Getting the most from South Korea's health insurance database

Kantar was able to offer our client the relevant know-how around successfully using and optimising South Korea's HIRA database.


A leading pharmaceutical company wanted to compare the demographic characteristics and the proportion of comorbidities between patients of a target disease with a matched cohort of patients without the target disease, using a healthcare real-world database in South Korea. They also wanted to examine treatment patterns and healthcare resource use (HCRU) over a 10-year period. Although health insurance claims data compiled by South Korea’s Health Insurance and Review Assessment (HIRA) is a valuable source of healthcare real-world data, the resource has many limitations regarding data analysis since the data was not loaded for this purpose. The company engaged Kantar in Korea because of its expertise in data analysis, protocol development, IRB approval, and reporting services that cover the entire scope of research projects from the start to the end of studies.


With the expansion of civilian accessibility to South Korea’s HIRA health insurance claims data five years ago, data analysis boundaries were widened for the health and medical-related database. Today, this access is being used to satisfy numerous real-world health objectives, and target study areas are being expanded and diversified. Kantar has the know-how when it comes to using and optimising the HIRA database, as we’ve built up our experience by communicating with HIRA officials and analysing HIRA data. From this groundwork, and our consultation with key opinion leaders (KOL), we were able to establish accurate patient definitions and proper data settings for the analysis.


To examine the 10-year data trend requested, we analysed targeted data from 2007, 2011 and 2016. And, since it was a comorbidity study, we also brought over the corresponding patients’ annual medical information. Once that was completed, we suggested an appropriate data extraction procedure for the matched control group setup, as well as an analysis method that relates to the structure of the database.


When using the HIRA database, patient definition can be the most important and most difficult aspect to meet. In this study, where there were many comorbidities and concomitant medications of keen interest, we had numerous meetings with KOLs, organised the definitions by searching published papers, checked the number of patients affected by each disease, and continued with our efforts to set up the proper definitions within the feasible boundaries. A full statistical analysis plan was written about the analysis and provided to our client.

Get in touch