Comparative Analysis of Patient Care Processes with Process Mining and Graph Edit Distance in Patient Data on the Indonesia National Health Insurance

Process mining is designed to support business process management in various sectors, such as manufacturing, finance, healthcare, etc. In the healthcare sector, process mining has been widely used in various case studies. This study will take sample data from the Health Social Security Administration Agency in 2015-2018. There are four types of sample data, namely Membership, Primary Health Facilities, Advanced Referral Health Facilities, and Non-Capitation Primary Level Health Facilities. Based on the sample data, this study aims to analyze the sequence of activities for each patient diagnosis using an inductive miner, which is one of the algorithms in process mining. The treatment process for each diagnosis is compared to one another using the Graph Edit Distance (GED) to see the similarities and differences between the treatment processes. This will be done to find out how efficient a treatment process is for all disease diagnoses from patients based on existing sample data by looking at the values of trace fitness, precision, generalization, and simplicity from process models for each diagnosis. The purpose of this paper is to look for differences or similarities in the process of patient treatments from the Health Social Security Administration Agency based on the diagnosis of one disease with another. The result of this study enables us to identify treatment processes with similarities based on the weighted node graph. Additionally, it provides insight into the relative efficiencies of different treatment processes, highlighting areas that should be improved.