Income

By linking with the NHID Eligibility database, we obtained information on national health insurance premiums at the end of the year (December 31) as a proxy for income levels in the NHSD. When the year-end health insurance premium was not available, the most recent national health insurance premiums were used instead. In a prior study conducted by the NHIS,20 the number of population of the NHID Eligibility database was almost the same as that from the Korean Statistical Information Service (KOSIS). In addition, the concordance correlation coefficients of the district-level crude mortality rate and life expectancy between the NHID Eligibility database and the KOSIS ranged between 0.997–0.999 and 0.914–0.990, respectively. NHID, therefore, is considered a useful data source for monitoring the health status of the entire Korean population at the level of district. The usefulness of the national health insurance premium as a proxy for income levels in monitoring the income gap in health outcomes has been demonstrated in previous studies.21,22 We calculated the equivalized health insurance premium to take into account the size of the household by using the following equation.

In the 2009–2013 CHS, income data were calculated from self-reported monthly or yearly household income information, which was collected as a continuous variable. In the 2014 CHS, monthly household income was measured as a categorical variable with eight categories (less than 500,000 Korean won [KRW], 500,000 to 1,000,000 KRW, 1,000,000 to 2,000,000 KRW, 2,000,000 to 3,000,000 KRW, 3,000,000 to 4,000,000 KRW, 4,000,000 to 5,000,000 KRW, 5,000,000 to 6,000,000 KRW, and over 6,000,000 KRW). The categorized monthly household income in the 2014 CHS was converted to a continuous variable by taking the median value of each category. Other details of income data for the CHS were described elsewhere.11