A pipeline for harmonising NHS Scotland laboratory data to enable national-level analyses

Area of Science:

• Health Informatics
• Bioinformatics
• Public Health

Background:

• Electronic health records (EHRs) are crucial for data-driven health research.
• Medical laboratory data in Scotland exhibits significant heterogeneity across local health boards.
• Existing national datasets (hospitalization, prescribing, death registers) lack standardized laboratory data.

Purpose of the Study:

• To develop an open-source software pipeline for harmonizing regional laboratory data in Scotland.
• To create a unified, national-level laboratory dataset for research purposes.
• To reduce the effort and time required for researchers working with diverse regional laboratory datasets.

Main Methods:

• Obtained sample laboratory data from four regional Safe Havens in Scotland.
• Compared variables and mapped them to 11 FHIR and 2 Scottish-specific standardized terms.
• Developed data transformations to standardize units and used Read Codes and SNOMED CT for test encoding.

Main Results:

• Identified 180 key test codes representing 98.7% of all laboratory tests performed in Scotland.
• Successfully developed transformations to standardize test result units.
• The pipeline effectively encoded tests using both Read Codes and SNOMED CT.

Conclusions:

• Validated a harmonisation pipeline for Scottish laboratory data.
• The pipeline facilitates the generation of clean, harmonized laboratory data at a national level.
• This open-source solution can be readily adopted by researchers and Safe Havens to improve data accessibility and usability.