“De-identified” does not mean that the data cannot be linked. It means that information such as your name, date of birth, and postcode is removed before researchers can access the data.
The linkage itself is carried out securely by authorised organisations using identifying information that researchers do not see. Once the different sources have been linked, identifying information is removed and researchers only have access to the de-identified linked dataset. This process us designed so that the people carrying out the linkage and the researchers using the data do not have access to the same information.
The process uses a “split-file” approach:
‘File 1’ contains the information needed to identify and match a #BeeWell participant (e.g. their name, date of birth, sex, and residential postcode) together with information about linkage permissions (i.e., what data sources can and cannot be linked, based on what participants were originally told). #BeeWell sends this securely to UK LLC’s trusted linkage partner, Digital Health and Care Wales (DHCW).
DHCW securely passes the relevant identifying information and a study ID to NHS England, where the Master Person Service uses the identifying information to find the participant’s NHS record.
Once the match has been made, NHS England creates ‘File 2’. This contains the relevant health information, but not the identifying information used to make the match. Instead, it contains an encrypted study ID that allows the health data to be connected to the corresponding #BeeWell data.
File 2 is then securely transferred to the UK LLC Trusted Research Environment, where researchers can analyse the linked, de-identified data.
So, the identifying information is used to make the match, but it does not form part of the dataset that researchers subsequently access. The File 1 and File 2 separation is an important safeguard: the information used for identification and linkage is kept separate from the health information provided to researchers.
In simple terms: File 1 is used to answer “which NHS record belongs to this #BeeWell participant?” and File 2 contains the resulting health data, linked using a study ID rather than their name, address, or other direct identifiers.