings in peer-reviewed publications is, for the purposes of this guidance, an important test of bona fide research.
Resources
- The study team has the expertise and capacity to support the request (including essential, ongoing facilitation).
- Funding is available in principle, from the study or the new user, to support the production of data for the user (or of running an analysis on their behalf).
Other criteria
- A study may specify other requirements, e.g. in relation to the management of intellectual property, deposition of derived data back with study, further use of derived data, and communication about the study. These will normally be a reflection of the terms of the study’s standard data sharing agreement.
Models for independent oversight of research data access
The study director/PI is responsible for the integrity of the process and for the decisions on data access. The access governance arrangements must be effective and proportionate for particular study. They must have an independent element and the key decisions must be transparent, equitable and documented. “Independent” for these purposes means ‘being impartial towards decisions about the study’s data’ and so will exclude, amongst others, individuals who collaborate substantively with the study.
Two following models are illustrative and can be adapted by Directors/PIs to the circumstances of their own studies.
Model 1 is appropriate for studies that have or anticipate a significant sharing activity. Model 2 is more appropriate for a study that anticipates only occasional requests, which are unlikely to give rise to scientific, technical, ethical or legal issues.
Model 1: All requests are considered by an Access Committee
- The committee considers all formal requests for access within an appropriate timescale, including new uses by the study team itself (beyond the funder-approved programme). It may also be tasked with reviewing the study’s access policy and procedures, and with annual review of the performance in terms of outcomes and service.
- The committee comprises a balanced set of scientists with expertise of the purposes for which data are likely to be requested, supplemented ad hoc by further experts as required.
- There may also be roles for members with other kinds of expertise, such as legal and ethical, or experience as a study participant.
- The committee chairman is independent of the study (neither a co-investigator nor a collaborator). One or more other members is also independent.
- Experts who understand the study thoroughly can make a significant contribution to access decisions. Consequently, it may be justifiable to include members who have an interest in the study (through collaboration or otherwise using the data). Potential and actual conflicts of interest must be managed judiciously according to good practice.
- The committee is advisory to the study director/PI, who is not a member but attends. Other study team members are also observers and not members of the committee.
- The policy may provide for specified classes of “straight forward” request, considered highly likely to fulfil the criteria, to be triaged out and reviewed by a “fast/light” process and reported post-hoc to the committee. The policy may provide for those reviews to be conducted ad hoc by (i) a subset of the committee, or (ii) the study director / PI. To ensure transparency, high standards of triage, feedback and reporting to the independent committee must be maintained.
Model 2: Access decisions are periodically reviewed by an independent Access Advisor
- The study team formally reviews access requests for proposals. All significant decisions (approval, referral back for further information, and decline) are documented for subsequent independent review.
- An advisor (or committee) with appropriate expertise, independent of the study, is appointed to periodically review the outcomes of access requests post hoc. The reviewer may also be tasked to advise on a study’s access policy and procedures.
- Individual requests may be referred to the advisor for advice if difficult issues arise, e.g. a risk to the data, participants or study, or to depletable resources.
- It may be in the best interests of the study and requestors that an intention to decline a request is first referred, with a justification, to the advisor for advice.
- The advisor may advise the Director/PI that they need external expert advice in order to formulate their own advice to the study.
- The study director/PI is responsible for access decisions.