Ethnicity

In HERON-UK ethnicity is standardised to the NHS Ethnic Category. For considerations around mapping ethnictiy for research we suggest starting with this paper.

Source data has at most one ethnicity recorded per person

Map ethnicity to the person table using the following

  • race_source_concept_id: use NHS Ethnic Category codes using the NHS Ethnic Category.
  • race_concept_id: map from NHS Ethnic Category codes to standard Race codes (note, these mappings were introduced in the August 2025 vocabulary release.) If race_source_concept_id is missing for an individual, set race_concept_id as missing.

Source data has multiple ethnicity records per person

Map ethnicity to both the person and observation tables.

For the person table, use NHS Ethnic Category codes as above. The person table can only contain one record per person so to choose a single ethnicity per person, using the record considered to be highest-quality given your source data (e.g. choosing most common code per person and if two or more codes occur with equal frequency, select the most recent one.)

For the observation table, include each ethnicity record for individuals using NHS Ethnic Category codes for source values, mapping these to their standard concepts. This accommodates datasets with multiple ethnicity records per patient and allows for customisation at the analysis stage (although for most network studies we would expect to use ethnicity as recorded in the person table).