Can resident registration expiration statistics due to death be used for near-real-time mortality tracking? A validation study using 2023 data from South Korea

Real-time mortality tracking is essential for public health surveillance, especially during emergencies such as the COVID-19 pandemic. In South Korea, delayed availability of Vital Statistics (VS) data has hindered timely mortality monitoring. This study evaluates whether National Administrative Data (NAD) on resident registration expiration due to death, provided by the Ministry of Interior and Safety (MOIS), could serve as a reliable alternative for near real-time mortality surveillance. We compared mortality counts between VS and NAD for 2023 at multiple geographic levels (county, province, and nation) and across demographic strata (sex and age groups). The analysis was conducted in 3 stages: comparing overall mortality counts, analyzing county-level distributions, and assessing equivalence through correlation analyses, scatter plots, and density plots. NAD showed strong agreement with VS at the national level, reporting only 0.4% more deaths overall (0.2% for men, 0.6% for women). Notable differences were observed in early childhood mortality, with NAD showing 16.8% fewer deaths for neonates (age 0) and 14.8% more for ages 1-4, as well as in monthly variations (5¬-9%). Correlation analyses indicated extremely high consistency between the 2 data sources across all geographic levels (correlation coefficients ≥0.999), especially at the national and provincial levels. NAD provides a reliable alternative to VS for real-time mortality surveillance in South Korea, providing comparable accuracy with much-improved timeliness. Although some variations are present in specific age groups and monthly trends, these can be addressed through appropriate analytical strategies. The recent availability of sex- and age-specific data in NAD since 2023 establishes it as a valuable infrastructure for mortality surveillance.