Accidents are the dominant cause of high mortality, morbidity, disability and loss of labor both in Korea and abroad. According to studies, the Preventable Death Rate (PDR) in Korea has been improving to 35.2% in 2010. However, it is still much higher compared to 24% in 2010 in England and Wales, U.K., and 15% in 2003 in Montana, U.S. Since 50% of patients’ deaths occur during pre-hospital emergency care, there are increasing efforts to improving EMS.
Previous studies related to EMS for reducing PDR can be categorized into two parts: focusing on the fairness, and on the efficiency of EMS. Focusing on the fairness of EMS, researchers analyzed the vulnerable areas and suggested how to redeploy resources such as emergency centers or fire stations. When focusing on the efficiency of EMS, many studies applied mathematical or simulation based models to suggest dispatching resources such as ambulances. In order to manage the limitative resources efficiently and prepare for emergency calls beforehand, researchers attempted to find the factors behind accidents. A lot of empirical studies considered a number of factors and some of them are sometimes conflicted. That is because each study considers different variables and when those variables come together, the results differ. Thus, multivariate analysis should be considered.
This paper therefore suggests a multivariate approach model to extract environmental contexts of frequent emergency calls and offers an application to the real case. According to the Routine Activity Theory (RAT), the components of an environmental context are divided into two dimensions: temporal and weather. Association rule mining is used as a tool for this study and Call-Context ratio which is a modified version of lift is calculated. The decision maker sets the threshold on the Call-Context ratio in order to select the environmental contexts that they will concern.
In the application results, the extracted environmental contexts were found to i...