The Fives Safes is a framework to help make decisions about how to make data available for research that may be confidential or sensitive, while making sure that the volunteers who have provided it are protected at every opportunity.
The Five Safes framework helps guide us to allow confidential and sensitive data to be used for research, while protecting the rights of those whose data is being accessed.
The Five Safes were developed by the Office for National Statistics and is now used in most Trusted Research Environments in the UK and internationally.
We consider much broader issues about ethics, fairness and ensuring transparency. However, the Five Safes allow us to focus on how we allow “Safe” researcher access of data while protecting participant confidentiality.
For all our data sharing decisions, we consider the following:
Safe projects – is the use of data appropriate?
Yes.
All data use goes through multiple reviews to ensure the use of data is appropriate and that only the relevant data needed for the task is made available. These reviews are conducted by our partner Longitudinal Population Studies, public contributors and more generally by organisations such as the NHS, government regulators and independent auditors.
Our resource exists to support research in the public good. These data will enable researchers to answer important questions that aim to improve health and care treatments, public health policy and learn more about how our behaviours impact our everyday lives.
Researchers must demonstrate that their research question has the potential to improve the public good. They also have to justify that the data they use are relevant to their question. These processes include our public contributors. Our audit researchers check researchers are using the data appropriately.
Ethical review is important to ensure the use of data is fair and appropriate. We have our own independent research ethics. Where a researchers question relates to very sensitive topics or uses new and complex methodologies we ask that they gain their own independent ethical approval.
Safe people – can users be trusted to operate in the right way?
We make sure only approved researchers who are professionals with the appropriate training required can work in the Trusted Research Environment. Researchers must work for legitimate research organisations, for example, Universities, the NHS and wider government departments, and research charities.
Research users working in the Trusted Research Environment must have taken the ‘safe researcher’ professional training, and passed the exam set by the Office for National Statistics. This trains researchers in maintaining data confidentiality.
Researchers that have received a successful application, and their host institution, require to sign contracts to maintain confidentiality and integrity. We undertake audits to make sure that everyone is working appropriately within the Trusted Research Environment.
Safe data – is there a risk the data might allow individuals to be identified?
It is highly unlikely that any one individual could be identified.
All data held in our Trusted Research Environment is de-identified. Neither our staff or research user ever see your names, commonly used identifiers (such as NHS ID) or your address or neighbourhood data. We also make sure other identifiers – such as the GP practice you use, school you attend(ed) or who you work for are never accessible. See our privacy notice for more detail as to how this works.
Does the system limit use?
Yes.
Our Trusted Research Environment only allows research users to study and analyse data within the secure environment. They cannot copy or remove data. They are only permitted to access the minimum amount of data to carry out their research project.
Are checks in place to make sure findings meet confidentiality standards?
Yes.
Once a research user has completed their research they have to generate completely anonymous results – the statistical evidence from their research. They are trained to do this as part of their “safe researcher” training.
Once they are ready, they ask us to export these anonymous results so they can be used in reports, evidence to policy makers and academic publications. The anonymous results are checked by two highly trained and audited statisticians to confirm they are anonymous and safe to leave the Trusted Research Environment.