Augmenting maternal clinical cohort data with administrative laboratory dataset linkages: a validation study

Area of Science:

• Health Informatics
• Data Science
• Bioinformatics

Background:

• Big data and large language models offer potential in healthcare.
• Protecting patient confidentiality is a major challenge in healthcare data analysis.
• Augmenting clinical data with administrative laboratory data can avoid demographic information.

Purpose of the Study:

• To explore an alternative method for examining patient files from large administrative datasets.
• To link external data to the National Health Laboratory Services (NHLS) database using specimen barcodes.
• To quantify the performance metrics of this data linkage approach without using demographic information.

Main Methods:

• Linked external hospital data to the NHLS database using specimen barcodes.
• Utilized specimen barcodes for deterministic data linkages.
• Quantified linkage performance metrics on a validation sample of 1200 records.

Main Results:

• Achieved a 95% success rate in linking NHLS data to external hospital data.
• Obtained 87% exact matches and 9% matches with typographic correction.
• Identified 5% mismatches or duplicates in the administrative data.

Conclusions:

• Specimen barcodes are reliable for deterministic data linking without demographic identifiers.
• This approach effectively creates large, linked health datasets.
• Patient confidentiality is maintained while enabling robust data linkage.