This study analyzed the spatial distribution patterns of depression among vulnerable elderly across Republic of Korea. The average level of depression in the basic administrative districts was derived using the individual depression scores from the Health Interview Survey data. Results of the spatial autocorrelation analysis revealed that the Moran's I value was 0.3138, indicating the existence of a neighborhood effect in the depression of the vulnerable older adults at the regional level. Subsequently, cluster analysis and one-way ANOVA were conducted for the hot spots where vulnerable older adult depression was concentrated. Based on the cluster analysis results, hot spots were the areas where the facilities that are necessary for the daily lives of older adults were insufficient and were categorized into three types. The findings indicate that environmental characteristics at the regional level should be considered in addition to the environmental characteristics of the house and neighborhood, which have been primarily addressed in previous studies.
Keywords: cluster analysis; depression; hot spot; neighborhood effect; spatial autocorrelation.