Adding value to clinical data by linkage to a public death registry

Area of Science:

• Health Informatics
• Data Management
• Public Health

Background:

• Clinical data repositories (CDR) often lack comprehensive patient outcome data.
• Integrating external mortality data enhances patient record completeness.
• Accurate patient matching is crucial for data integration.

Purpose of the Study:

• To describe a methodology for merging statewide mortality data into a hospital's CDR.
• To assess the impact and efficiency of this data integration process.
• To demonstrate the utility of enriched CDR data for identifying at-risk patient groups.

Main Methods:

• Implemented three linkage passes using Social Security Number, Last Name/Birth Date, and Last Name/First Name.
• Refined initial matches using a scoring algorithm based on identifier matching quality.
• Validated the scoring algorithm with a cohort of patients who died at the health system.

Main Results:

• Achieved a 97% update rate for deaths from the state mortality files into the CDR.
• The scoring algorithm effectively refined initial, broadly inclusive matches.
• Demonstrated the system's ability to identify at-risk patients, such as those who committed suicide.

Conclusions:

• The described approach efficiently and inexpensively enriches hospital data with vital outcomes information.
• This method significantly improves the completeness of patient records within the CDR.
• Enriched data facilitates better identification of patient populations requiring targeted care.