Validation of operational definitions of mortality in a nationwide hemodialysis population using the Health Insurance Review and Assessment Service databases of Korea

Background Health Insurance Review and Assessment Service’s (HIRA) claims data have been used in studies of hemodialysis patients even though information about mortality is not provided in this database. Mortality analysis using HIRA data has been conducted using various operational definitions that have not been validated. This study aimed to validate operational definitions of mortality for maintenance hemodialysis patients that have been used when analyzing the Korean HIRA database. Methods This study utilized claims data of the Korean National Health Insurance Service (NHIS) between January 2008 and December 2019. We estimated mortality based on operational definitions applied in previous studies using the HIRA database and compared it with NHIS mortality information to validate accuracy. Results A total of 128,876 patients who started maintenance hemodialysis between January 2009 and December 2019 were analyzed. The accuracy of estimated mortality was the highest at 96% in the group where mortality was defined as an absence of claims data for 150 days. If the period of no claims data was set to 90 days or less, there was a risk of overestimating the mortality for the entire study period. When it was set to 180 days or more, there was a risk of underestimating the mortality, as the follow-up time was close to the end of the study period. Conclusion When mortality analysis of maintenance hemodialysis patients is performed using HIRA data, it is most accurate to set the operational definition period as the absence of claims data for 150 days.