CPRD GOLD and linked ONS mortality records: Reconciling guidelines

Area of Science:

• Epidemiology
• Health Informatics
• Data Science

Background:

• The Clinical Practice Research Datalink (CPRD) GOLD is a key UK primary care dataset for epidemiological research.
• Office for National Statistics (ONS) death data for England is the gold standard for mortality data and can be linked to CPRD GOLD.
• While CPRD GOLD's death data accuracy is sufficient for most studies, challenges exist in managing linked ONS and CPRD GOLD datasets.

Purpose of the Study:

• To provide guidelines for reconciling mortality information from both CPRD GOLD and linked ONS datasets.
• To enhance the quality, reproducibility, transparency, and comparability of clinical research using linked mortality data.
• To address challenges including linkage coverage, correctness, regional limitations, and data discrepancies.

Main Methods:

• Developed recommendations for managing mortality data from CPRD GOLD and linked ONS datasets.
• Considered linkage coverage period, correctness likelihood, regional limitations, and data discrepancies.
• Implemented these guidelines in a practical SQL algorithm for researcher use.

Main Results:

• Provided detailed guidelines for reconciling mortality data between CPRD GOLD and linked ONS death datasets.
• Addressed the strengths and limitations of both datasets.
• Developed an SQL algorithm to facilitate the consistent application of these guidelines.

Conclusions:

• Detailed guidelines for reconciling CPRD GOLD and ONS linked death datasets have been established.
• The consistent application of these guidelines, supported by an SQL algorithm, can significantly improve clinical research.
• Potential improvements include enhanced data quality, reproducibility, transparency, and comparability in clinical research.