In this study, we investigated the association between social environment characteristics and suicide rate in Korea, using nationwide longitudinal data covering all 229 districts over 11 years (2008–2018). We found that 1) poor socioeconomic conditions and isolation characteristics (higher “% population aged 65 and older eligible for the basic pension”, “% vacant houses in the area”,” % divorce”, and “% detached houses”) were associated with higher suicide rates; 2) higher smoking rates were associated with higher suicide rates; and 3) higher religious activity and greater access to recreational opportunities (sports facilities and parks) were associated with lower suicide rate. Associations with social environment characteristics were found to differ by age group; in general, the associations with socioeconomic status and health behavior characteristics were more evident in older age groups, whereas the associations with indicators of isolation and recreational opportunities were more prominent in younger age groups.
Lower socioeconomic status has been suggested as a major risk factor of suicide. The suicide rate increased during economic depressions [10, 11], and the higher suicide rates among people with lower income and education levels have been reported globally and consistently [12–15]. This study also showed that poor socioeconomic levels were associated with higher suicide rates at a district-level, as “% population aged 65 and older eligible for the basic pension” and “% vacant houses in the area” showed a positive association with suicide rate. Moreover, we found that the association between “% population aged 65 and older eligible for the basic pension” and suicide rate is the highest in the people aged 60 and older (Table 3). Previous studies also suggested that poverty and economic difficulties are dominant factors of suicide in the elderly, along with poor physical health [6, 16]. Thus, this result has crucial implications for suicide prevention policies in Korea, which showed a distinctively higher suicide rate in the elderly population than in other age groups.
Further, social isolation together with lower socioeconomic levels has been identified as another major risk factor of suicide [17–19]. Results from our study were consistent with this finding, as a positive association was found between suicide rate and “% divorce” and “% detached houses”. Moreover, we found that the effects of poor socioeconomic status and social isolation on suicide rate were more prominent in males than in females. We postulate that men’s higher levels of participation in economic activity in Korea might be related to this gender difference. According to the Korea National Statistics Office, the labour force in 2019 comprised 73.5% men and 53.5% women. In addition, these results can be partly explained by the gender difference in social relationship patterns. Previous studies revealed that males were more susceptible to social isolation than females [20, 21] and females usually have bigger social networks, receive more social support, and engage more actively in their social relationships than males [22–24]. Although further studies are required, our results suggest the need for gender-differentiated suicide prevention policies that focus on different social vulnerability factors.
This study also found that regional variables related to physical exercise and park availability can affect a reduction in the suicide rate. Numerous studies reported that increased levels of physical exercise lead to a reduction in stress and depressive disorders that may be related to suicide [25, 26]. Further, although existing results are mixed, a recent systematic review study reported a statistically significant negative association between physical activity and suicidal ideation . A Korean study also revealed that more physical activity is associated with less suicidal thoughts and attempts in adolescents . In addition, previous studies have reported that parks and green space provide positive effects, leading to fewer suicidal outcomes (suicide mortality, suicide ideation, and suicide attempts) [29–31], while improving health by encouraging physical and social activities [32, 33]. The effects of physical exercise and park availability on suicide rate in this study were more prominent in the youngest age group (aged 10–39) than in other, older, age groups. We considered that this may be associated with outdoor activity patterns in younger people. In other words, young people may be more likely to engage in outdoor and physical activities than older adults, and thus the average time spent using parks and sports facilities may also be higher. Future research is merited to further explore how physical exercise and park use affect suicide in relation to age, and these results can be useful for establishing effective suicide prevention policies for the young generation.
This study also found a positive association between “% current smokers” and suicide rate, and this association was more evident in the older populations (aged 40–59 and aged 60 and older), than in people aged 10–39. Previous studies have addressed smoking as one of the important risk factors of suicide  and reported that higher smoking was significantly associated with higher risks of suicidal ideation, planning, and attempt, as well as suicide death . We could not find a positive significant relationship between “% people exhibiting a high risk drinking” and suicide rate at a district level, and even the male population showed a negative association between high risk drinking behavior and suicide rate (Table 3). Because numerous existing studies consistently reported the detrimental effects of drinking on suicidal behavior [36, 37], the results of this study should be addressed carefully. Firstly, this result can be related to Korean socioeconomic culture. Korean men, especially young males, tend to build social capital mainly at the workplace and through economic activities , and a group dinner after work is a major part of the drinking culture of Korea. It implies that there is a possibility that more alcohol consumption at group dinners can be beneficial to develop social networks and therefore reduce the social isolation that may lead to an increase in suicide risk. Secondly, the results are estimates at a district level using aggregated data; thus, this association should be examined in greater detail in future studies.
Finally, this study found that the association between social environment characteristics and suicide rate was different by regional urbanicity, with the relationships between social environment characteristics and suicide rate being generally more evident in high-density areas (i.e. more urbanised areas) than other areas (i.e. less urbanised areas). We speculate that higher percentages of young and mid-aged populations present in more urbanised areas, which suggested the isolation experienced in areas with higher social activity levels might have a detrimental impact on the suicide rate as we have found there are more evident associations of suicide rate with isolation characteristics. In addition, the association between suicide rate and % basic pension was more evident in high- and mid-density areas than in low-density areas, although the “% population aged 65 and older eligible for the basic pension” was highest in low-density areas. We postulate this result can be related to larger relative deprivation in mid- and high-density areas; however, this study provided limited epidemiological evidence, and further investigations should be performed to determine the regional differences in relation to poor economic status of the population aged 65 years and older.
The study had several limitations. Firstly, as we mentioned earlier, the study results have a limited interpretation with respect to the individual-level association between social environment characteristics and suicide. Because the mortality data provided by the Korea National Statistics Office does not include individuals’ socioeconomic status and residential addresses, we were unable to examine the specific effects of individual-level socioeconomic status and individual-level environmental exposure data. Therefore, our study results reflected aggregated community-level results. Secondly, collection of several social environment characteristic variables (“% people who regularly participated in religious activities”, “% current smokers”, “% people exhibiting a high risk drinking”, “% population with recognized stress”, and “% population that is obese”) was limited to self-reporting, as these variables were obtained from the Korean Community Health Survey (KCHS) . Although previous studies have reported the good quality of self-reported data, and quality control assessments have been performed for KCHS , there may be underlying problems in misclassifications and recall bias. Given these possible shortcomings, our study needs to be complemented by data from future individual-level cohort studies.
Nevertheless, our study has some key strengths that can offset its limitations. Firstly, the study analyzed a large nationwide database of suicide deaths in Korea, with more than 154,866 cases over 11 years. Moreover, we collected data for a total of 12 annual social environment indicators of regional-level socioeconomic, demographic, urbanicity, general health behaviors, and other environmental characteristics, and analyzed the associations between these annual variables and suicide rates using advanced statistical methods. Finally, by sub-population analyses, we found distinct roles these social environment characteristics perform in reducing or increasing suicide rate measured across densification of areas, sex, and age. These results can be used for establishing evidence-based and targeted suicide prevention policies for each sub-population. To our knowledge, this is the largest study investigating the complex roles of social environment characteristics on suicide rate in Korea.