Identifying a national death index match

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• National Death Index (NDI) data are vital for determining survival status in research.
• NDI provides candidate matches, but accuracy can vary, especially with multiple potential matches.
• Established Populations for Epidemiologic Studies of the Elderly (EPESE) data are frequently used in long-term health research.

Purpose of the Study:

• To develop and validate an algorithm for improving the accuracy of National Death Index (NDI) matches.
• To assess the agreement between NDI and Social Security Death Index (SSDI) death dates using specific matching criteria.

Main Methods:

• Utilized data from the Duke University EPESE cohort, NDI, and SSDI for 1,896 individuals born before 1922 and alive as of January 1, 1999.
• Developed a matching algorithm based on Social Security number and combinations of personal information (name, date of birth).
• Compared NDI and SSDI dates of death for cases with and without the required matching information.

Main Results:

• An algorithm matching Social Security number plus specific personal information achieved 94.7% agreement between NDI and SSDI death dates.
• Candidate decedents lacking this specific information combination showed only 12.3% agreement between NDI and SSDI death dates.
• The proposed algorithm significantly enhances the precision of NDI match identification.

Conclusions:

• An easily applicable algorithm significantly improves the accuracy of identifying National Death Index matches.
• This method is essential for reliable survival status determination in epidemiological and clinical studies.
• Enhanced NDI matching contributes to more robust research findings in aging and public health.