Introduction, objective and scope

Linking information about the same person within a data set and across data sets is essential for many analytical uses. Different data sets can include different identifying fields, data submitters may record information differently about the same person and submitted data can include data quality errors.

To verify the submitted person details and maximise opportunities for linkage between data sets, NHS England has developed the Person_ID, a key that identifies individuals. This uses a system known as the Master Person Service (MPS) to trace and verify key identifiers against the Personal Demographics Service (PDS). MPS can match a person’s details where identifiers may be partially missing or slightly different, allowing the assignment of a consistent Person_ID.

NHS England data sets that assign a Person_ID to each record during processing include

• Hospital Episode Statistics (HES)
• Mental Health Services Data Set (MHSDS)
• Maternity Services Data Set (MSDS)
• Community Services Data Set (CSDS)

The Person_ID can be either the NHS number if a match for the record was found in PDS, or the MPS_ID (a unique identifier from the MPS record bucket) if the record was not found in PDS but sufficient demographic information was provided to create a persistent unmatched record. Future occurrences of such records may be matched against the MPS record bucket if it is again not possible to find a matching record in PDS. MPS_ID is also sometimes known as UPRI 1 (unmatched person record identifier 1). If the queried record could return neither an NHS number, nor an MPS_ID, then a one-time-use identifier is generated, sometimes known as UPRI 2 (unmatched person record identifier 2).

The NHS number and MPS_ID might be recurring identifiers across data sets, and therefore can be used to link patients.

A consistent Person_ID across data sets can be tokenised allowing users to count people or perform linked analysis without any need to access identifiable information. The masked identifier is sometimes known as the Token_Person_ID. Most users will work with the Token_Person_ID rather than the Person_ID. Specialist users from the Trusted Research Environment (TRE) can only see this tokenised identifier.

The aim of this handbook is to provide users of the Person_ID in the HES database with reference documentation on what the Person_ID is, how it is derived via the MPS, how the data flows between services (Data Processing Services (DPS) and Spine), and how to interpret the output information associated to the linked Person_ID.

This handbook contains some technical details of the MPS so users can understand how records are matched to PDS, however, it is not intended to be a complete technical guide for MPS, nor to provide recommendations on how to improve the quality of the linked records.

When a person’s identity is traced by the MPS, additional information from PDS such as registered GP practice can be stored to enrich a data set. This use of the MPS is outside the scope of this document – information on the use of this functionality for a specific data set may be found in the specification for the relevant data set.

MPS is used in other settings. While this handbook may include information relevant to other uses of MPS, it should not be relied upon as a complete or accurate representation of use of the MPS in other contexts.

Please direct questions to [email protected].

This email address should be used to give feedback about this document or to ask for clarification about any of the content of this document.

Please note, we cannot investigate specific anomalies in data returned by MPS. However, the Empirical examples section contains some specific examples of edge cases, which may be helpful.