Geographical inequalities in suicide rates and area deprivation in South Korea.

BACKGROUND

Recent years have witnessed growing interest in determinants of suicide in South Korea. While much of the research has paid attention to the role of individual factors in explaining suicide mortality, another potentially important dimension of health inequality is geography.

AIMS OF THE STUDY

To describe the geographical distribution of suicide rates in South Korea, and to examine to what extent the distribution is shaped by area deprivation.

METHODS

Using 2004-2006 mortality data and 2005 population census data, the study first described the spatial mapping of age-standardised suicide rates and level of area deprivation across 250 districts in South Korea. A spatial lag model was also employed to examine the association between the two, taking into account the spatial dependence and interactions between neighbouring districts, as indicated by Moran's I and Geary' C statistics. In addition to level of area deprivation, population density, welfare spending, and rates of divorce, marriage and fertility at district-level were added to the model.

RESULTS

The spatial mapping of suicide rates across 250 districts exhibited a different pattern between males and females. While the highest suicide rates tended to cluster in the north-east region of South Korea for males, no clear pattern was observed for females. On the other hand, the lowest suicide rates for both males and females were found in most districts of the capital city. The results of the spatial lag model revealed a significant association between suicide rates and area deprivation, particularly for male suicide. Compared to the least deprived area, there were about 12-13 more male suicide cases (per 100,000 males) in highly deprived areas. The association with area deprivation was less clear for females. Population density was negatively associated with suicide rates for both males and females. While suicide rates were in general positively associated with divorce rates and negatively associated with fertility rates, the coefficients of both were not significant in the model for females.

DISCUSSION AND LIMITATIONS

The study highlighted substantial geographical variation in suicide rates across South Korea. Such variation can be largely explained by level of area deprivation, especially for males. The level of excess suicides (i.e. 12-13 per 100,000 males) in highly deprived areas was similar to the average suicide rates of OECD countries. Due to the paucity of individual-level data however, the study was unable to test whether or not these associations were independent of population composition.

IMPLICATIONS FOR HEALTH POLICIES

Prioritising development in relatively more deprived areas would potentially minimise resource barriers. Since such area-based approaches focus on a body with clear responsibility (i.e. local authority) and a greater relevance for local residents, programmes and services can be more effectively tailored to local needs and also more effectively delivered to local residents.

IMPLICATIONS FOR FURTHER RESEARCH

Alternative sources of data with individual-level data should be explored to strengthen the basis of the present findings, and also to examine the exact mechanism underlying the associations between suicide rates and area deprivation.